Characterization of the cis-Acting Contributions to Avian Hepadnavirus RNA Encapsidation

Abstract

Previous analysis of duck hepatitis B virus (DHBV) indicated the presence of at least two cis-acting sequences required for efficient encapsidation of its pregenomic RNA (pgRNA), ɛ and region II. ɛ, an RNA stem-loop near the 5′ end of the pgRNA, has been characterized in detail, while region II, located in the middle of the pgRNA, is not as well defined. Our initial aim was to identify the sequence important for the function of region II in DHBV. We scanned region II and the surrounding sequence by using a quantitative encapsidation assay. We found that the sequence between nucleotides (nt) 438 and 720 contributed to efficient pgRNA encapsidation, while the sequence between nt 538 and 610 made the largest contribution to encapsidation. Additionally, deletions between the two encapsidation sequences, ɛ and region II, had variable effects on encapsidation, while substitutions of heterologous sequence between ɛ and region II disrupted the ability of the pgRNA to be encapsidated efficiently. Overall, these data indicate that the intervening sequences between ɛ and region II play a role in encapsidation. We also analyzed heron hepatitis B virus (HHBV) for the presence of region II and found features similar to DHBV: a broad region necessary for efficient encapsidation that contained a critical region II sequence. Furthermore, we analyzed variants of DHBV that were substituted with HHBV sequence over region II and found that the chimeras were not fully functional for RNA encapsidation. These results indicate that sequences within region II may need to be compatible with other viral components in order to function in pgRNA encapsidation.

Hepadnaviruses are liver-tropic, small DNA viruses that replicate through an RNA intermediate called the pregenomic RNA (pgRNA) (for a review, see references 7 and 22). Hepadnaviruses characteristically cause acute and chronic hepatitis in their hosts. Members of the hepadnavirus family include the prototype, human hepatitis B virus (HBV), as well as the distantly related avian viruses duck hepatitis B virus (DHBV) and heron hepatitis B virus (HHBV). Although they are both avihepadnaviruses, DHBV and HHBV share only 79% nucleotide identity. Studying the avian hepadnaviruses has provided significant information about the molecular mechanisms of replication for this family of viruses.

When a hepadnavirus infects a hepatocyte, its relaxed circular DNA genome is deposited into the nucleus and converted into covalently closed circular DNA (26). The pgRNA is transcribed from covalently closed circular DNA, and pgRNA is longer than genome length and contains a terminal redundancy (4) (Fig. 1). Once exported from the nucleus, the pgRNA is the mRNA for the viral polymerase (P) and core (C) proteins (6, 21). Additionally, pgRNA is encapsidated into capsids along with the P protein (1, 11), and the pgRNA becomes the template for reverse transcription (23). DNA synthesis occurs within capsids, and capsids containing mature viral DNA can be secreted by the cell as virions (29).

FIG. 1.

The DHBV pgRNA is the mRNA for the translation of the viral C and P genes and is the template for reverse transcription. At top are shown the nucleotide coordinates of the C and P open reading frames. At bottom are shown the nucleotide positions of features on the pgRNA. The 5′ end of the RNA is at nt 2529, and terminal redundancies (designated R) are approximately 270 nt. ɛ is an encapsidation signal located at the 5′ end of the pgRNA, and the ɛ secondary structure is located at nt 2560 to 2617 (3). Region II is a second encapsidation signal in the middle of the pregenome between nt 551 and 719 (5).

There are several requirements for the selective encapsidation of the hepadnavirus pgRNA. The P protein is essential for pgRNA encapsidation (1, 11), and it has been demonstrated that P interacts with ɛ, an encapsidation signal on the pgRNA, to facilitate encapsidation (19). ɛ is located near the 5′ end of the pgRNA and forms a phylogenetically conserved secondary structure that is important for encapsidation (14, 18, 19). ɛ also contains the origin for minus-strand DNA synthesis (17, 24, 27). Although the P protein is critical for encapsidation, DNA synthesis is not required for encapsidation (1, 11, 28, 30). Mutations that eliminate the ability of P to synthesize DNA, such as substitutions of the catalytic domain or of the amino acid that primes minus-strand DNA, do not disrupt encapsidation.

Both HBV and DHBV have ɛ encapsidation elements that interact with the P protein to encapsidate pgRNA. The HBV ɛ is within the first 132 nt of the pgRNA, and fusing HBV ɛ to a heterologous sequence (lacZ) was sufficient for the encapsidation of lacZ into HBV capsids (14, 18). On the other hand, the cis-acting requirements of DHBV pgRNA encapsidation are more complex. Hirsch and colleagues (12) demonstrated that DHBV ɛ was not sufficient for encapsidation. They showed that the first 136 nt of the DHBV pregenome, which contains ɛ, fused onto lacZ RNA was not encapsidated. To define a contiguous sequence that was necessary and sufficient for RNA encapsidation, they analyzed a series of DHBV/lacZ fusion RNAs with progressively larger 5′ segments of DHBV pregenome appended to lacZ. They found that only fusions containing at least the first 1,200 nt of DHBV were encapsidated. By analyzing a series of overlapping deletions within DHBV, Calvert and Summers (5) identified a second cis-acting requirement for DHBV encapsidation. This requirement, called region II, was mapped to a 200-nt region in DHBV located 900 nt 3′ of ɛ. (Fig. 1). These experiments indicated that DHBV has two discrete encapsidation elements that are required for RNA encapsidation.

It is not clear how region II functions in encapsidation. In order to gain insight into the function of region II, we used a quantitative analysis to analyze deletion variants within and surrounding region II. We characterized sequences which contribute to encapsidation and identified a sequence from nt 539 to 610 that significantly contributed to encapsidation. Further quantitative analysis revealed that deletions and substitutions of heterologous sequence between ɛ and region II had different effects on encapsidation. We identified similar cis-acting contributions to encapsidation for HHBV and examined the ability of HHBV region II to functionally substitute for the DHBV region II. Overall, these data characterize the cis requirements for avian hepadnavirus encapsidation and demonstrate their complex nature.

MATERIALS AND METHODS

Plasmid construction.

Details of the plasmid construction will be provided upon request. All variants were given names to describe the mutations they contain. For example, Δ438-487 is a DHBV variant that has nt 438 to 487 deleted. The DHBV deletion variants (Fig. 2 C and D) Δ438-487, Δ488-537, Δ538-587, Δ588-637, Δ638-687, Δ688-737, Δ738-787, Δ720-748, Δ720-906, Δ505-524, Δ515-538, Δ539-551, Δ552-575, Δ576-599, Δ600-623, Δ610-629, and Δ621-640 were generated in the parental plasmid p503 (9). p503 contains 1.5 tandem copies of the DHBV3 genome and expresses a P-null version of DHBV due to a 1-nt deletion at nt 424 in the P gene. The variant Δ2548-2580 also contains this P-null mutation.

FIG. 2.

DHBV nt 438 to 720 contributes to efficient encapsidation. (A) Test plasmids (wild-type reference or deletion variants) were cotransfected with an internal standard plasmid. All test plasmids (P⁻ C⁺) expressed the core protein, but they did not express P because of a frameshift at nt 424 which is upstream of the deletion mutation in all variants. Deletions were introduced between nt 438 and 906, as indicated. The internal standard plasmid (DHBV P^Y96F C⁻) provided a DNA synthesis mutant of P. The probe in the RPA contained minus-sense DHBV sequences from 2810 to 3021/1 to 84. The probe annealed the target RNA over the C gene variation, and different-size fragments were protected by the test RNAs (295 nt) and the internal standard RNAs (260 nt). (B) RPA analysis of a series of deletions across DHBV region II. Cytoplasmic poly(A) RNA (fraction A) and capsid RNA (fraction C) were isolated from cotransfected LMH cells. RPA was performed, and products were electrophoresed in a 5% acrylamide gel. Lane 1, undigested probe which represented 1/20 of the amount used in the RPA; lane 2, digested probe; lanes 3 and 4, digestion of a test RNA alone and the internal standard RNA alone, respectively, isolated from the poly(A) RNA fraction of transfected LMH cells; lanes 5 to 20, protected fragments from the A fractions (odd numbered lanes) and the C fractions (even numbered lanes) of the indicated test plasmid cotransfected with the internal standard plasmid. WT, wild type. (C) Encapsidation efficiency of variants containing deletions from nt 438 to 901. Encapsidation efficiency was calculated as the level of test RNA in the C fraction normalized to the level of internal standard (IS) RNA in the C fraction divided by the level of test RNA in the A fraction normalized to the level of internal standard RNA in the A fraction, or (C_test/C_IS)/(A_test/A_IS). The encapsidation level of the DHBV deleted variants was normalized to that of the WT reference and plotted as the percent encapsidation efficiency of the WT reference. (D) Encapsidation efficiency of DHBV variants containing smaller deletions from nt 505 to 640. (C and D) The results represent the means and standard deviations (error bars) of three (C) and two (D) independent analyses.

The DHBV/lacZ and DHBV/β-globin variants (see Fig. 4A) were in a DHBV P⁻ C⁻ background. DHBV P⁻ C⁻ contains the same P-null mutation as p503 and has a 4-nt deletion from nt 2846 to 2849 which destroys an NsiI site and creates a C-null mutation (11). The substituted variants were named based on the DHBV positions where the heterologous sequences were substituted. For example, β-globin 2672-420 has a 770-nt segment of β-globin sequence replacing DHBV from nt 2672 to 3021/1 to 420. β-Globin sequence was from the Glob-GAPDH-Glob plasmid (10), and nt 1 of β-globin corresponds to the 5′ terminus of the human β-globin mRNA. lacZ sequence was from pON249 (8), and nt 1 of lacZ represents the A of codon 1 of the lacZ gene. β-Globin 2672-420, β-globin 2672-110, LacZ 2672-38, LacZ 33-420, and β-globin 108-420 contain a novel XhoI site at nt 2672 created by site-directed mutagenesis. β-Globin 2672-420 was generated by replacing DHBV nt 2672 to 3021/1 to 420 with β-globin nt 1 to 769. β-Globin 2672-110 has β-globin nt 1 to 462 replacing DHBV nt 2672 to 3021/1 to 110. LacZ 2672-38 was generated by substituting DHBV nt 2672 to 3021/1 to 38 with LacZ nt 51 to 438. LacZ 33-420 has lacZ sequence from nt 433 to 819 replacing DHBV nt 33 to 420. β-Globin 108-420 has β-globin sequence from nt 457 to 769 replacing DHBV nt 108 to 420. β-Globin 906-1658 contains β-globin sequence from nt 1 to 769 replacing the DHBV sequence from PflFI (nt 901) to BamHI (nt 1658) in the parental vector p503.

FIG. 4.

Substitutions between ɛ and region II cause encapsidation defects. (A) β-Globin or lacZ sequence was substituted into the P⁻ C⁻ DHBV parental vector at the indicated DHBV positions to generate a series of DHBV substitution variants. (B) RPA of DHBV substitution variants. RPA was performed as described in Fig. 3A. Lanes 1 to 4 contain the indicated controls; lanes 5 to 16 contain the RPA results of the isolated A and C fractions from the indicated test plasmid cotransfected with the internal standard (IS). WT, wild type; β-glo, β-globin. (C) Encapsidation efficiency of the substitution variants normalized to the encapsidation efficiency of the WT reference and plotted as the percent encapsidation efficiency of the WT reference. The results represent the means and standard deviations (error bars) of three independent analyses.

The variants with sequences between ɛ and region II deleted (Fig. 3), Δ2672-109, Δ2672-36, Δ33-425, and Δ2672-424, were created from the similar substitution variants. Δ2672-109 was made by removing the XhoI (nt 2672)-to-EagI (nt 108) β-globin fragment from β-globin 2672-110. Δ2672-36 was made by removing the XhoI (nt 2672)-to-HincII (nt 32) LacZ fragment from LacZ 3672-38. Δ33-425 was made by removing the HincII (nt 33)-to-PstI (nt 420) fragment from LacZ 33-420. Δ2672-424 was created by removing the XhoI (nt 2672)-to-PstI (nt 420) fragment from β-globin 2672-420.

FIG. 3.

Deletion of the intervening sequences between ɛ and region II has variable effects on encapsidation. (A) Deletions were introduced within nt 2672 to 3021/1 to 425, as indicated. The test plasmids (P⁻ C⁻ WT reference or the deleted variants) were cotransfected with an internal standard (DHBV P^Y96F C⁺) that provided the replication proteins. Additionally, the P^Y96F mutation within the internal standard contained 5 out of 7 nt substituted from nt 451 to 457 that created a difference between the test RNAs and the internal standard RNAs. The probe annealed to DHBV sequences from nt 420 to 720, and the protected fragments of the test RNAs and the internal standard RNAs were different sizes (300 and 263 nt, respectively). (B) Encapsidation efficiency of variants containing deletions between ɛ and region II. The results represent the means and standard deviations (error bars) of three independent analyses. WT, wild type.

The HHBV variants (see Fig. 5B) Δ44-243, Δ198-397, and Δ326-525 were created using the parental plasmid HHBV P⁻ C⁻. HHBV P⁻ C⁻ is a C-null and P-null derivative of 413-2, a plasmid that expresses wild-type HHBV (16). The C-null mutation of HHBV P⁻ C⁻ is a 1-nt deletion at nt 2690 which introduces a premature termination codon in the core gene, and the P-null mutation is a 1-nt substitution at nt 182 from T to A which generates a premature termination codon in P. The HHBV variants (see Fig. 5B) Δ427-626, Δ562-761, Δ694-893, Δ832-1031, and Δ965-1164 contain deletions that were introduced into the parental vector HHBV P⁻ C⁺. HHBV P⁻ C⁺ is a P-null derivative of 413-2 that contains the same P-null mutation as HHBV P⁻ C⁻. Δ2574-2580, or ENHHBV, has been described previously (16). The HHBV/DHBV chimeras (see Fig. 6A) have HHBV sequence replacing the analogously located DHBV sequence. The chimeras HHBV 426-720, HHBV 426-901, HHBV 426-1364, and HHBV 906-1364 contain HHBV4 substitutions within the indicated sequences. The variant HHBVɛ contains HHBV sequences from nt 1787 to 2652 replacing the DHBV sequence from nt 1660 to 2646 in the 5′ copy of the redundancy in the DHBV plasmid p503. The double chimera, HHBVɛ and HHBV region II, has the HHBV substitution from nt 426 to 720 in the parental construct HHBVɛ.

FIG. 5.

HHBV also has a region II encapsidation signal. (A) The HHBV test plasmids (the P⁻ C⁺ or P⁻ C⁻ wild-type reference or the HHBV variants) were cotransfected with an internal standard plasmid that provided the HHBV C and P^Y96F proteins (HHBV P^Y96FC⁺). The probe in the RPA annealed HHBV sequences from nt 2840 to 3027/1 to 44, which overlapped a region containing a variation between the test RNA and internal standard RNA (see Materials and Methods). Therefore, the HHBV test RNAs and the HHBV internal standard RNAs protected different-size fragments in the RPA. (B) Encapsidation efficiency of the HHBV deletion variants normalized to the wild-type reference and plotted as the percent encapsidation efficiency of the wild-type reference (WT Ref.). The results represent the means and standard deviations (error bars) of three independent analyses.

FIG. 6.

DHBV with an HHBV region II has an encapsidation defect. (A) The three chimeras, HHBV 426-720, HHBV 426-901, and HHBV 426- 1364, have HHBV region II substituting for the analogous DHBV sequence. Another chimera had a substitution within a sequence that is not required for encapsidation (HHBV 906-1364). The pgRNA of HHBVɛ contains an HHBV substitution from nt 2535 to 2652, and the double chimera contains HHBVɛ and HHBV 426-720. All test chimeras were P⁻ C⁺ and were cotransfected with the DHBV internal standard as described in Fig. 2A. (B) Representative RPA of chimeric analysis. Lanes 1 to 4 contain the indicated controls; lanes 5 to 12 contain the RPA results from the A and C fractions of the indicated test plasmid and internal standard (IS) plasmid cotransfection. WT, wild type. (C) Encapsidation efficiency of the chimeras was normalized to the encapsidation efficiency of the WT reference and expressed as the percent encapsidation efficiency of the WT reference. The results represent the means and standard deviations (error bars) of three independent analyses.

Two different internal standards were used in the experiments with DHBV. Both were expressed from plasmids containing 1.5 tandem copies of DHBV3. The plasmid called DHBV P^Y96F C⁻ (Fig. 2A) expressed a pregenome that contains the same C-null mutation as DHBV P⁻ C⁻ and expressed a P protein with amino acid 96 changed from tyrosine to phenylalanine by the substitution CTA to GTT beginning at nt 454. The C-to-G change at nt 454 was silent in the leucine codon at amino acid 95. The internal standard called DHBV P^Y96F C⁺ (Fig. 3A) expressed a pregenome with an intact C gene and expressed a P protein with amino acid 96 changed from tyrosine to phenylalanine by substituting 5 nt beginning at nt 451, from TCTCTAT to CTTGTTC. The substitutions from nt 451 to 454 did not alter the sequence of amino acids 94 or 95 of P. These substitutions introduced a difference between the internal standard and test RNAs that could be distinguished in the RNase protection assay (RPA). The internal standard used in the HHBV analysis (see Fig. 5A) was expressed from a plasmid containing 1.4 tandem copies of HHBV4, 413-2 (16). This internal standard was called HHBV P^Y96F C⁺ and contained two different substitutions. It expressed a wild-type C protein and contained a 5-nt substitution beginning at nt 2871 (CCTCCGA to TTTAAGG) that did not alter the amino acid sequence of C. These substitutions introduced a difference that could be distinguished in the RPA. The second mutation in HHBV P^Y96F C⁺ substituted 4 nt beginning at nt 452 (TTGTAT to CTCTTC) and changed amino acid 96 in P protein from a tyrosine to phenylalanine. Finally, all constructed plasmids were confirmed by DNA sequencing to have the desired mutations.

Cell cultures and transfection.

The chicken hepatoma cell line, LMH, was cultured as described previously (15). DNA transfections were performed using the calcium phosphate precipitation method. In each cotransfection, 7 μg of test plasmid and 3 μg of internal standard plasmid were transfected.

Isolation of cytoplasmic poly(A) RNA and capsid RNA.

Three days posttransfection, LMH cells were washed with HBS-EGTA (2 mM HEPES [pH 7.45], 150 mM NaCl, and 0.5 mM EGTA). Cells were lysed with Nonidet P-40 lysis buffer (10 mM Tris [pH 8.0], 1 mM EDTA, 0.2% Nonidet P-40). Half of the lysate was used to isolate the cytoplasmic polyadenylated [poly(A)] RNA, and the other half of the lysate was used to isolate the capsid RNA. For the isolation of poly(A) RNA, the cytoplasmic lysate was adjusted to 10 mM Tris-HCl (pH 7.5), 0.5 M NaCl, 1 mM EDTA, 0.5% sodium dodecyl sulfate (SDS), and proteinase K (100 μg/ml) and incubated at 37°C for 30 min. Next, oligo(dT) cellulose (Roche) was added, and samples were incubated at room temperature for 1 h. Oligo(dT)-bound RNA was washed three times with solution I (10 mM Tris-HCl [pH 7.5], 0.5 M NaCl, 1 mM EDTA, 0.5% SDS) and once with solution II (10 mM Tris-HCl [pH 7.5], 0.1 M NaCl, 1 mM EDTA, 0.5% SDS). Poly(A) RNA was eluted twice with 200 μl of TE (10 mM Tris-HCl [pH 7.5], 1 mM EDTA) and stored at −20°C in 0.3 M sodium acetate and ethanol.

For the isolation of capsid RNA, the cytoplasmic lysate was adjusted to 5 mM CaCl₂, and 22 U of micrococcal nuclease (Boehringer) was added. The samples were incubated at 37°C for 90 min. The solution was adjusted to 10 mM EDTA, 0.2% SDS, 50 mM NaCl, and pronase (0.2 mg/ml) to inactivate the micrococcal nuclease and digest proteins. The capsid RNA was extracted once with phenol-chloroform, once with chloroform, and stored in ethanol.

RPA.

Riboprobes were transcribed in vitro from linearized DNA templates using T7 or T3 RNA polymerase and labeled with [³²P]UTP. Riboprobes were purified from 5% polyacrylamide-7.6 M urea gels and eluted in a solution of 0.5 M ammonium acetate, 1 mM EDTA, and 0.1% SDS.

An aliquot equivalent to 1/20 of the transfected plate of cytoplasmic poly(A) RNA and capsid RNA was coprecipitated with approximately 10⁵ cpm of purified riboprobe. Pellets were resuspended in 10 μl of hybridization buffer (40 mM PIPES [pH 6.8], 1 mM EDTA [pH 8.0], 0.4 M NaCl, 80% deionized formamide), heated at 95°C for 3 min, and incubated overnight at 42°C. Next, 150 μl of digestion mix (300 mM NaCl, 10 mM Tris [pH 7.5], 5 mM EDTA, RNase A [7 ng/μl], and RNase T₁ [7 U/μl]) was added, and samples were incubated for 30 min at 37°C. For a negative control with probe only, no RNases were added to the digestion buffer. After digestion, RNases were inactivated with 160 μl of solution D (4 M guanidinium thiocyanate, 25 mM sodium citrate, 0.5% sodium lauryl sarcosinate, and 0.1 M β-mercaptoethanol), and RNA was precipitated with 15 μg of glycogen, 75 μl of ethanol, and 325 μl of isopropanol. Pellets were resuspended in formamide loading buffer, heat denatured at 95°C for 3 min, and loaded onto a 5% polyacrylamide-7.6 M urea sequencing gel. After electrophoresis, gels were dried and analyzed using a Molecular Dynamics Storm apparatus and ImageQuant software.

RESULTS

Quantitative analysis of RNA encapsidation.

We devised an experimental strategy to accurately and reliably measure the level of encapsidation of pgRNA variants. Figures 2A, 3A, and 5A show the three strategies used in these experiments. In general, LMH cells were cotransfected with a test plasmid and an internal standard plasmid. The test plasmid expressed either the pgRNA of the wild-type reference or of a variant whose encapsidation efficiency was being measured. Since deletions or substitutions were introduced into sequences that contained the P gene, a premature stop codon was introduced in the 5′ portion of the P gene so that the variant pgRNAs did not express P protein (P null or P⁻). Also, since some variants contained deletions or substitutions in the C gene, some variants had a mutation that eliminated C protein synthesis (C null or C⁻). The wild-type reference in each analysis was a derivative of wild-type pgRNA that contained either the P-null mutation (P⁻ C⁺) or the C-null and P-null mutations (P⁻ C⁻), as specified below. The internal standard plasmid expressed a pgRNA that was encapsidation competent and provided either P protein or both P and C proteins. The P protein expressed from the internal standard plasmid had the tyrosine at position 96 changed to a phenylalanine. The variant protein, P^Y96F, is functional for encapsidation but cannot support the priming of minus-strand DNA synthesis (28, 30). Also, Southern blotting analysis of intracellular capsids did not detect DNA for DHBV and HHBV variants expressing P^Y96F (data not shown). Consequently, the loss of RNA-containing capsids resulting from DNA synthesis did not occur in our experiments. Three days after the transfection, cytoplasmic poly(A) RNA (fraction A) and capsid RNA (fraction C) were isolated from each plate of cells. The levels of test RNA and internal standard RNA were determined in each A and C fraction using an RPA. In the RPA, we used a probe that annealed to a region that contained a difference between the test and internal standard RNAs. Thus, different-size fragments were protected by the test RNA and the internal standard RNA. Test RNA alone or internal standard RNA alone isolated from the cytoplasmic poly(A) fraction of transfected LMH cells was included in each RPA to demonstrate that RNase digestion was complete. Additionally, the probe detected only pgRNA and not any subgenomic RNA. The encapsidation efficiency was calculated as the level of test RNA normalized to internal standard RNA detected in the capsid RNA fraction (fraction C) divided by the level of test RNA normalized to internal standard RNA detected in the cytoplasmic poly(A) fraction (fraction A). The encapsidation efficiency of the variants was expressed as the percent encapsidation efficiency of the wild-type reference.

Encapsidation analysis of DHBV variants containing deletions within region II.

Previous studies located the DHBV region II encapsidation signal between nt 551 and 719 (5). In order to map region II more precisely, we generated a series of deletions from nt 438 to 906 in DHBV3. Figure 2A shows the RPA strategy used to analyze the DHBV variants with deletions from nt 438 to 906. Since the region II variants had deletions within the P gene, all test plasmids were null for P protein production, and the internal standard plasmid, DHBV P^Y96F C⁻, donated P^Y96F in trans. Additionally, the test plasmids expressed the C protein, but the internal standard had an early termination codon in the C gene due to a 4-nt deletion at nt 2846. The RNA probe annealed to a 295-nt region containing the C gene variation between the test and internal standard RNAs from nt 2810 to 3021/1 to 84. Thus, the probe fragments protected by the test RNA and internal standard RNA were different sizes, as shown in Fig. 2B (295 and 260 nt, respectively).

Figure 2C shows the quantification of the encapsidation efficiency of the deleted variants. Variants with deletions 3′ of nt 720 (Δ738-787, Δ720-748, and Δ720-906) were not defective for encapsidation. On the other hand, Δ438-487, Δ488-537, Δ538-587, Δ588-637, Δ638-687, and Δ688-737 encapsidated RNA at lower levels than did the wild-type reference. However, variants containing deletions within the middle of the region analyzed had more-severe defects in encapsidation than variants containing deletions within the sequences flanking the middle. The variants with deletions in the middle, Δ538-587 and Δ588-637, encapsidated RNA at 28% ± 7% and 23% ± 3% of the level of the wild-type reference, respectively (values are presented as means ± standard deviations). The magnitude of the encapsidation defect for Δ538-587 and Δ588-637 was similar to that for a variant containing a deletion in ɛ (Δ2549-2580 [Fig. 2C]). Δ2549-2580 contains the same deletion in ɛ as a previously described variant (called Δ5 in reference 12) that was interpreted as encapsidation negative. Variants with deletions flanking nt 538 to 637, or Δ488-537 and Δ638-687, were not as defective in encapsidation as was the ɛ mutant. These results indicate that DHBV sequences from nt 438 to 719 contributed to encapsidation, but nt 538 to 637 made a significant contribution to encapsidation, because deleting these sequences had a similar effect on encapsidation as did deleting sequences within ɛ.

Next we analyzed the sequences within and adjacent to nt 538 to 637 by making a series of smaller deletions across this region (Fig. 2D). Δ539-551, Δ552-575, Δ576-599, and Δ600-623 were the variants most defective for RNA encapsidation; they encapsidated RNA at 35% ± 6%, 24% ± 1%, 24% ± 2%, and 22% ± 1% of the level of the wild-type reference, respectively. Variants with deletions either 5′ (Δ505-524 and Δ515-538) or 3′ (Δ610-629 and Δ621-640) of this region had less severe encapsidation defects. The results of the experiments in Fig. 2D demonstrate that sequences from nt 505 to 640 are important for efficient DHBV encapsidation. However, when deleted, sequences from nt 539 to 610 have a dramatic effect on RNA encapsidation, and therefore, they define a critical sequence of region II.

Encapsidation analysis of DHBV variants containing deletions between ɛ and region II.

We examined RNAs containing deletions between ɛ and region II to determine if the intervening sequences contributed to DHBV encapsidation. The DHBV variants with deletions between ɛ and region II were analyzed using the RPA strategy in Fig. 3A. Since all variants had deletions in the C gene and/or P genes, the test RNAs were complemented in trans with C and P^Y96F from the internal standard. The test RNAs and internal standard RNA were different from each other over nt 451 to 457 as a result of a 5-nt substitution in the internal standard. The probe, which was from nt 420 to 720, protected different-size RNA fragments from the test RNA (300 nt) and internal standard RNA (263 nt). The results demonstrate that particular deletions between ɛ and region II have different effects on encapsidation (Fig. 3B). For example, Δ2672-109 (459 nt deleted), Δ2672-36 (386 nt deleted), and Δ33-425 (393 nt deleted) encapsidated RNA at a level similar to the wild-type reference, DHBV P⁻ C⁻. These smaller deletions indicate that the 3′ boundary of ɛ is located upstream of nt 2672 and the 5′ boundary of region II is located downstream of nt 425. However, complete removal of these intervening sequences impaired encapsidation. Δ2672-424 (774 nt deleted) encapsidated at 32% ± 7% of the level of the wild-type reference, which is similar to the encapsidation efficiency of an ɛ or region II deletion variant. Therefore, complete removal of the intervening sequences between ɛ or region II affects RNA encapsidation, but smaller deletions within the intervening sequences have no effect on encapsidation.

Substitution of the sequence between ɛ and region II disrupts encapsidation.

The reduced encapsidation of Δ2672-424 indicated that there could be a distance requirement between ɛ and region II. Substituting heterologous sequence between ɛ and region II would remove the intervening DHBV sequences but maintain the size of the pregenome. Therefore, we made a variant with β-globin sequence substituting the intervening DHBV sequence from nt 2672 to 3021/1 to 420 (β-globin 2672-420). Additionally, we made several other substitution variants with β-globin or LacZ sequence replacing the DHBV sequence between ɛ and region II, as shown in Fig. 4A, in order to understand the effect of substitutions between ɛ and region II. Additionally, we generated a variant with a β-globin sequence substitution 3′ of region II (β-globin 906-1658). Previous studies have shown sequences in DHBV 3′ of nt 901 are disposable for encapsidation (12). Northern blotting of the cytoplasmic poly(A) RNA expressed by the variants depicted in Fig. 4A indicated that all synthesized a pgRNA of the correct size (data not shown). Next, the substituted variants were cotransfected with the internal standard, DHBV P^Y96F C⁺, and RPA was performed based on the strategy shown in Fig. 3A.

Restoring the size of Δ2672-424 by replacing the sequence between ɛ and region II with heterologous sequence did not restore the ability of the RNA to be encapsidated (Fig. 4B and C). The variant β-globin 2672-420, containing a 770-nt substitution, was deficient for encapsidation. Substituting the identical β-globin sequence 3′ of region II, as in β-globin 906-1658, did not have a negative effect on encapsidation. This result indicated that the β-globin sequence did not have a general negative effect on encapsidation. Surprisingly, variants with smaller substitutions between ɛ and region II (β-globin 2672-110 [460-nt substitution], LacZ 2672-38 [388-nt substitution], LacZ 33-420 [388-nt substitution], and β-globin 108-420 [313-nt substitution]) were similarly deficient in encapsidation. In fact, all of these variants, except β-globin 108-420, encapsidated RNA at a level similar to variants containing deletions within ɛ (Δ2549-2580, Fig. 2C) or within the critical sequence of region II from nt 539 to 610 (Fig. 2D). Conversely, variants containing deletions at nearly identical positions on the pregenome did not cause defects in encapsidation (Fig. 3B, Δ2672-109, Δ2672-36, and Δ33-425). Overall, these data demonstrate that replacing the DHBV sequence between ɛ and region II with heterologous sequence disrupts encapsidation.

HHBV has encapsidation requirements similar to DHBV.

In order to understand if region II is a general feature of avian hepadnaviruses, we examined the encapsidation of variants of HHBV, the avihepadnavirus that is most distantly related to DHBV (20, 25). We made a series of overlapping 200-nt deletions in HHBV from nt 44 to 1164, which includes the sequence located analogously to region II of DHBV (Fig. 5). Within this region of the genome, the nucleotide coordinates represent the same position on DHBV and HHBV. The strategy used to examine these HHBV variants (Fig. 5A) was similar to that described for the DHBV variants. In particular, the test RNAs with deletions in the C and P genes (Δ44-243, Δ198-397, and Δ326-525) were null for C and P production, and the test RNAs with deletions in only the P gene (Δ427-626, Δ562-761, Δ694-893, Δ832-1031, and Δ965-1164) were null only for P production. The HHBV internal standard, HHBV P^Y96F C⁺, provided P^Y96F and C. Additionally, the C gene of the internal standard contained five silent substitutions at nt 2870 that created a difference to distinguish between internal standard and test RNAs in the RPA.

We quantified the encapsidation efficiency of the HHBV variants (Fig. 5B). Interestingly, we observed trends in the HHBV deletion variants that were similar to the observations we made for the deletion variants of DHBV (compare Fig. 5B with 2C). The HHBV variants with deletions 3′ of 832, Δ832-1031 and Δ965-1164, encapsidated RNA at levels similar to the wild-type reference. However, all HHBV variants with 200-nt deletions 5′ of nt 832 encapsidated RNA less efficiently than the wild-type reference. The HHBV variants with deletions in the 5′ region, Δ44-243 and Δ198-397, had the least-severe defects in RNA encapsidation; these variants encapsidated RNA at 71% ± 7% and 64% ± 6% of the level of the wild-type reference, respectively. Variants with deletions in the middle of this region, Δ326-525, Δ427-626, Δ562-761, and Δ694-893, encapsidated RNA at lower levels. The magnitude of the encapsidation defect for deletions within this region was similar to that of an HHBV ɛ deletion mutant, Δ2574-2580 (Fig. 5B). Overall, this analysis defines a region within nt 398 to 831 required for efficient HHBV encapsidation which is analogous to region II of DHBV. Also similar to DHBV, sequences upstream of region II are required for encapsidation, but deleting these sequences has less drastic effects on encapsidation. Thus, DHBV and a distantly related avihepadnavirus, HHBV, share similar requirements for efficient encapsidation.

DHBV/HHBV chimeras test the function of region II.

A well-established function of ɛ in encapsidation is to interact with P and promote the assembly of replication competent capsids. It is not understood how region II contributes to encapsidation. Possibly, region II functions by interacting with other viral components, such as interacting with other viral sequences or binding to viral proteins. In order to gain an understanding of how region II might be functioning, we tested whether the HHBV region II could functionally substitute for the DHBV region II. If region II functions by interacting with other viral components in a species-specific manner, then replacing the region II of DHBV with that of HHBV might interfere with the ability of region II to function.

We created region II chimeras with HHBV sequence from nt 426 to 720, nt 426 to 901, and nt 426 to 1364 substituted into DHBV (Fig. 6A). Also, a chimera was tested which had an HHBV substitution into DHBV within a region that is not required for encapsidation (HHBV 906-1364). The nucleotide identity between DHBV and HHBV in each of these regions is as follows: nt 426 to 720, 76%; nt 426 to 901, 74%; nt 426 to 1364, 64%; and nt 906 to 1364, 56%). The strategy to analyze the chimeras was the same as described in Fig. 2A. All chimeras produced DHBV C protein, and the DHBV P^Y96F protein was provided from the internal standard plasmid. It was found that replacing the DHBV region II sequence with HHBV region II sequence caused measurable defects in encapsidation (Fig. 6B and C). Chimeras HHBV 426-720 and HHBV 426-901 encapsidated RNA at levels that were 60% ± 10% and 68% ± 4% of the wild-type reference, respectively. On the other hand, the chimera with an HHBV substitution in a region that was not required for encapsidation, HHBV 906-1364, encapsidated RNA at 122% ± 12% of the wild-type reference. The largest substitution, HHBV 426-1364, encapsidated RNA more efficiently than HHBV 426-720 and HHBV 426-901, but encapsidation was still not at wild-type levels. These results demonstrate that the HHBV region II sequence that was substituted does not completely substitute for the DHBV region II. If region II interacts with ɛ in order to function, then adding back HHBV ɛ to a region II chimera could restore encapsidation efficiency of the chimera. Figure 6B and C show that this is not the case. The chimera with only HHBV ɛ, HHBVɛ, encapsidated at wild-type levels, but the double chimera with HHBVɛ and region II had a defect in encapsidation similar to the region II chimera. Therefore, HHBV region II is not completely functional in DHBV, possibly due to an incompatibility of HHBV region II with one or more DHBV encapsidation components.

DISCUSSION

Earlier studies of hepadnavirus encapsidation identified important cis-acting sequences for hepadnavirus encapsidation. For our analyses, we developed a quantitative assay to measure the level of encapsidation of hepadnavirus variants. Our encapsidation assay has several important features. Measuring encapsidation as the ratio of capsid RNA to cytoplasmic poly(A) RNA accounts for variations in the RNA accumulation for different variants. The inclusion of an internal standard allows us to eliminate variations due to transfection efficiency or sample preparation. Additionally, the internal standard in each experiment monitors unanticipated influences of any given variant. For example, if we see an encapsidation defect due to the inability of a variant to make sufficient C protein, then we would expect encapsidation of the internal standard to be affected as well. Likewise, if variants fortuitously express a dominant negative protein that inhibits encapsidation, then we would also expect the internal standard to be encapsidated less efficiently. The consistent levels of encapsidation of the internal standard observed for each experiment increases our confidence that the deficiencies detected in encapsidation of the test RNAs are cis acting.

This work has characterized the cis-acting contributions to hepadnavirus encapsidation. Initially, we mapped a region within the previously described region II from DHBV nt 539 to 610 that makes a significant contribution to encapsidation. Currently, we do not understand how this critical region contributes to encapsidation. It might function similarly to ɛ by forming a structure that interacts with viral proteins to promote encapsidation. We have analyzed potential local RNA secondary structures within this sequence but have thus far been unsuccessful in identifying a structure in the critical element important for encapsidation (data not shown). On the other hand, this critical element could function in encapsidation by interacting with cellular components to promote encapsidation of the pgRNA.

Deletions within DHBV nt 539 to 610 produced a similar encapsidation phenotype as a variant containing a deletion within ɛ, Δ2548-2580 (Fig. 2C). Δ2548-2580 contains the same deletion as a previously described variant, Δ5 (12). In that work, Δ5 produced no detectable encapsidated RNA and was interpreted as an encapsidation-negative variant. For our experiments, variants containing deletions between nt 539 and 610 and Δ2548-2580 encapsidated RNA at a level that was approximately 20 to 30% of the wild-type reference. It is unclear why deletion of these cis-acting encapsidation sequences does not have a larger effect on encapsidation. One possibility is that 20% of the level of the wild-type reference represents the lowest level of encapsidation detected by our assay. Therefore, detecting lower encapsidation levels would be beyond the limits of the assay. Another possibility is that variants containing deletions within ɛ or region II retain residual encapsidation activity. Perhaps a variant containing a deletion in both ɛ and region II would have an even greater defect in encapsidation. On the other hand, low levels of encapsidation may be possible independent of the presence of ɛ or region II.

We found that deletions and substitutions between ɛ and region II had variable effects on encapsidation. Smaller deletions between ɛ and region II had little effect on encapsidation, while complete removal of the intervening sequence affected encapsidation. Restoring the number of nucleotides between ɛ and region II by substituting heterologous sequence did not restore encapsidation. In fact, all substitutions between ɛ and region II significantly affected the ability of the variant pregenomes to be encapsidated. The difference in phenotypes between these two classes of variants is unclear. Possibly, the intervening sequence between ɛ and region II has evolved to not interfere with the encapsidation process. Therefore, removal of the intervening sequence does not affect encapsidation, but substitutions provide sequences that interfere with the function of ɛ and region II in encapsidation. For example, the intervening sequence between ɛ and region II permits the formation of a specific ɛ and/or region II structure or interaction, and substitution of the intervening sequence prohibits the formation of important interactions.

A distantly related avihepadnavirus, HHBV, demonstrated similar features for encapsidation. HHBV also has a region II, but, like DHBV, sequences 5′ of region II contribute to encapsidation. DHBV/HHBV chimeras demonstrated that the HHBV region II element cannot completely substitute for the DHBV region II. One possibility is that an insufficient amount of HHBV region II sequence was replaced in the chimera in order for HHBV region II to function in a DHBV context. Another interpretation is that the HHBV region II is incompatible with other DHBV elements or factors that are required for encapsidation. However, adding back HHBV ɛ to the region II chimera did not restore encapsidation of the chimera. Interestingly, the chimera with only HHBV ɛ was able to encapsidate RNA as efficiently as the DHBV wild-type reference. These results indicated that HHBV ɛ is completely functional for encapsidation with DHBV P and C, despite a 50% decrease in the ability of DHBV P to bind to HHBV ɛ (2). Possibly, adding back additional HHBV sequence or providing HHBV replication proteins could define how region II functions in encapsidation.

It is not clear how region II contributes to the encapsidation of pgRNA. Region II may function in encapsidation in a manner similar to ɛ; that is, region II may bind to a viral protein in order to promote assembly of the viral particle. Conversely, region II may have a role in making the RNA template more suitable for encapsidation. For example, region II may function to direct translating ribosomes off the pgRNA and thus make the RNA accessible to the viral P and C proteins. Experiments done by Howe and Tyrrell (13) indicate that a DHBV sequence containing region II interacts with P and suppresses C translation. This activity could move ribosomes off the pgRNA and render it competent for encapsidation. However, if P does interact with region II, it is possible that interaction has multiple effects, such as inhibiting translation and promoting encapsidation. Why HBV has simpler cis-acting requirements for encapsidation indicates that HBV uses an alternative mechanism to substitute for the role of the region II element in avian hepadnaviruses. Understanding how region II functions in avian hepadnaviruses should provide insight into the mechanism HBV uses to replace the role of a region II element.