Discovery of novel disease-causing mutation in SSBP1 and its correction using adenine base editor to improve mitochondrial function

Abstract

Mutations in nuclear genes regulating mitochondrial DNA (mtDNA) replication are associated with mtDNA depletion syndromes. Using whole-genome sequencing, we identified a heterozygous mutation (c.272G>A:p.Arg91Gln) in single-stranded DNA-binding protein 1 (SSBP1), a crucial protein involved in mtDNA replisome. The proband manifested symptoms including sensorineural deafness, congenital cataract, optic atrophy, macular dystrophy, and myopathy. This mutation impeded multimer formation and DNA-binding affinity, leading to reduced efficiency of mtDNA replication, altered mitochondria dynamics, and compromised mitochondrial function. To correct this mutation, we tested two adenine base editor (ABE) variants on patient-derived fibroblasts. One variant, NG-Cas9-based ABE8e (NG-ABE8e), showed higher editing efficacy (≤30%) and enhanced mitochondrial replication and function, despite off-target editing frequencies; however, risks from bystander editing were limited due to silent mutations and off-target sites in non-translated regions. The other variant, NG-Cas9-based ABE8eWQ (NG-ABE8eWQ), had a safer therapeutic profile with very few off-target effects, but this came at the cost of lower editing efficacy (≤10% editing). Despite this, NG-ABE8eWQ-edited cells still restored replication and improved mtDNA copy number, which in turn recovery of compromised mitochondrial function. Taken together, base editing-based gene therapies may be a promising treatment for mitochondrial diseases, including those associated with SSBP1 mutations.

Keywords: MT: RNA/ DNA editing; NG-ABE8e; NG-ABE8eWQ; NG-Cas9-based ABE8e; NG-Cas9-based ABE8eWQ; SSBP1; editing efficacy; mitochondrial diseases; myopathy; off-target effects; optic atrophy; sensorineural deafness; single-stranded binding protein 1.

Graphical abstract

Figure 1

Characterization of de novo mutation in SSBP1 (A) The pedigree of the Korean family (SNUH547) harboring a de novo heterozygous SSBP1 mutation (c.272G>A:p.Arg91Gln). Filled symbols and opened symbols indicate affected and unaffected individuals, respectively. (B) The auditory steady-state response for the proband (SNUH547-1105) exhibits bilateral, symmetric, moderately severe sensorineural deafness. Meanwhile, pure-tone audiometry for both parents shows normal hearing. (C) Sanger sequencing chromatograms of the SSBP1 c.272G>A:p.Arg91Gln mutation in the SNUH547 family. The arrow indicates the site of the mutation. (D) The sequence domain (top) and conservation maps (bottom) showcase SSBP1 mutations previously reported in the literature (blue circles), with the inclusion of the novel mutation (c.272G>A:p.Arg91Gln) from this study (red circle). In the conservation map, gray regions denote untranslated regions, while green regions signify the coding sequence. All the mutations’ residues, including Arg91 residue, were highly conserved among the SSBP1 orthologs in various species (highlighted in yellow).

Figure 2

Overexpression of mutant SSBP1 induces mtDNA depletion and mitochondrial dysfunction in A549 cells through the inhibition of mtDNA replication (A) Quantification of mtDNA was performed using genomic DNA isolated from A549 cells overexpressing SSBP1. Relative copy numbers of mtDNA genes ND1 and ND5 were measured using quantitative real-time PCR. SLCO2B1 and SERPINA DNA levels were used to normalize the results; ∗∗p < 0.01, ∗∗∗p < 0.001 (mean ± SEM, n = 3 independent experiments). (B) Ribbon diagram showing intrachain interaction of SSBP1 Arg91 with Ala134 (PDB: 3ULL). The Ala134 residue forms a hydrogen bond with the Arg91 residue; however, the interaction with the Ala134 residue was lost with the Arg91Gln mutant. (C and D) Images of Coomassie Brilliant Blue staining (left) and SSBP1 immunoblotting for SSBP1 recombinant proteins after oligomerization. Each immunoblot band intensity was quantified using ImageJ, ∗∗∗p < 0.001 (means ± SEM, n = 3 independent experiments, unpaired Student’s t test). (E) Ribbon diagram of SSBP1 tetramer structure with aligned ssDNA (PDB: 1EYG). The negatively charged ssDNA interacts with the positively charged Arg91 residue; however, the Arg91Gln residue is predicted not to interact. (F) Electrophoretic mobility shift assay analysis of SSBP1 was conducted using a ssDNA probe at a concentration of 250 nM, which was incubated with 1 μM of both wild-type and mutant SSBP1 tetramers. (G and H) Representative confocal images of A549 cells with SSBP1 overexpression with EdU incorporation (green). SSBP1 was stained using Anti-Flag (blue), while mitochondria were stained with MitoTracker (red) (also see Figure S7) and nuclei were stained with DAPI (white). The number of EdU foci per cell was counted manually. ∗∗∗p < 0.001 (mean ± SEM, n = 14–18, unpaired Student’s t test). Scale bar, 10 μm. (I) 7S DNA was quantified in A549 cells overexpressing SSBP1 using quantitative real-time PCR. The results were normalized to mtDNA levels; ∗∗∗p < 0.001 (means ± SEM, n = 3 independent experiments, unpaired Student’s t test). (J) Representative images of dynamics of the mitochondrial network in A549, HEI-OC1, and HeLa cells stained with MitoTracker (red), Flag (green), and DAPI (blue). Images processed with Fiji software, visualized in white. Scale bar, 10 μm. (K) OCR in HeLa cells overexpressing SSBP1 wild-type and mutant under basal conditions and after injection of oligomycin (O), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP; F), rotenone (R) and antimycin A (AA). (L) Basal respiration, maximal respiration, and ATP production in HeLa cells were calculated from OCR traces and are reported in the graph; ∗p < 0.05 (mean ± SEM, n = 5, unpaired Student’s t test).

Figure 3

Analysis of sgRNA and ABE variant efficacy for SSBP1 mutation and their off-target effects (A) Positions and editing efficiencies of sgRNA candidates within the SSBP1 gene. The red colored A indicates the target nucleotide (c.272G>A) and blue colored A indicates bystander As. A-to-G conversion rate for each nucleotide is shown using heatmaps. Underlined base identifies the PAM of each sgRNA. The editing frequency displayed on heatmap is an average value between three experimental replicates. (B) A-to-G or C-to-other conversion or indel rate after using NG-sg1. The position of base is defined as counting the end distal to the PAM as position 1. (C) Potential sgRNA-dependent off-target sites investigated by Cas-OFFinder software. (D) sgRNA-dependent off-target editing outcomes of ABE-treated patient-derived fibroblast cells. The dashed line indicates 50% frequency because of the heterozygosity of on-target.

Figure 4

Functional recovery of mitochondria in patient-derived fibroblasts using ABEs (A) Genomic DNA was isolated from the fibroblast cell lines and subjected to quantitative real-time PCR for ND1 and ND5. SLCO2B1 or SERPINA was used to normalize the results; ∗∗∗p < 0.001 (mean ± SEM, n = 3 independent experiments). (B) Representative confocal images showed control, patient-derived, and genome-edited fibroblasts stained with MitoTracker (red) and anti-DNA (green). The boxes in the merged images point to the enlarged sections displayed at the bottom of each panel. The relative amounts of DNA that co-localized with MitoTracker were quantified; ∗p < 0.05, ∗∗∗p < 0.001 (means ± SEM, n = 5–17, one-way ANOVA followed by the Bonferroni post hoc test). Scale bar, 10 μm. (C) Same as (B), except that immunofluorescence with EdU incorporation. The co-localization of MitoTracker (red) and EdU (green) foci was employed to confirm the replication of mtDNA. The EdU incorporation into the mitochondria of different fibroblast cell lines was calculated and statistically analyzed; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 (mean ± SEM, n = 12–18, one-way ANOVA followed by the Bonferroni post hoc test). Scale bar, 10 μm. (D) Quantification of 7S DNA within the genomic DNA from fibroblast cell lines by quantitative real-time PCR. The mtDNA level was used to normalize the result (mean ± SEM, n = 3 independent experiments, one-way ANOVA followed by the Bonferroni post hoc test). (E) Representative confocal images depict mitochondrial fission in fibroblast cell lines. Mitochondrial network formation was assessed through MitoTracker staining (red). The boxes in the merged images indicate the enlarged sections shown at the bottom of each panel. Scale bar, 10 μm. (F) Transmission electron microscopy images of mitochondria with normal (blue arrow), large vacuole (yellow arrow), and abnormal (red) morphology in fibroblast cell lines. The relative proportion of normal mitochondria within the total mitochondria population was calculated; ∗∗∗p < 0.001 (mean ± SEM, n = 7–29, one-way ANOVA followed by the Bonferroni post hoc test). (G) OCR in fibroblasts under basal conditions and after injection of oligomycin (O), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP; F), rotenone (R), and antimycin A (AA). (H) Alterations in basal respiration, maximal respiration, and ATP production in fibroblasts were derived from OCR traces; ∗∗p < 0.01, ∗∗∗p < 0.001 (mean ± SEM, n = 3–6 one-way ANOVA followed by the Bonferroni post hoc test).

Figure 5

Characterizations of SSBP1 mutations in patients and predictions of base editing efficacy (A) Type and distribution of the 114 single nucleotide variations (SNVs) reported in the literature for SSBP1. Among them, 89.5% display G-to-A single base transitions, whereas 8.7% manifest A-to-G transitions, indicating the dominance of single base transition mutations. (B) Assessment the number of patients and the corresponding SSBP1 rescue score. A total of 8 G-to-A mutations associated with mtDNA depletion syndrome for potential base editing applications. Predictions were carried out using the DeepBE and BE-Hive software. The SSBP1 rescue score (or frequency) indicates that the mutation is corrected without any other amino acid mutations.