Structural basis for nuclear import of adeno-associated virus serotype 6 capsid protein

Abstract

Adeno-associated viruses (AAVs) are the most extensively researched viral vectors for gene therapy globally. The AAV viral protein 1 (VP1) N-terminus controls the capsid's ability to translocate into the cell nucleus; however, the exact mechanism of this process is largely unknown. In this study, we sought to elucidate the precise interactions between AAV serotype 6 (AAV6), a promising vector for immune disorders, and host transport receptors responsible for vector nuclear localization. Focusing on the positively charged basic areas within the N-terminus of AAV6 VP1, we identified a 53-amino acid region that interacts with nuclear import receptors. We measured the binding affinities between this region and various nuclear import receptors, discovering a notably strong interaction with IMPα5 and IMPα7 in the low nanomolar range. We also elucidated the X-ray crystal structure of this region in complex with an importin alpha (IMPα) isoform, uncovering its binding as a bipartite nuclear localization signal (NLS). Furthermore, we show that using this bipartite NLS, AAV6 VP1 capsid protein can localize to the nucleus of mammalian cells in a manner dependent on the IMPα/IMPβ nuclear import pathway. This study provides detailed insights into the interaction between the AAV6 VP1 capsid protein and nuclear import receptors, deepening our knowledge of AAV nuclear import mechanisms and establishing a basis for the improvement of AAV6-based gene therapy vectors.IMPORTANCEAAVs, recognized as the most extensively researched viral vectors for gene therapy globally, offer significant advantages over alternatives due to their small size, non-pathogenic nature, and innate ability for tissue-specific targeting. AAVs are required to localize to the nucleus to perform their role as a gene therapy vector; however, the precise mechanisms that facilitate this process remain unknown. Despite sharing overt genomic similarities with AAV1 and AAV2, AAV6 is a unique serotype. It is currently recognized for its ability to effectively transduce hematopoietic cell lineages and, consequently, is considered promising for the treatment of immune disorders. Identifying the exact mechanisms that permit AAV6 to access the nucleus can open up new avenues for gene therapy vector engineering, which can ultimately lead to increased therapeutic benefits.

Keywords: adeno-associated virus; importin; karyopherin; nuclear import.

Fig 1

AAV6 VP1-BR binds to importin alpha isoforms but not importin beta 1. (A) AAV6 VP1-BR FITC-tagged peptide sequence used for EMSA and FP. The peptide spans residues 120–172 and contains an N-terminal FITC tag and Ahx linker. (B) EMSA showing AAV6 VP1-BR with ΔIBB-IMPα isoforms spanning members from each of the three subfamilies (SF1, IMPα1/IMPα2; SF2, IMPα3; SF3, and IMPα5/7) and IMPβ1. The FITC peptide is shown false colored in red (middle panel). Proteins were stained using Coomassie blue stain (top panel, blue). The overlay is represented at the bottom panel, where FITC peptide (red) overlays with protein (blue) to indicate co-migration (binding) of AAV6 VP1-BR with all IMPα proteins but not IMPβ1. EMSA results are representative of three independent experiments. (C) FP assay measuring the direct binding between AAV6 VP1-BR FITC peptide and respective importin isoforms. Binding was observed with IMPα1 (K_D = 557 nM), IMPα2 (K_D = 265 nM), IMPα3 (K_D = 198 nM), IMPα5 (K_D = 71 nM), and IMPα7 (K_D = 23 nM). Binding with IMPβ1 was so low that an accurate K_D could not be determined (N/D). Error bars were calculated using the standard error of the mean of three independent experiments. The error for the K_D values was the standard error of the mean. (D) Statistical results of a one-way ANOVA test using Tukey’s post-test for multiple comparisons of the FP K_D results. All P values of comparison between IMPα isoforms are displayed as follows: ns as P > 0.05; *P ≤ 0.05, **P ≤ 0.01, ****P ≤ 0.0001. ns, non-significant.

Fig 2

Crystal structure reveals AAV6 VP1-BR as a bipartite NLS in complex with IMPα2. (A) Schematic overview of the AAV6 VP1 protein and structure of the AAV2 VP1-BR (purple surface) and IMPα2 (gray surface) complex resolved to 2.3-Å resolution. The sequence of AAV6 VP1-BR bound to IMPα2 is detailed in the box (BRs determined by sequence alignment are underlined, with residues seen within the structure in bold purple, and residues not able to be resolved in the structure are denoted in gray). This structure has been deposited to the PDB and givPen the code 9CFT. (B) Major and minor binding sites of IMPα2 (gray surface) with labeled AAV6 VP1-BR residues (purple sticks) occupying sites and P positions indicated. (C) Simplified representation of IMPα2 and AAV6 VP1-BR binding interactions. The AAV6 VP1-BR (purple line) residues forming bonds with IMPα2 residues (gray box) are indicated through complementary arrows. Salt bridges are indicated via underlined IMPα2 residues, and non-underlined residues indicate hydrogen bonds. binding interactions were determined by the PDBsum server. P4 position Leu¹⁷¹ is not depicted as it does not form any hydrogen bonds or salt bridge interactions with IMPα2.

Fig 3

AAV6 VP1 nuclear localization is mediated by the IMPα/IMPβ heterodimer. (A) HEK 293A cells were seeded on glass coverslips and transfected to express the indicated GFP fusion proteins in the presence or absence of mcherry-Bimax2. Sequences include mutated residues in red. (B) Twenty-four hours later, cells were treated with DRAQ5 to stain cell nuclei, fixed, and mounted on microscope slides to allow quantitative CLSM analysis. Representative images of the indicated GFP fusion proteins expressed in the absence (left panels, no add) or presence (right panels, +mcherry-Bimax2) of mcherry-Bimax2. Images of the indicated channels are shown, along with a merged image. (C) Micrographs such as those shown in panel B were quantitatively analyzed to calculate Fn/c relative to each GFP fusion protein at the single-cell level. Individual measurements are shown, along with the mean (black horizontal bars) and standard deviation of the mean (gray vertical bars) from three independent experiments. Results from the Welch and Brown-Forsythe ANOVA test of significance are shown. ***P < 0.0005, ****P < 0.0001. (D) The percentage of cells relative to each indicated GFP fusion protein, displaying the indicated subcellular localization, is shown. N, nuclear, Fn/c  ≥ 10; N  >  C, nuclear more than cytosolic, 2  ≤  Fn/c  <  10; U, ubiquitous, 1  ≤  Fn/c  <  2; C  >  N, more cytosolic than nuclear, Fn/c  <  1. ns, non-significant.