De Novo Development of mtDNA Deletion Due to Decreased POLG and SSBP1 Expression in Humans

Defects in the mitochondrial genome (mitochondrial DNA (mtDNA)) are associated with both congenital and acquired disorders in humans. Nuclear-encoded DNA polymerase subunit gamma (POLG) plays an important role in mtDNA replication, and proofreading and mutations in POLG have been linked with increased mtDNA deletions. SSBP1 is also a crucial gene for mtDNA replication. Here, we describe a patient diagnosed with Pearson syndrome with large mtDNA deletions that were not detected in the somatic cells of the mother. Exome sequencing was used to evaluate the nuclear factors associated with the patient and his family, which revealed a paternal POLG mutation (c.868C > T) and a maternal SSBP1 mutation (c.320G > A). The patient showed lower POLG and SSBP1 expression than his healthy brothers and the general population of a similar age. Notably, c.868C in the wild-type allele was highly methylated in the patient compared to the same site in both his healthy brothers. These results suggest that the co- deficient expression of POLG and SSBP1 genes could contribute to the development of mtDNA deletion.


Introduction
Mitochondria are cellular organelles that are surrounded by a double membrane, which is found in almost all eukaryotic cells. They contain a circular genome, mitochondrial DNA (mtDNA), which encodes proteins, rRNAs, and tRNAs necessary for mitochondrial function and exhibits maternal inheritance patterns [1]. Mitochondria play an essential role in energy production by producing adenosine triphosphate (ATP). They have been implicated in many physiological processes, including the production of reactive oxygen species (ROS), pyrimidine and lipid biosynthesis, regulation of cellular substrate levels, apoptosis, metal metabolism, calcium homeostasis and flux, neurotransmitter synthesis, heat production, and insulin secretion [2].
Defects in mtDNA have been linked to multiple human disorders, including multisystem congenital diseases and acquired degenerative disorders [3]. Additionally, they have been related to human aging [4] and are deemed the genetic cause of various human diseases, such as Pearson syndrome (PS), progressive external ophthalmoplegia (PEO), Kearns-Sayre syndrome (KSS), and Leigh syndrome [5,6]. The common mtDNA defect is a specific 4977 bp (mt8468-13446) deletion, which has been identified in various diseases and tissues associated with aging [7,8]. Similar to this deletion, most mtDNA deletions are seen in the major arc between replication origins of heavy and light strands (O H and O L , respectively) [9] and predominantly flanked by short direct mtDNA repeats. Mainly, these deletions tend to develop during the repair of damaged mtDNA rather than replication [9,10].
The genes required to regulate and facilitate mitochondrial function are encoded by both mitochondrial and nuclear DNA, which include 37 mtDNA genes and more than 1000 nuclear genes [11]. Among the nuclear genes, the DNA polymerase subunit gamma (POLG) encodes a catalytic subunit of the mitochondrial DNA polymerase, which plays an important role in mtDNA replication and proofreading [12,13]. Mutations in POLG are associated with the development of mtDNA point mutations and various large mtDNA deletions in different somatic tissues [6,13,14]. Additionally, a previous study using transgenic Polg mutator mice could evidence the link between POLG expression and mtDNA deletion [15].
SSBP1 is also an important protein for mtDNA replication. It must stabilize singlestranded mtDNA and stimulate DNA synthesis by POLG [16]. SSBP1 mutation impairs the amount of mtDNA and its replication [17].
In this study, we described a patient diagnosed with PS who exhibited a large deletion in his mtDNA. Noteworthy, this large deletion was not detected in maternal somatic cells. Furthermore, the association between POLG and SSBP1 expression and de novo mtDNA deletion was investigated, thus gaining further insight into the contributions of the nuclear genome in the development of mtDNA defects.

PCR Assay for Large mtDNA Deletions
DNAs were extracted from multiple tissues of both the PS patient and selected family members. The region surrounding the mtDNA deletion site was amplified by PCR using the following primer sets designed to detect major deletion and wild mtDNA in PS: mt8249F GGGCCCGTATTTACCCTATAG, mt9028R GGCCTGCAGTAATGTTAGCG, and mt13737R GAGAAATCCTGCGAATAGGC. PCR reactions were performed under the following conditions: 1 cycle at 94 • C for 5 min; then 35 cycles at 94 • C for 30 s, 56 • C for 20 s, and 68 • C for 1 min; followed by 1 cycle at 68 • C for 3 min. To detect multiple mtDNA deletions, the following primer sets were used: mt3163F GCCTTCCCCCGTAAATGATA and mt14342R GGGTGGTGGTTGTGGTAAAC. PCR reactions were performed under the following conditions: 1 cycle at 94 • C for 5 min; then 35 cycles at 94 • C for 30 s, 56 • C for 20 s, and 68 • C for 30 s, 1 min 30 s, and 2 min 30 s; followed by 1 cycle at 68 • C for 3 min.

Quantitative PCR for the Detection of mtDNA Deletions
Heteroplasmy of the mtDNA deletions was evaluated using a qPCR assay. DNA was extracted from multiple tissues, and regions of the mtDNA, including deletions (mt8921-13316) in the PS patient, were amplified by PCR using the following primer set: F-8822 CTATAAACCTAGCCATGGCC and R-13436 GGTATGGTTTTGAGTAGTCC. PCR reactions were performed under the following conditions: 1 cycle at 94 • C for 10 min, then 40 cycles at 95 • C for 15 s, and finally 1 cycle at 60 • C for 1 min. Common mtDNA regions were also amplified using PCR with the following primer set: F-16483 GTGAACTGTATCC-GACATCTG and R-100 CAGCGTCTCGCAATGCTATC. The PCR reaction conditions used here were identical to those used for the mtDNA deletion PCR. Heteroplasmy for each deletion was calculated as follows: heteroplasmy of deletion (%) = (quantity of deletion mtDNA/quantity of common mtDNA) × 100.

Skin Fibroblast Isolation and Culture, and Single-Cell Sub-cloning
A skin biopsy sample was taken from the upper arm and dissected into small pieces, incubated in 0.1% collagenase (type IV) for 30 min, and plated into 60 mm dishes containing Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12 (DMEM/F12 medium, Life Technologies, Eugene, OR, USA) supplemented with 10% fetal bovine serum (FBS; Life Technologies) and 4 ng/mL of basic fibroblast growth factor (bFGF; Sigma, Saint Louis, MO, USA). For single-cell sub-cloning, 500 cells were plated in 100 mm dishes, confirmed, and marked for every single cell at a sufficient distance from another cell. These single cells then grew into colonies that were randomly selected and re-plated in six-well culture dishes, cultured, and then subjected to DNA extraction.

Mitochondrial Respiration
Mitochondrial respiration was measured using the XF Cell Mito Stress Test kit in an XF24 Extracellular Flux Analyzer (Seahorse Bioscience, Billerica, MA, USA) according to previously described methods [18]. The oxygen consumption rates (OCRs) for the mitochondria were measured using a serial addition of oligomycin (0.5 µg/mL) for ATP production (Subtraction basal OCR from oligomycin OCR), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 1 µM) for maximal respiration and reserve capacity (Subtraction basal OCR from maximal OCR), and antimycin A (1 µM) and rotenone (0.5 µM) for nonmitochondrial oxygen usage. These values were then normalized against baseline oxygen consumption.

Sanger Sequencing
The c.868 position in POLG and the c.320 position in SSBP1 were amplified using PCR with the following primer set: POLG F-CCTGGAGGTCCCTACTGGTG and R-TGTCCTTCATGGTGCCCTTC and SSBP1 F-GGGCTAGATCCTTCCAAAAG and R-CCGTAAGAGAAATAGCTGAC. PCR reactions were performed under the following conditions: 1 cycle at 94 • C for 5 min; then 35 cycles at 94 • C for 30 s, 56 • C for 20 s, and 68 • C for 30 s; followed by 1 cycle at 68 • C for 3 min. PCR products were purified, sequenced, and analyzed using Sequencher v5.0.

Quantitative RT-PCR (qRT-PCR) for POLG and DNMT Expression
Total RNA was extracted using the RNeasy Mini kit (Qiagen, Hilden, Germany), and complementary DNA (cDNA) was synthesized using the SuperScript TM VILO TM cDNA Synthesis kit (Invitrogen, Carlsbad, California, USA) according to the manufacturer's specifications. qPCR analysis was performed according to previously described methods [20] with minor modification using the following gene-specific primers: POLG F-CTCTTGACCAGGTGCATGTTTG and R-TGCACTGAAAAAGGCGACTG, SSBP1 F-TCAGGACCCTGTCTTGAGA and R-GATATTCTGTGCCATGTTGTC, DNMT1 F-CGAGG ACGAAGATGGAGACG and R-CTGAATGCACTTGGGAGGGT, DNMT3A F-ATCTACGA GGTCCTGCAGGT and R-CACCCACATGTCCGTGTACA, DNMT3B F-TGCTCTTCCTCA GCTGTGTG and R-CTGTCGGCACTGTGGTTTTG, and β-ACTIN F-TGCTATCCCTGTACG CCTCT and R-CTCCTTAATGTCACGCACGA. The expression levels for each gene were measured by qRT-PCR using SYBR Premix Ex Taq (Takara, Kyoto, Japan). The reaction parameters for qRT-PCR were as follows: 1 cycle at 95 • C for 15 min; followed by 35 cycles at 95 • C for 10 s, 60 • C for 15 s, and 72 • C for 20 s; and a final elongation step at 72 • C for 5 min. Gene expression levels were expressed as relative 2 -∆∆Ct values.

Statistical Analysis
The data are presented as the mean ± standard error of the mean (SEM). Independentgroup t-tests for unpairwise comparisons or ANOVA with Tukey's analysis for multiple comparisons were used to evaluate statistical significance (GraphPad Prism, San Diego, CA, USA), and a p-value of <0.05 was considered as significant.

Pearson Syndrome (PS) with a Large mtDNA Deletion
The PS patient was the fifth child of a family with four healthy male children from nonconsanguineous parents ( Figure 1A). At 5 months of age, he presented with sideroblastic anemia, pancytopenia, and bone marrow failure. At 2 years of age, he was identified with proximal renal tubular acidosis (Table 1). When he reached the age of 3, he was affected by chronic kidney disease, exocrine pancreatic insufficiency, adrenal cortical insufficiency, and developmental delay. He was finally diagnosed with PS with mtDNA deletions in peripheral leukocytes. The mtDNA sequence analysis of the PS patient showed that the major deletion was 4395 base pairs (bp) located between ATPase6 (mt8921) and ND5 (mt13316) genes ( Figure 1B,C). However, this deletion has not been previously recorded in the human mitochondrial genome database MitoMAP (https://www.mitomap.org, accessed on 13 February 2021). Nevertheless, in a manner similar to previous cases of deletion phenomena, short direct mtDNA repeats were detected in this novel deletion [10]. The repeats were 5 bp in length (GCACA) and placed at both ends of the deletion (mt8921-8925 and mt13317-13321, Figure 1B). Further investigation on mtDNA deletions in the blood cells of the PS patient and his family members revealed that the major deletion was exhibited only in the PS patient but not in family members ( Figure 1D). For confirmation, multiple tissues from the PS patient were examined, including blood, skin, and both buccal and urine epithelia. As a result, a major deletion was detected in these tissues ( Figure 1E). For investigating maternal transmission of mtDNA deletion, multiple tissues from the mother were also examined. However, no mtDNA deletions were detected in the maternal tissues ( Figure 1E). Even so, there was a possibility that the mother had an extremely low heteroplasmic state or that the deletion had occurred in only a limited number of individual cells [18]. Accordingly, skin fibroblasts from the mother were isolated and evaluated. A total of 10 fibroblast colonies (FC) were selected at random and subjected to single-cell analysis. However, no individual colonies carried the deletion. Although the mother's somatic cell was evaluated due to a lack of sample from the mother's germ cells, these findings could not exclude the possibility that the patient's mtDNA deletion was maternally inherited. The incidence of the major deletion in the PS patient and his family members was quantified within a 0.05% detection limit. Among the family members, only the PS patient harbored the mtDNA deletion and exhibited heteroplasmy between 9% and 70% in various tissues ( Figure 1F). Additional examinations were performed to investigate other mtDNA deletions in the PS patient. We expected that the short direct repeats could be associated with multiple mtDNA deletions. Therefore, the area between mt3163 and mt14342 was explored because there were 15 sites of 5 bp direct repeats out of 17 sites. Multiple mtDNA deletion bands, including major deletions, appeared at different extension times during the PCR cycle ( Figure 1G). Among the multiple bands, the band appearing in 30 s of extension time was selected because of the more abundant amplification than other bands, resulting in mtDNA deletion located between mt3192 and mt14017 ( Figure 1H). This large deletion of 10,825 bp was place-encoded 16S rRNA through ND5 genes; however, this deletion was not flanked by direct repeats, unlike the major deletion ( Figure 1H). in the PS patient were significantly lower than those in other family members, except brother 1.   The oxygen consumption rate (OCR) is an indicator of mitochondrial respiration and energy production [18,23]. Hence, the OCRs of skin fibroblasts isolated from the PS patient and his family were measured and compared. Mainly, the OCR of the PS patient's skin fibroblasts was lower than that of his family members (Figure 2A). The PS patient showed significantly lower basal respiration and ATP production compared to the mother and brother 2 ( Figure 2B). Further, the maximal respiration and the spare respiratory capacity in the PS patient were significantly lower than those in other family members, except brother 1. The significantly lower level of mitochondrial metabolism in the PS patient. Basal respiration and ATP production were significantly reduced in the PS patient compared to the mother and brother 2. The maximal respiration and the spare respiratory capacity in the PS patient were significantly lower than in the father, mother, and brother 2. Each assay was performed in triplicate, and data are represented as the mean ± SEM. a, b, and c indicate significant (p < 0.05) differences among family members.

Identification of Nuclear POLG and SSBP1 Mutations by Exome Sequencing
Since there was no deletion detected in the mtDNA of maternal somatic tissues, the nuclear genome's relation with the major deletion was investigated. Nuclear exome sequencing was performed using the peripheral leukocytes of the PS patient and his parents. In comparison, a total of 2328 variants were detected in the PS patient, 53 of them de novo ( Figure 3A). However, none of the unique variants was recorded to be related to the mitochondria in the Universal Protein Resource (UniProt), based on the functional description available (https://www.uniprot.org). Therefore, the remaining 2275 inherited variants were targeted for further investigation, and 6 of them were selected since they are located in the mitochondria-related genes POLG, SSBP1, MTCH2, ATAD3A, and PAM16, with a The significantly lower level of mitochondrial metabolism in the PS patient. Basal respiration and ATP production were significantly reduced in the PS patient compared to the mother and brother 2. The maximal respiration and the spare respiratory capacity in the PS patient were significantly lower than in the father, mother, and brother 2. Each assay was performed in triplicate, and data are represented as the mean ± SEM. a, b, and c indicate significant (p < 0.05) differences among family members.

Identification of Nuclear POLG and SSBP1 Mutations by Exome Sequencing
Since there was no deletion detected in the mtDNA of maternal somatic tissues, the nuclear genome's relation with the major deletion was investigated. Nuclear exome sequencing was performed using the peripheral leukocytes of the PS patient and his parents. In comparison, a total of 2328 variants were detected in the PS patient, 53 of them de novo ( Figure 3A). However, none of the unique variants was recorded to be related to the mitochondria in the Universal Protein Resource (UniProt), based on the functional description available (https://www.uniprot.org, accessed on 13 February 2021). Therefore, the remaining 2275 inherited variants were targeted for further investigation, and 6 of them were selected since they are located in the mitochondria-related genes POLG, SSBP1, MTCH2, ATAD3A, and PAM16, with a minor allele frequency of <0.03 in the general population ( Figure 3B,C). Among these six variants, c.868C > T (p. [Arg290Cys]) within the POLG gene was predicted to have some pathogenic effect, such as Leigh syndrome, based on the GATK analysis pipeline (https://gatk.broadinstitute.org, accessed on 13 February 2021) and the highest-rank score for pathogenicity prediction using in silico prediction tools ( Figure 3B). been reported as an autosomal dominant mutation associated with optic atrophy disorder and severe and progressive mitochondrial disease manifestations across the full Pearson, Kearns-Sayre, and Leigh syndrome spectrum [17,[25][26][27].
Hence, POLG c.868C>T (p. [Arg290Cys]) and SSBP1 c.320G>A (p. [Arg107Gln]) were evaluated as candidates for a pathogenic factor in the PS patient according to the pathogenicity predictions using exome sequencing analysis and previous reports.   This variant was reported as a homozygous variant in a 2-year-old PS patient [24]. This patient was diagnosed with infantile spasms and showed multiple clinical symptoms, including myoclonic seizures, tonic-clonic seizures, and a global development delay. However, mtDNA analysis was not performed for this patient. Besides, several reports have demonstrated that various POLG mutations are associated with mtDNA-deletion-related diseases, including sensory ataxic neuropathy, dysarthria and ophthalmoparesis (SANDO), and progressive external ophthalmoplegia (PEO) ( Table 2).
The second-highest-rank variant, c.320G > A (p. [Arg107Gln]), in the SSBP1 gene has been reported as an autosomal dominant mutation associated with optic atrophy disorder and severe and progressive mitochondrial disease manifestations across the full Pearson, Kearns-Sayre, and Leigh syndrome spectrum [17,[25][26][27].
[Arg290Cys]) and SSBP1 c.320G>A (p. [Arg107Gln]) were evaluated as candidates for a pathogenic factor in the PS patient according to the pathogenicity predictions using exome sequencing analysis and previous reports.

Deficient Expression of POLG and SSBP1 Induced by Mutations in POLG and SSBP1 Genes
A heterozygous c.868C>T POLG mutation in the PS patient was confirmed using Sanger sequencing. This mutation was inherited from his father ( Figure 4A). Besides, his healthy older brother (brother 1) had the same heterozygous mutation, while the other healthy brother (brother 2) and the mother presented with the wild-type sequence ( Figure 4A). Although the same heterozygous mutation (c.868C>T) was shared by the PS patient and two family members (father and brother 1), the mtDNA deletion was only detected in the PS patient. To understand the difference in this disease's penetrance between the subjects with the POLG mutation, the cDNA converted from their mRNA, including the c.868 position, was sequenced using Sanger sequencing. We noted that the wild-type C residue was expressed at lower levels in the PS patient than that observed in brother 1; however, it was expressed at a similar level to that observed in the father ( Figure 4A).
Then the expression of POLG in mRNA and protein levels were evaluated in all family members. The PS patient and the carrier group demonstrated lower POLG mRNA expression than the wild-type group ( Figure 4B). The highest POLG mRNA expression was observed in brother 2, followed by the mother. In an age-grouped comparison, the PS patient and brother 1 showed significantly lower POLG expression than that observed in brother 2 and the father displayed lower expression levels than those exhibited by the mother (p < 0.05). In addition, in an age-dependent comparison within groups, the father's POLG expression level was significantly lower than that of brother 1 in the carrier group, and the mother's POLG expression level was also lower than that expressed by brother 2. To detect POLG protein expression, two different proteins, β-ACTIN and GAPDH, were used as reference proteins to exclude the possibility of varying reference protein levels caused by aging. β-ACTIN and GAPDH proteins showed a similar expression level in the PS patient and his family members. When POLG protein levels were analyzed using these two reference proteins, POLG protein expression was similar to that of POLG mRNA expression ( Figure 4C). For confirmation of age-dependent POLG gene expression, we evaluated the POLG mRNA levels in the general population (18 individuals, 20-79 years old) and healthy family members of the PS patient and found that POLG expression was reduced by aging ( Figure 4D). The PS patient showed lower POLG expression compared to other people of a similar age. Based on these results, it can be assumed that both the mutation in the POLG gene and aging could be critical factors in decreased POLG expression.
We also analyzed the expression of SSBP1 mRNA in the PS patient and his family members. The PS patient and his parents had already been investigated for their genotype in the c.320 position of the SSBP1 gene in exome sequencing; therefore, we analyzed the same position in brothers 1 and 2 by Sanger sequencing and found that both had no mutation ( Figure 4E). The PS patient and his mother, the carriers, demonstrated significantly lower SSBP1 mRNA expression than the wild-type members, the father, brother 1, and brother 2 ( Figure 4F). The SSPB1 protein level was significantly lower in the PS patient than in all other healthy family members ( Figure 4G). Among the healthy family members, the parents had significantly lower SSBP1 mRNA and protein expression than brothers 1 and 2 ( Figure 4F,G). ers 1 and 2 ( Figure 4F,G). Based on these results, SSBP1 expression was analyzed in the general population, and results showed decreased age-dependent SSBP1 expression like POLG expression ( Figure 4H). The PS patient showed lower SSBP1 expression than other people of a similar age; however, healthy members had comparable SSBP1 expression ( Figure 4H). These results demonstrate that both mutations in the SSBP1 gene and aging could induce reduced SSBP1 expression.
For the PS patient, the expression of POLG and SSBP1 was significantly lower than other people of a similar age group, which could lead to mtDNA deletion and disease.  The PS patient and carriers showed lower POLG mRNA expression than the wild-type group when comparing within similar age of family members (PS patient and brother 1 vs. brother 2; father vs. mother). Further, the father showed lower POLG expression than brother 1 in the carrier group, and the mother displayed lower POLG expression than brother 2 in the wild-type group. The relative POLG mRNA expression was normalized by β-ACTIN mRNA. Each assay was completed in triplicate, and the data are represented as the mean ± SEM. The letters a, b, and c indicate significant (p < 0.05) differences among family members. (C) Deficient POLG protein expression induced by POLG mutation and aging. The PS patient and carriers showed lower POLG protein expression than the wild-type group when comparing within similar age of family members (PS patient and brother 1 vs. brother 2; father vs. mother). The father showed lower POLG expression than brother 1 in the carrier group, and the mother displayed lower POLG expression than brother 2 in the wild-type group. The arrow in a blot of POLG indicates the bands that were quantified for relative expression. The relative POLG protein expression was normalized by β-ACTIN and GAPDH protein. Each assay was completed in triplicate, and the data are represented as the mean ± SEM. The letters a, b, and c indicate significant (p < 0.05) differences among family members. (D) Age-dependent POLG expression. Eighteen individuals in the general population and healthy family members of the PS patient showed that POLG expression reduces by aging. However, the PS patient displayed lower POLG expression compared to others of a similar age. R 2 means Pearson's r correlation. Red and blue dots represent the patient and carriers, respectively. (E) Chromatograph of position c.302 in the SSBP1 gene in brother 1 and brother 2. Both had no mutation. (F) Deficient SSBP1 mRNA expression induced by SSBP1 mutation and aging. The PS patient and his mother, the carriers, demonstrated significantly lower SSBP1 mRNA expression than the wild-type members, the father, brother 1, and brother 2. The relative SSBP1 mRNA expression was normalized by β-ACTIN mRNA. Each assay was completed in triplicate, and the data are represented as the mean ± SEM; * p < 0.05. Among the healthy family members, the parents showed lower SSBP1 expression than brothers 1 and 2. (G) Deficient SSBP1 protein expression induced by SSBP1 mutation and aging. The PS patient showed lower SSBP1 protein expression than the healthy family members. The parents had significantly lower SSBP1 protein expression than brothers 1 and 2. The relative SSBP1 protein expression was normalized by GAPDH protein. Each assay was completed in triplicate, and the data are represented as the mean ± SEM. The letters a, b, and c indicate significant (p < 0.05) differences among family members. (H) Age-dependent SSBP1 expression. SSBP1 expression was reduced by aging in the general population. The PS patient displayed lower SSBP1 expression compared to others of a similar age. R 2 means Pearson's r correlation. Red and blue dots represent the patient and carriers, respectively.
Based on these results, SSBP1 expression was analyzed in the general population, and results showed decreased age-dependent SSBP1 expression like POLG expression ( Figure 4H). The PS patient showed lower SSBP1 expression than other people of a similar age; however, healthy members had comparable SSBP1 expression ( Figure 4H). These results demonstrate that both mutations in the SSBP1 gene and aging could induce reduced SSBP1 expression.
For the PS patient, the expression of POLG and SSBP1 was significantly lower than other people of a similar age group, which could lead to mtDNA deletion and disease.

De Novo mtDNA Deletions with Methylation Changes
We investigated another factor that reduces gene expressions in patients and focused on the degree of methylation in the mutation position in the POLG gene for the penetration of pathogenicity. As children, the PS patient and brother 1 harbored the same mutation; however, the PS patient showed a significantly lower POLG expression although PS patient was younger than brother 1. Several CpG sites were selected, including those at the mutation position (c.868) and several other sites (c.502-613, c.862, c.925, and c.1022) in exon 4 of the POLG gene ( Figure 5A), and analyzed in the PS patient, brother 1, and the father. The methylation status at each site was assessed using the pyrosequencing of the genomic DNA from skin cells. The methylation statuses of c.502-613, c.925, and c.1022 were similar between the subjects (79-89%) ( Figure 5B). In the mutation position, c.868C was changed to c.868T in the carriers' mutant allele. Hence only c.868C could be methylated, and the methylation score predicted by the pyrosequencing reflected that c.868T was not methylated, eventually resulting in a 36% methylation status at position c.868 for the PS patient, while it was 21% and 23% for brother 1 and the father, respectively ( Figure 5B). If the methylation score reflects only the methylation status of the wild-type C residue, then the methylation scores could be assumed to be 72%, 42%, and 46%, respectively ( Figure 5B). The c.862 position showed that 70% was methylated in the PS patient, while the methylation levels were 29% and 39% in brother 1 and the father, respectively ( Figure 5B). The methylation level was measured in wild-type family members, in whom both alleles were C in c.868, resulting in methylation levels of 64% and 72% in brother 2 and the mother, respectively ( Figure 5C). Even though the methylation scores of the mother and brother 1 were similar to or less than that of the PS patient, these wild-type family members showed significantly higher POLG expression, suggesting that the methylation level and POLG expression had a different correlation between wild-type and mutated family members. According to the results, the PS patient displayed hypermethylation in the mutation position compared to brother 1, which suggests that the hypermethylation of the wild-type allele could be a critical factor in the penetrance of the mtDNA deletion only in the PS patient.  brother 1, and the father. c.862 and c.868 were hypermethylated in the PS patient, while the CpG island and other CpG sites were comparable to those of the wild type. Only the wild-type allele at c.868 could be methylated, with the pyrosequencing results confirming no methylation at the mutated allele at c.868. m C in the pyrosequencing peak map represents a methylated cytosine. Each assay was completed in duplicate, and samples were collected from individual cell culture plates. (C) The methylation level of the mutation site in brother 2 and the mother. The assay was performed one time. (D) DNMT expression levels in the mRNA of the PS patient and his family members. Expression of DNMT1 and DNMT3A was higher in the PS patient compared to that in other family members (p < 0.05). Relative DNMT mRNA expression was normalized against β-ACTIN. The data were normalized with healthy control (brother 2). Each assay was repeated in quadruplicate, and the data are represented as the mean ± SEM; * p < 0.05. The letters a, b, and c indicate significant (p < 0.05) differences among family members.
Next, we evaluated whether the expression of methyltransferases, which regulate methylation of the genome [30], was also different in the PS patient compared to the carriers. The mRNA expression levels of DNMT1, DNMT3A, and DNMT3B were measured by qRT-PCR. The expression levels of DNMT1, expected to maintain methylation patterns in newly synthesized DNA during replication, and those of DNMT3A, which functions as a de novo methyltransferase [30], were significantly higher in the PS patient than those in brother 1 ( Figure 5D). These significant differences were also seen in comparison with other family members. DNMT3B levels were similar among all subjects. Taken together, this suggests that hypermethylation in c.868 mutation in the POLG gene acts concurrently to support the penetrance of the large mtDNA deletion in the PS patient.

Discussion
The PS patient harbored multiple mtDNA deletions in several tissues, including the major mtDNA deletion (mt8921-13316), which is a novel mutation that has not been previously reported. This deletion affected mitochondrial function and occurred at various heteroplasmic levels in multiple organs. We hypothesized that the multiple mtDNA deletions in the PS patient could be derived from 5 bp direct repeats in mtDNA deletion based on previous reports that state that mtDNA deletions are often flanked by short direct repeats [10,31]. In the PS patient, a large 10,825 bp deletion (mt3192-14017) was also detected; however, this deletion was not flanked by direct repeats like the major deletion. POLG mutations have been related to diseases that involve multiple mtDNA deletions [6,14,28,29]. Even though the father and brother 1 carried the same POLG mutation without mtDNA deletion, we concluded that the multiple mtDNA deletions in the PS patient could be related to POLG mutation rather than direct repeats. However, the possibility of an association between multiple mtDNA deletions and direct repeats cannot be ignored.
In the PS patient, we could not find any evidence of maternal inheritance for the major deletion; hence, we evaluated the nuclear factors by performing exome sequencing in the PS patient and his parents. As expected, we were able to identify an inherited variant in the POLG gene that could affect mtDNA proofreading. The POLG gene variant c.
[868C>T] (p. [Arg290Cys]) was identified as having the most potent disease risk following in silico evaluations; this mutation was shown to be paternally inherited, and both his father and his elder brother also presented with this variation. However, the mtDNA deletion occurred exclusively in the PS patient.
POLG mRNA and protein expression was evaluated to determine the effect of the c.868 mutation. Among the children, the POLG expression in mRNA and protein levels were significantly lower in the PS patient and the carrier (brother 1) than the wild-type child (brother 2). Further, the father, with the POLG mutation, also exhibited lower POLG protein expression compared to the wild-type mother. Therefore, we confirmed that this POLG mutation reduced POLG expression. Following this, we analyzed POLG expression in the general population. The results showed that POLG expression decreases with age. Healthy family members of the PS patient showed similar phenomena compared to the general population. On the other hand, the PS patient displayed significantly lower expression compared to others of a similar age. Even though the PS patient and the father showed an equivalent level of POLG expression, only the PS patient had a disease phenotype. We supposed that the threshold of the POLG requirement could be different according to age. Cells of juveniles could require higher POLG expression to avoid pathogenicity, while an adult could demand relatively lower POLG expression. Therefore, only the PS patient had the disease, but the father did not under a similar level of POLG expression.
SSBP1 also showed age-dependent decreased expression in the general population. We compared SSBP1 expression in the PS patient and his family to that in the general population. The PS patient showed lower SSBP1 expression than others of a similar age, like POLG expression. Therefore, we concluded that the lower expression of both POLG and SSBP1 genes could contribute to the large de novo mtDNA deletion, leading to the disease in the patient. The PS patient carried mutations in both POLG and SSBP1 genes, while his parents harbored only one of the two mutations without clinical symptoms. A similar digenic case was observed in a PEO patient harboring multiple mtDNA deletions, who carried the co-occurrence of substitution in POLG and C10orf2/Twinkle [32]. The POLG mutation alone in this patient did not cause the clinical phenotype, which was a similar phenomenon to that for one parent of the PS patient.
The PS patient and brother 1 harbored the same mutation; however, only the PS patient showed mtDNA deletion with lower POLG expression than brother 1. There must be additional contributing factors in reducing POLG expression. The methylation of the genome can regulate gene expression [33]; accordingly, we evaluated methylation levels at the mutation sites. As expected, the mutation position was hypermethylated in the PS patient in comparison to brother 1, which could be a critical factor in the induction of lower POLG expression in the PS patient.

Conclusions
Mutations in POLG and SSBP1 genes can negatively affect their mRNA and protein expression ( Figure 6); however, they cannot cause mtDNA deletion and related diseases independently. Co-occurring mutations in both POLG and SSBP1 genes lead to deficient expression, which may contribute to the mtDNA deletion development described in this study. Hypermethylation of the wild-type allele within the same region in the POLG gene might further destabilize mtDNA integrity. The results suggest that POLG and SSBP1 expression might be associated with the penetrance of mtDNA mutations, but not individually. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The exome sequencing data were deposited in GenBank under accession numbers PRJNA555464.