Deep mutational scanning of hepatitis B virus reveals a mechanism for cis-preferential reverse transcription

Abstract

Hepatitis B virus (HBV) is a small double-stranded DNA virus that chronically infects 296 million people. Over half of its compact genome encodes proteins in two overlapping reading frames, and during evolution, multiple selective pressures can act on shared nucleotides. This study combines an RNA-based HBV cell culture system with deep mutational scanning (DMS) to uncouple cis- and trans-acting sequence requirements in the HBV genome. The results support a leaky ribosome scanning model for polymerase translation, provide a fitness map of the HBV polymerase at single-nucleotide resolution, and identify conserved prolines adjacent to the HBV polymerase termination codon that stall ribosomes. Further experiments indicated that stalled ribosomes tether the nascent polymerase to its template RNA, ensuring cis-preferential RNA packaging and reverse transcription of the HBV genome.

Keywords: cis preference; deep mutational scanning; diproline; hepatitis B virus; reverse transcription; ribosome stalling.

Figure 1.. HBV pgRNA transfection can uncouple selection from overlapping reading frames.

(A) Schematic of the HBV genome with colored boxes representing ORFs and arrows indicating Pol II transcriptional start sites. (B) HBV DNA copy number measured by qPCR two days after transfecting cells with WT or mutant pgRNAs. Core- (T33*), HBs- (C69*), HBx- (G27*), Pol-(YMHH active site mutant). Data plotted are mean, with each replicate plotted as a dot. Lower limit of quantification (LLOQ). (C) Schematic to illustrate trans (Core) and cis (Pol) activity when multiple pgRNAs are present in a single cell. Left, pgRNA encoding defective Core protein can be encapsidated in trans by functional Core from another pgRNA and reverse transcribed to produce HBV DNA. Right, Pol works in cis, therefore, pgRNAs encoding defective Pol protein are not reverse transcribed by functional Pol encoded by another pgRNA. (D) WT and mutant pgRNAs encoding Pol⁻ or Core⁻ were co-transfected at the ratios indicated on the x-axis. The percentage of WT HBV DNA from total DNA recovered, as determined by amplicon sequencing, is plotted on the y-axis. The horizontal dotted line indicates the expected percentage of WT HBV DNA, assuming an absolute cis preference. A diagonal dotted line indicates the expected percentage of WT HBV DNA, assuming no cis preference. Each symbol represents the mean of three replicates. Error bars are ±SEM. See also Table S1.

Figure 2.. Pol translation model and deep mutational scanning the Core region.

(A) A Pol translation model that involves ribosome re-initiation is shown in the context of pgRNA and its open reading frames (ORFs). First, a ribosome translates the C0 ORF, enabling it to bypass the Core start codon. The ribosome then re-initiates scanning downstream of C0, bypasses the C1 methionine codon due to poor Kozak context, and initiates translation at the J ORF, which allows the ribosome to bypass the C2 methionine codon. Lastly, the ribosome re-initiates scanning after translating the J ORF and initiates translation at the Pol start codon. (B) DMS heatmap of Core, codons 2 – 186. The 64 codons are depicted on the y-axis, with start and termination (stop) codons highlighted. Site positions are portrayed on the x-axis. The color of heatmap squares correlates to the log₂ enrichment factor of a given variant. WT codons are symbolized with black dots. White squares are filtered-out variants one G-to-T or C-to-A transversion away from the WT codon and are likely affected by oxidative damage during sequencing library preparation. The size of heatmap squares is reduced if read counts for the plasmid library reference were low. (C) In-frame start and stop variants in Core are separated into groups for statistical comparison with WT. Data are mean with 95% confidence intervals. (D) Variants that improve or reduce the Kozak sequence one codon upstream of the C1 or the J ORF start codon are grouped for statistical analysis. Variants that mutate or preserve the J ORF start or stop codon are grouped as well. Data are mean with 95% confidence intervals. **, p<0.01. ***, p<0.001. ****, p<0.0001 by Mann-Whitney U test. An interactive version of the heatmap in panel B is at https://hbv-dms.github.io/R1/Fig/Fig_2B.html. See also Figures S1 and S2.

Figure 3.. Polymerase deep mutational scanning.

(A) Heatmap depicting the fitness of polymerase variants is plotted as in Fig. 2B. Pol and its overlapping reading frames are shown as colored boxes (top). Catalytic site amino acids (triangles) and regions of interest (lines) are indicated below the heatmap. (B) AlphaFold 2 structural predictions of HBV gtA Pol. TP, RT, and RNase H domains are indicated by color. The Spacer domain is unstructured and is not shown. N and C termini are labeled. Catalytic site amino acids are circled and labeled. The region of interaction between the RT extension and insertion is labeled. (C) Close-up view of the predicted interaction between the RT extension and insertion. RT extension and insertion are indicated by color. Amino acids involved in two putative zinc fingers are colored yellow and labeled. The possible locations of zinc ions are shown as gray circles. (D) Normalized log₂ selective pressure. The absolute value of the log₂ fold change in frequency of all sense codons at each aa position was added to calculate the cumulative selective pressure. All values were normalized to the highest and lowest value in the data set. Each aa position is represented by a dot. Specific aa were subsetted and plotted as indicated. RT extension (aa 325-354) and insertion (aa 456-508), excluding aa in the putative zinc finger motifs. Spacer (aa 179-324). Active sites (Y65, D553, D554, D702, E731, D750, D790). Remaining (all aa excluding those plotted in other categories). Mean and 95% confidence intervals are shown. An interactive version of the heatmap in panel A is at https://hbv-dms.github.io/R1/Fig/Fig_3A.html. See also Figures S3 and S4, Table S3, and Data S1.

Figure 4.. HBV Pol C terminus encodes a ribosome stalling motif.

(A) AlphaFold 2 structure prediction of the HBV RNase H domain. The classical RNase H fold is labeled, and the unstructured tail (aa 810-845) is colored pink and labeled. N and C termini are labeled. (B) Enlarged heatmap of the unstructured tail (aa 810-845) is plotted as in Fig. 2B. The predicted locations of aa in the ribosome exit tunnel are indicated below the heatmap. Amino acids with strong selective pressure are colored and labeled. L4 and L17; ribosomal proteins. Direct repeat 2 (DR2) is labeled. (C) WT and mutant HBV RNase H constructs (top), produced for the indicated periods in vitro in the presence of ³⁵S-methionine, were separated by SDS-PAGE and analyzed by autoradiography. A solid triangle indicates the expected size of the full-length product. An open triangle indicates tRNA-bound intermediates. (D) WT and mutant HBV RNase H constructs, produced as in panel C for the indicated periods of time, were treated with RNase OUT or RNase A for 15 min to protect or degrade covalently bound tRNA, respectively. (E) WT and mutant HBV RNase H translation products produced as in panel C (15 min) were separated by centrifugation through a 1M sucrose cushion, as depicted on the top. Pellets containing high-density protein/nucleic acids were resuspended and then treated with RNase OUT or RNase A for 15 min to protect or degrade covalently bound tRNA, respectively. The ³⁵S-methionine labeled proteins were analyzed by SDS-PAGE, as shown on the bottom. (F) Northern blot to detect RNase I protected fragments from in vitro translation assays (15 min) using WT or Mut (PP->AA) templates as in panel C. ³²P-labeled probes that hybridized either near the start codon (left) or the termination codon (right) were used to detect protected RNA fragments. The samples loaded in each of the three lanes are on the right of the blots. The approximate size of monosome- and disome-protected RNA fragments are indicated on the right. See also Figure S5, Tables S1, S3, and Data S1.

Figure 5.. C-terminal prolines contribute to HBV polymerase’s cis preference.

(A) Schematic of HBV Pol mutations. (B) The effect of mutations on HBV Pol’s ability to package and reverse transcribe pgRNA.; qPCR quantification of HBV DNA copies per well of a 6-well plate. Replicates are plotted, and column height indicates the mean. Error bars are SEM. LLOQ, lower limit of quantification. (C) The pgRNAs indicated were co-transfected at equal ratios (50:50) and the percentage of YMHH mutant HBV DNA recovered is plotted. The dotted line indicates the percentage of YMHH pgRNA transfected. Each replicate is plotted, and column height indicates the mean. Error bars are SEM. See also Figure S6 and Table S1.

Figure 6.. Ribosome stalling leads to cis-preferential reverse transcription of HBV pgRNA.

A model to explain the molecular basis of HBV Pol’s cis preference. The model posits that the proline codons encoding the C terminus of Pol stalls ribosomes, thereby tethering Pol to the pgRNA template to facilitate cis-preferential binding of the Pol protein to the packaging signal (epsilon), ultimately resulting in cis-preferential pgRNA reverse transcription.