Structure and nucleic acid interactions of the SΔ60 domain of the hepatitis delta virus small antigen

Abstract

Infection with hepatitis delta virus (HDV) causes the most severe form of viral hepatitis, affecting more than 15 million people worldwide. HDV is a small RNA satellite virus of the hepatitis B virus (HBV) that relies on the HBV envelope for viral particle assembly. The only specific HDV component is the ribonucleoprotein (RNP), which consists of viral RNA (vRNA) associated with the small (S) and large (L) delta antigens (HDAg). While the structure of the HDAg N-terminal assembly domain is known, here we address the structure of the remaining S^Δ60 protein using NMR. We show that S^Δ60 contains two intrinsically disordered regions separated by a helix-loop-helix motif and that this structure is conserved in the full-length protein. Solution NMR analysis revealed that S^Δ60 binds to both full-length and truncated vRNA, highlighting the role of the helical regions in submicromolar affinity interactions. The resulting complex contains approximately 120 S^Δ60 proteins per RNA. Our results provide a model for the arginine-rich domains in RNP assembly and RNA interactions. In addition, we show that a cluster of acidic residues within the structured region of S^Δ60 is critical for HDV replication, possibly mimicking the nucleosome acidic patch involved in the recruitment of chromatin remodelers. Our work thus provides the molecular basis for understanding the role of the C-terminal RNA-binding domain of S-HDAg in HDV infection.

Keywords: NMR; S-HDAg; acidic cluster; hepatitis delta virus; protein–RNA interaction.

Fig. 1.

Sequences and alignments of the S-HDAg from eight different human genotypes retrieved from GenBank, highlighting the assembly domain, the short linear motif (SLiM)-like domain, the nuclear localization signal (NLS) and the two arginine-rich domains (ARMs). Posttranslational modifications are indicated by circled letters (P, phosphorylation; M, methylation; A, acetylation). The numbers 1 to 8 correspond to the following GenBank accession numbers of the strain sequences used: AJ000558.1, BAD02974.1, AAB02595.1, AAF22831.1, CAJ66095.1, CAJ66096.1, AXF50952.1, and CAJ66094.1, respectively. Amino acid numbering is according to genotype 1. Genotype 1 is the genotype used in this study.

Fig. 2.

3D structure of the central HLH motif of S^Δ60. (A) Ensemble of the final 20 structures of S^HLH (residues 89 to 148) shown in light pink. The N- and C-terminal regions are fully disordered and are not shown here. (B) Lowest-energy structure of S^HLH showing a predominantly α-helical structure, with residues G118-S123 forming a short turn. The ARM1 (E98-K106) and ARM2 (E138-A145) regions are shown in green. (C) Side view (90° rotation) of the S^HLH structure, showing an angle of approximately 55° between the two helices. (D) Structure with side chains of all Arg residues represented as sticks. (E) Close-up of the hydrophobic core of the Leu residues involved in the leucine zipper, shown with green side chains. (F) Close-up of the side chains of charged residue stabilizing the hairpin via salt bridges involving R104 and K112 with the negatively charged residues E129 and E126, respectively. The distances between the NH groups of R104 and K112 are given with the closest oxygen atoms of the glutamates. (G) Close-up of other charged residue side chains stabilizing the hairpin involving R140 and E98 as well as D101, R105, and D137. Figure generated using ChimeraX 1.4 (34).

Fig. 3.

Comparison of S-HDAg and S^Δ60 NMR spectra. Overlay of 2D solid-state ¹H-¹⁵N INEPT NMR spectrum of S-HDAg (gray) and solution-state ¹H-¹⁵N band-Selective Optimized Flip-Angle Short-Transient Heteronuclear Multiple Quantum Coherence NMR spectrum of S^Δ60 (blue). A schematic representation of both constructs is shown on Top, with helices represented as spheres. Resonances from S^Δ60 corresponding to the HLH region are surrounded by green circles.

Fig. 4.

Characterization of HDV RNA and binding to S^Δ60. (A) AFM image of full-length RNA^1.7kb. The vertical scale of the Right Inset is −1.0 to 1.5 nm. (B) Histogram and smoothed length distribution of HDV RNA molecular skeletons extracted from eight AFM images as described in SI Appendix, Fig. S3 and calculated for 130 molecules. The average RNA length is 217 ± 5 nm with a SD of 61 nm. (C) RNA^1.7kb visualized by TEM with positive staining, using darkfield imaging mode: the electron micrograph on the Left corresponds to a higher magnification of the area shown on the Right. (D) Overlay of nine 2D HN-SOFAST spectra of S^Δ60 with increasing amounts of RNA^1.7kb added to the protein solution, ranging from 40,000:1 protein:RNA molar ratio (light blue) to 80:1 (dark blue). The control spectrum is shown in pale blue. Full spectra are shown in SI Appendix, Fig. S5. (E) Relative intensities of NMR signals from S^Δ60 residues for different amounts of vRNA added to the protein.

Fig. 5.

Binding profile of S^Δ60 and S^HLH upon interaction with nucleic acids. (A) Binding parameter obtained for S^Δ60 (in dark blue) and for S^HLH (in dark red) with RNA^1.7kb derived from the fit of the intensity decrease. (B) Binding parameter obtained with RNA^1.7kb plotted on the S^Δ60 structure with a color gradient from white (weak binding) to dark red (high binding), and with the unassigned/overlapping residues colored in gray. (C) Binding parameter values obtained for S^Δ60 (in light blue) and for S^HLH (in light red) with RNA^394nt. (D) Binding parameter obtained with RNA^394nt plotted on the S^Δ60 structure with a color gradient from white (weak binding) to dark red (high binding). The Right panels of (B and C) show a close-up view with the side chains of residues showing binding above the average calculated in the helices.

Fig. 6.

Models for the interaction of S^HLH with RNA. (A) Matching of the S^HLH motif (wheat) on the HIV Rev ARM peptide (cyan) in complex with its stem-loop IIB RNA (transparent gray) (PDB 4PMI) using ChimeraX. (B) Close-up of the matched motif and the conserved amino acids (dark blue and green for S^HLH and cyan for Rev). (C) Predicted structural model obtained by AlphaFold3 (40) of S^HLH in complex with the hairpin portion of the HDV RNA, from nucleotide 26 in 3′ to 1,656 in 5′. (D) Restraint-driven docking model obtained by HADDOCK of S^HLH with the predicted model of the HDV RNA fragment shown in panel C. The best cluster is shown, with active residues indicated by red sticks.

Fig. 7.

Identification and role of the S-HDAg acidic cluster. (A) Sequence alignment of H2A histone and S^Δ60, highlighting the three residues involved in the nucleosome acidic patch of H2A and aligning with S^Δ60. (B) Structure of the nucleosome core particle H2A (cyan)-H2B (blue), H3-H4, and 146 bp DNA fragment (PDB ID: 1KX5). A close-up view displays the side chain of the acidic patch residues of H2A and H2B. The electrostatic potential is projected onto the molecular surface of one chain of H2A and H2B to highlight the acidic patch (49). (C) Structure of S^HLH in surface representation, with the electrostatic potential showing the acidic cluster. A close-up view of the side chain of the ¹²⁵EEEEE¹²⁹ stretch is shown. (D) PHHs (Left) and HepG2-NTCP cells (Right) infected with WT HDV or recombinant HDV mutants. Viral particles were inoculated into PHHs and HepG2-NTCP cells at an m.o.i. of 10. Cells were collected 5 d postinfection, total RNA was extracted and intracellular HDV RNA levels were quantified by qPCR and expressed as HDV genomes per 2 µg of total RNA. Error bars in both PHH cells and HepG2-NTCP indicate the SD of three independent experiments. **P < 0.005; ***P < 0.0005; (one-way ANOVA test). Mock = mock infected; WT = wild-type HDV; S/A/A+S = HDV mutant carrying in S-HDAg the R75A mutation in the SLiM (S); the E98A, E127A, and E129A mutations in the acidic cluster (A), and all four mutations combined (A+S).