Abstract

Background

   The ATP synthase F1 subunit α (ATP5F1A) gene encodes a critical
   structural subunit of mitochondrial complex V. ATP5F1A mutations are
   linked to mitochondrial complex V deficiency diseases. Although only 14
   cases have been reported globally, the genotype-phenotype correlations
   and underlying molecular mechanisms remain poorly understood.

Objective

   To investigate the pathogenic mechanisms of ATP5F1A deficiency through
   functional analysis of a recurrent missense variant.

Method

   A Han Chinese family with developmental delay and motor dysfunction was
   studied. Whole-exome sequencing and trio analysis identified the
   causative variant. Pathogenicity was evaluated using bioinformatic
   predictions and structural modeling. HEK293T cells were transfected
   with wild-type or mutant-type ATP5F1A plasmids for Western blot and
   immunofluorescence analysis. Morpholino (MO) oligonucleotides were
   microinjected into zebrafish embryos for gene knockdown. Motor neuron
   development was observed in Tg(mnx1:eGFP) zebrafish, with accompanying
   behavioral assessments. RNA sequencing was conducted to explore the
   underlying molecular pathways.

Results

   A de novo missense variant (c.1252G > A, p.Gly418Arg) in ATP5F1A was
   identified and shown to segregate with the disease phenotype. The
   mutation reduced protein stability and expression. In HEK293T cells,
   the mutant protein exhibited reduced expression without affecting
   mitochondrial localization. In zebrafish, atp5fa1 knockdown caused
   growth retardation, motor dysfunction, and impaired motor neuron axon
   development. Rescue experiments with human wild-type ATP5F1A mRNA
   partially restored motor neuron morphology. Transcriptomic analysis
   identified 2,261 differentially expressed genes, enriched in
   neurotransmission and apelin signaling pathways. qPCR confirmed
   downregulation of autophagy-related genes (apln, becn1, map1lc3b) in
   knockdown larvae. Western blot showed that atp5fa1 knockdown increased
   P62 and decreased Lc3b-II expression in zebrafish models.

Conclusion

   This study is the first to report pathogenic ATP5F1A mutations in the
   Chinese population. Atp5fa1 dysfunction leads to multi-system defects
   and disease phenotypes in a zebrafish model, possibly mediated through
   inhibiting autophagy activation mechanisms.

   Keywords: ATP5F1A, Complex v deficiency, Zebrafish, Neurodevelopmental
   deficits

Introduction

   Primary mitochondrial diseases (PMDs) are inherited metabolic disorders
   caused by mutations in mitochondrial DNA (mtDNA) or nuclear DNA (nDNA)
   genes encoding components of the electron transport chain (ETC), with
   an incidence of approximately 1 in 5,000 [[32]1]. PMDs result from
   dysfunction in the mitochondrial oxidative phosphorylation system. As
   mitochondria are present in nearly all tissues and organs, clinical
   manifestations are highly heterogeneous and may affect multiple
   systems, including the nervous system, muscles, and heart [[33]2,
   [34]3]. Notably, childhood-onset PMDs are mainly due to nDNA mutations
   and are often characterized by early onset, severe symptoms, and high
   mortality [[35]4].

   Mitochondrial complex V (ATP synthase) is the final effector of
   oxidative phosphorylation. Structural or functional abnormalities in
   this complex directly impair cellular energy metabolism. Complex V
   catalyzes the final step of oxidative phosphorylation during aerobic
   respiration by transferring protons across the mitochondrial
   electrochemical gradient and synthesizing ATP from ADP, inorganic
   phosphate (Pi), and Mg^2+ [[36]5]. It is encoded by two mtDNA
   genes—ATP6 and ATP8 (encoding subunits α and A6L, respectively)—and 27
   nuclear genes. Its F1 head comprises three α and three β subunits
   arranged alternately to form the catalytic core [[37]6]. During ATP
   synthesis or hydrolysis, the α subunit provides essential arginine
   residues that assist the β subunit in catalysis, playing a key role in
   maintaining enzymatic activity [[38]7]. Mitochondrial complex V
   deficiency, a subtype of PMDs, is typically caused by mutations in
   structural subunits or assembly factors [[39]8]. Common clinical
   features include developmental delay, lactic acidosis, and
   hyperammonemia. Neurological symptoms may include epilepsy, various
   dystonias (including spasms and myoclonus), intellectual disability,
   and motor impairments [[40]9].

   The ATP5F1A gene (OMIM#164360) located on chromosome 18q21.1, encodes
   the α subunit of the catalytic F1 domain of complex V, a core component
   of the mitochondrial ATP synthase complex [[41]10]. In 2013, ATP5F1A
   mutations were first reported to reduce complex V content and impair
   mitochondrial function, leading to fatal mitochondrial encephalopathy
   in infancy [[42]11]. Such mutations can disrupt ATP synthase activity,
   impair ATP production, and compromise cellular energy metabolism,
   resulting in various clinical symptoms [[43]12]. Currently, 14 cases
   linking ATP5F1A mutations to human diseases have been reported, with
   significant phenotypic heterogeneity depending on mutation sites and
   types. Recently, Harrer et al. [[44]13] reported the recurrent
   c.1252G > A variant in three individuals from two families with
   hereditary spastic paraplegia, providing the first clinical
   characterization of this mutation. Comprehensive molecular analysis,
   including fibroblast studies and structural modeling, confirmed its
   pathogenicity. However, the precise molecular mechanisms of ATP5F1A
   deficiency and its role in neurodevelopment remain unclear. In this
   study, we characterize the ATP5F1A c.1252G > A (p.Gly418Arg) variant,
   identified in a Han Chinese child with multi-system symptoms including
   growth retardation, movement disorders, and metabolic abnormalities.
   Through overexpression of the mutant gene and a zebrafish knockdown
   model combined with transcriptomic analysis, we found that ATP5F1A
   c.1252G > A is a loss-of-function (LoF) variant. The atp5fa1 MO
   zebrafish model replicated key phenotypes seen in patients. Disruption
   of the autophagy pathway may be a central mechanism in ATP5F1A-related
   neurodevelopmental disorders. Therefore, this study provides new
   insights into the genotype-phenotype correlations of ATP5F1A-associated
   diseases.

Materials and methods

Patient samples

   Blood samples were collected from the patients and their parents, and
   informed consent was obtained from the parents. The study was approved
   by the Ethical Review Committee of the Affiliated Hospital of Southwest
   Medical University (Ethics approval number: KY2025375).

Molecular analysis

   Whole-exome sequencing of the family was used for the children and
   parents, followed by Sanger sequencing for validation. Briefly, DNA
   samples were extracted from peripheral blood, sequenced using
   high-throughput sequencing, and then filtered and screened for
   variants. Subsequently, primers were designed and Sanger sequencing was
   utilized to confirm the variant sites. Forward primer:
   5′-GCTTGCTAGCGTGCGTATTT-3′; reverse primer: 5′-ACAACACGTAGTACAGGCCG-3′.

In silico analysis

   For better clinical interpretation of variants, allele frequency data
   were obtained from the database gnomAD
   ([45]http://gnomad.broadinstitute.org/). We used the SIFT
   ([46]http://sift.jcvi.org), PolyPhen-2
   ([47]http://genetics.bwh.harvard.edu/pph2) and MutationTaster
   ([48]http://www.mutationtaster.org) prediction tools to assess the
   pathogenicity of the variants and classified the variants according to
   the American College of Medical Genetics and Genomics (ACMG). We
   analyzed the amino acid conservation of the two mutation sites across
   species according to NCBI HomoloGene
   ([49]https://www.ncbi.nlm.nih.gov/homologene) and used the alignment
   function of MEGA ([50]https://www.megasoftware.net/) to compare
   multiple target sequences. Normal and mutant proteins were modeled
   using SwissModel ([51]http://swissmodel.expasy.org/) and visualized
   with the PyMOL package ([52]https://pymol.org/2/).

Cell culture and transfection

   Human embryonic kidney 293T (HEK 293T) cells were cultured in
   Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal
   bovine serum (FBS) and 1% penicillin-streptomycin-amphotericin B at 37℃
   with 5% CO[2]. The wild-type and c.1252G > A mutant full-length
   sequences encoding human ATP5F1A were cloned into the GV657 vector with
   a FLAG tag. For transfection experiments, HEK 293T cells (passages 3–7)
   were seeded in 6-well plates at a density of 3 × 10⁵ cells per well and
   cultured overnight until reaching a density of approximately 70%-80%.
   Transfection was performed using Lipofectamine 2000 (Invitrogen,
   Carlsbad, California, USA) according to the manufacturer’s
   instructions. Briefly, 2.5 µg of plasmid DNA was diluted in 200 µL of
   DMEM, and 5 µL of Lipofectamine 2000 was diluted in 200 µL of DMEM.
   After a 5-minute incubation at room temperature, the diluted DNA was
   mixed with Lipofectamine 2000 and incubated for an additional 20 min to
   form DNA-Lipofectamine complexes. The transfection mixture was then
   added dropwise to the cells cultured in antibiotic-free medium. Cells
   were collected for subsequent analysis 48–72 h post-transfection.

qPCR

   At 72 h post-fertilization (hpf), zebrafish larvae were collected and
   homogenized for total RNA extraction using the TRNzol Universal Reagent
   (TianGen Biotech, Beijing, China), following the manufacturer’s
   instructions. First-strand cDNA was synthesized from 1 µg of total RNA
   using the PrimeScript™ RT Reagent Kit (Takara Bio Inc., Tokyo, Japan).
   Quantitative real-time PCR (RT-qPCR) was conducted using ChamQ
   Universal SYBR qPCR Master Mix (Vazyme Biotech, Nanjing, China) on a
   LightCycler^® 96 System (Roche Diagnostics, Basel, Switzerland). Each
   experimental group consisted of 15 zebrafish larvae, with three
   independent biological replicates and three technical replicates per
   sample. The relative mRNA expression levels were normalized to β-actin,
   which served as an internal reference gene, and were calculated using
   the 2^(-ΔΔCT) method. qPCR primers are shown in Table S1.

Western blotting

   HEK 293T cells transfected with plasmids and zebrafish larvae at 72hpf
   were lysed in RIPA buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1%
   Triton X-100, 1% sodium deoxycholate, 0.1% SDS; Beyotime, Shanghai,
   China) supplemented with protease inhibitor cocktail (Beyotime).
   Following centrifugation at 12,000 × g for 10 min at 4 °C, protein
   concentrations were quantified using an Enhanced BCA Protein Assay Kit
   (NCM, Suzhou, China). Equal amounts of protein (30 µg per lane) were
   resolved on 10% SDS-PAGE and transferred onto PVDF membranes
   (Immobilon-P, 0.22 μm; Merck Millipore). Membranes were blocked for 1 h
   at room temperature in TBS-T (Tris-buffered saline containing 0.1%
   Tween-20) supplemented with 5% non-fat dry milk. Primary antibody
   incubation was performed overnight at 4 °C using the following
   antibodies: rabbit anti-ATP5F1A (1:1000; Abcam, ab176569), mouse
   anti-DYKDDDDK Tag (1:1000; Proteintech, 66008-4), rabbit anti-LC3B
   (1:2000; Abcam, ab192890), rabbit anti-P62 (1:1000; Abcam, ab207305),
   and mouse anti-GAPDH (1:5000; Servicebio, GB15002-100). Following
   primary antibody incubation, membranes were washed three times with
   TBS-T and then incubated for 1 h at room temperature with
   HRP-conjugated secondary antibodies: goat anti-rabbit IgG (1:15000;
   Abcam, ab97051) and goat anti-mouse IgG (1:5000; Abclonal, AS003).
   After extensive washing with TBS-T, protein bands were visualized using
   Ultra-Sensitive ECL Chemiluminescent Substrate Kit (Biosharp, Beijing,
   China) and detected with the Tanon imaging system. Densitometric
   analysis of protein bands was performed using ImageJ software (National
   Institutes of Health, USA).

Immunofluorescence

   HEK 293T cells were grown on coverslips and transfected as described.
   Cells were washed with PBS and fixed with 4% paraformaldehyde in PBS
   for 15 min at room temperature. Following fixation, cells were washed
   three times with PBS and permeabilized with 0.5% Triton X-100 in PBS
   for 15 min. After blocking with 1% bovine serum albumin for 1 h at room
   temperature, cells were incubated with primary antibodies overnight at
   4 °C: anti-IMMT (1:500; Proteintech, 68226-1) and anti-DYKDDDDK tag
   (1:500; Proteintech, 80801-2). Subsequently, cells were washed three
   times and incubated with fluorophore-conjugated secondary antibodies
   for 1 h at room temperature in the dark: goat anti-mouse Multi-rAb
   CoraLite^® Plus 555 (1:500; Proteintech, RGAM003) and goat anti-rabbit
   Alexa Fluor^® 647 (1:500; Abcam, ab150079). Nuclei were counterstained
   with DAPI, and coverslips were mounted using anti-fade mounting medium
   (Beyotime, Shanghai, China). Images were acquired using a confocal
   microscope (Nikon AXR, Japan).

Zebrafish maintenance and microinjection

   Wild-type (AB) and Tg (mnx1: eGFP) transgenic lines of zebrafish were
   raised at the Zebrafish Technology Platform of Southwest Medical
   University. All zebrafish were maintained at 28 °C with a 14 h light
   and 10 h dark cycle. The atp5fa1-e3i4-MO sequence
   (5′-GCCTAGAAGTTTGTCCCTCACCTTC-3′) and the standard control MO sequence
   (5′-CCTCTTACCTCAGTTACAATTTATA-3′) were synthesized by GeneTools
   (Philomat, OR, USA). According to standard procedures, 2 ng/nL was
   injected into the yolk pole of 300 zebrafish embryos at the one-cell
   stage. Microinjected embryos were placed in E3 medium (5 mM NaCl, 0.17
   mM KCl, 0.33 mM CaCl[2], 0.33 mM MgSO[4] and 0.00001% methylene blue)
   and cultured in a 28.5 °C incubator. Zebrafish embryos used for
   fluorescence imaging were treated with E3 medium containing 0.05 mmol/L
   1-phenyl-2-thiourea (PTU) at 24hpf to inhibit skin melanin production.

Behavioral recording

   The zebrafish larvae were tested for their motor behavior using a
   zebrafish behavior analysis system (DanioVision, Noldus, Netherlands).
   A clean 24-well plate was used with 500 uL of culture medium per well
   and one 120hpf larva was tested (three independent experiments, with 8
   larvae per group). Each test of dark field free-swimming was preceded
   by 10 min of adaptation and 30 min of tracking. The alternating light
   and dark test consisted of: (1) a 10-minute acclimatization period, (2)
   a sudden light change test, alternating between 10 min of bright light
   (white LED, ~ 1900 lx) and 10 min of dark (0 lx), totaling 60 min. Data
   were recorded as travel distance (cm) and speed (cm/min). Data were
   analyzed using EthoVision XT software.

Transmission electron microscopy (TEM)

   Zebrafish larvae were pre-fixed with 3% glutaraldehyde and then
   post-fixed with 1% osmium tetroxide. Samples were then dehydrated
   through a graded acetone series (30%, 50%, 70%, 80%, 90%, 95%, and
   100%, with three changes at 100%). Following dehydration, specimens
   were infiltrated with acetone and Epon-812 resin mixtures at ratios of
   3:1, 1:1, and 1:3, then embedded in pure Epon-812 resin. Ultrathin
   Sects. (60–90 nm) were prepared using an ultramicrotome and collected
   on copper grids. Sections were double-stained with uranyl acetate
   (10–15 min) followed by lead citrate (1–2 min) at room temperature. The
   ultrastructural observation and analysis of mitochondria in zebrafish
   larvae neurons were conducted using a JEM-1400FLASH transmission
   electron microscope (JEOL, Tokyo, Japan).

Mitochondrial complex V enzyme activity assay

   The mitochondrial complex V enzyme activity was determined using the
   mitochondrial respiratory chain complex V/ATP synthase activity
   detection kit (Solarbio, Beijing, China). Zebrafish larvae at 72 hpf
   (n = 15 larvae per group) were homogenized in cold isolation buffer,
   and mitochondria were extracted according to the manufacturer’s
   instructions. The assay principle involves complex V hydrolyzing ATP to
   produce ADP and Pi, with the rate of Pi increase measured
   colorimetrically at 340 nm. The calculation of complex V activity was
   normalized by protein concentration.

RNA sequencing analysis

   Transcriptome RNA sequencing (RNA-seq) experiments were performed on
   72hpf larvae in the control and experimental groups. There were 4
   samples in each group, and 15 larvae in each sample. After washing, the
   larvae were snap-frozen in liquid nitrogen and sent to a commercial
   supplier for RNA extraction and high-throughput sequencing. The
   sequencing results were analyzed and plotted on the Majorbio cloud
   platform ([53]https://cloud.majorbio.com/). |log2FC| ≥ 1 and adjusted
   P < 0.05 were the screening criteria. In GO enrichment analysis and
   KEGG PATHWAY enrichment analysis, enrichment was considered significant
   when the adjusted P value (FDR) was < 0.05.

mRNA microinjection

   For the rescue experiment, wild-type human ATP5F1A mRNA was synthesized
   using the mMESSAGE mMACHINE T7 Ultra Kit (Thermo Fisher, AM1345)
   according to the manufacturer’s instructions. Linearized plasmid DNA
   (plasmid construction described in Sect. [54]2.4) was used as a
   template for in vitro transcription, followed by DNase treatment and
   lithium chloride purification. The quantitative analysis was performed
   using NanoDrop. The mRNA was diluted to 200 ng/µL in 0.5× Danieau
   solution containing 0.05% phenol red as a tracer. Morpholino
   oligonucleotides and mRNA were co-injected into one-cell stage embryos.
   After injection, the embryos were transferred to E3 medium and
   incubated at 28.5 °C.

Statistical analysis

   Analyses were performed using GraphPad Prism 9 software. Data normality
   was assessed using Shapiro-Wilk test. Statistical significance was
   calculated using Student t-test or analysis of variance (ANOVA), as
   appropriate. Data are presented as mean ± standard deviation (SD).
   Asterisks indicate the level of statistical significance, * indicates
   P < 0.05, ** indicates P < 0.01, *** indicates P < 0.001 and ****
   indicates P < 0.0001.

Results

Clinical presentation and laboratory findings

   The proband was a 2 years and 4 months oldfemale, the first child of
   non-consanguineous parents (Fig. [55]1A). She was referred to our
   pediatric rehabilitation clinic at 1 year and 10 months old due to
   global developmental delay. Physical examination revealed growth
   retardation, generalized hypotonia, markedly weakened support strength,
   inability to maintain lower limb extension, and impaired standing
   balance. Biochemical analysis demonstrated elevated serum lactate (3.05
   mmol/L; normal: 0.5–2.2 mmol/L) and lactate dehydrogenase (LDH: 268.2
   U/L; normal: 120–250 U/L). Routine blood tests, hepatic/renal function
   panels, and creatine kinase levels were within normal ranges. Metabolic
   screening via liquid chromatography-tandem mass spectrometry (LC-MS/MS)
   showed no significant abnormalities in amino acid or acylcarnitine
   profiles. Brain and full-spine MRI revealed no structural anomalies.
   Video electroencephalography (VEEG) during wake-sleep cycles detected
   intermittent 2.8–3.5 Hz slow-wave bursts localized to bilateral
   fronto-central-parietal regions, with right hemispheric predominance.
   At 2-years-and-4-month-old, motor activity was limited to crawling and
   supported standing, with independent ambulation unachieved. Speech
   development was restricted to monosyllabic utterances. Gesell
   Developmental Schedules evaluation demonstrated profound delays across
   domains: The developmental quotient (DQ) of adaptation, gross motor and
   fine motor was 29.0 (equivalent to 8.4 months); the DQ of language was
   28.0 (equivalent to 8.1 months); and the DQ of personal-social behavior
   was 27.0 (equivalent to 7.5 months). Sensory evaluation revealed severe
   proprioceptive and balance deficits. Respiratory chain enzyme analysis
   was deferred due to parental refusal of muscle biopsy.

Fig. 1.

   [56]Fig. 1
   [57]Open in a new tab

   Identification of a de novo ATP5F1A mutation. A Pedigree chart. Arrow
   indicates the proband (II-1). The solid symbol indicates the affected
   individual (square for males and round for females). B Sanger
   sequencing confirming the de novo heterozygous c.1252G > A (p.
   Gly418Arg) variant in ATP5F1A ([58]NM_001001937.2). (C) Evolutionary
   conservation analysis of glycine-418 demonstrates complete conservation
   at this residue position across vertebrate species. (D) Structural
   simulation of ATP5F1A wild-type and mutant(p. Gly418Arg)protein. Models
   generated using SwissModel and visualized with PyMOL

   Whole exome sequencing identified a heterozygous c.1252G > A (p.
   Gly418Arg) variant in ATP5F1A ([59]NM_001001937.2), located in exon 10.
   Sanger sequencing confirmed this as a de novo mutation, absent in both
   parents (Fig. [60]1B). The variant was unreported in gnomAD population
   databases and submitted to ClinVar (accession: VCV002572074.1;
   [61]https://www.ncbi.nlm.nih.gov/clinvar). Evolutionary conservation
   analysis demonstrated high conservation of glycine at ATP5F1A position
   418 in several species (Fig. [62]1C). Structural modeling revealed
   significant conformational alterations in the mutant protein:
   substitution of non-polar glycine with positively charged arginine at
   residue 418 introduced novel hydrogen bonding with valine-417
   (Fig. [63]1D), potentially disrupting ATP synthase α-subunit folding.
   Bioinformatics prediction software unanimously supported pathogenicity:
   SIFT score = 0, damaging; PolyPhen-2 score = 1, probably damaging;
   MutationTaster score = 0.999, disease-causing. Following ACMG
   guidelines, this variant was classified as pathogenic.

ATP5F1A mutations and human disease phenotypes

   This study summarizes all reported ATP5F1A variants that lead to human
   diseases (Table [64]1). A total of seven mutation sites were
   identified, including six missense mutations and one frameshift
   mutation (Fig. [65]2A). Patients inherit these mutations in both
   autosomal recessive (homozygous) and autosomal dominant (de novo
   heterozygous) patterns. Variants are distributed in both the
   nucleotide-binding domain and the C-terminal catalytic domain of
   ATP5F1A (Fig. [66]2B). Conservation analyses across species showed that
   all missense mutations occurred at highly conserved loci with a ConSurf
   score of at least 7 (Fig. [67]2C).

Table 1.

   Clinical phenotype and genetic characteristics of ATP5F1A related
   diseases patients
   Patient Sex Age at
   Onset Clinical feature Variants Zygosity ETC biochemistry prognosis
   CI CII CIII CIV CV
   P1 M 18d failure-to-thrive, chronic diarrhoea, anaemia, hyperammonemia,
   lactic acidemia c.620 G>A（p.Arg207His） heterozygous NR normal normal
   normal ↓ alive（symptom resolution at12m）
   P2 F 9d failure-to-thrive, hyperammonemia, lactic acidemia c.620
   G>A（p.Arg207His） heterozygous NR NR NR NR ↓ alive（symptom resolution
   at10m）
   P3 F 13d failure-to-thrive, chronic diarrhoea, anaemia, hyperammonemia,
   lactic acidemia, encephalopathy c.620 G>A（p.Arg207His） heterozygous
   normal normal normal normal ↓ alive（symptom resolution at15-18m）
   P4 F Newborn failure to thrive, microcephaly, encephalopathy, IUGR,
   hypotonia, pulmonaryhypertension, heart failure c.926 A>G（p.Tyr321cys）
   homozygous normal normal normal ↓ NR died at 3m
   P5 F Newborn NR c.926 A>G（p.Tyr321cys） homozygous normal normal ↓ ↓ NR
   died at 15m
   P6 NR Newborn intractable seizures, encephalopathy, high-pitched cry,
   horizontal and vertical nystagmus, abnormal primitive reflexes, tonus
   dysregulation c.985 C>T（p.Arg329cys） heterozygous normal normal normal
   normal ↓ died in the first weeks of life
   P7 NR Newborn c.985 C>T（p.Arg329cys） heterozygous normal normal normal
   normal ↓ died in the first weeks of life
   P8 F

   Early

   infancy
   developmental delay, failure-to-thrive, lactic acidosis c.620
   G>A（p.Arg207His） heterozygous normal normal normal normal ↓
   alive（symptom resolution）
   P9 M NR psychomotor delay, intellectual disability, ataxia, spastic
   paraparesis, dystonia c.545 G>A（p.Arg182Gln） heterozygous NR NR NR NR
   NR alive
   P10 F NR psychomotor retardation, spastic tetraparesis, generalized
   dystonia, absent speech, swallowing problems, increased blood lactate
   concentrations c.1037 C>T（p.Ser346Phe） heterozygous NR NR NR NR NR
   alive
   P11 M NR developmental delay, dystonia, cognitive impairment,
   myoclonus, pyramidal syndrome c.1404del (p.Glu469Serfs*3) heterozygous
   NR NR NR NR ↓ alive
   P12 F NR spastic gait impairment, mild intellectual disability
   c.1252G>A (p.Gly418Arg) heterozygous NR NR NR NR NR alive
   P13 F NR spastic paraparesis, mild neurodevelopmental impairments,
   delayed language development, borderline intellectual disability,
   autism spectrum disorder c.1252G>A (p.Gly418Arg) heterozygous NR NR NR
   NR NR alive
   P14 F NR bilateral lower-extremity spasticity c.1252G>A (p.Gly418Arg)
   heterozygous NR NR NR NR NR alive
   [68]Open in a new tab

   Abbreviations: ETC, electron transport chain; CI, complex I; CII,
   complex II; CIII, complex III; CIV, complex IV; CV, complex V; NR, not
   report; IUGR, intrauterine growth restriction

Fig. 2.

   [69]Fig. 2
   [70]Open in a new tab

   Distribution and conservation analysis of ATP5F1A pathogenic variants.
   A) Spatial distribution of 7 variants in the three-dimensional
   structure of human ATP5F1A protein. B Distribution of mutation sites in
   the ATP5F1A protein domains. Domain definitions are based on the
   InterPro database: ATP-synt_ab_N (β-barrel domain, InterPro:
   IPR000568), F1-ATPase_alpha_CD (nucleotide binding domain, InterPro:
   IPR000793), ATP-synt_ab_C (C-terminal catalytic domain, InterPro:
   IPR000194). Generated using IBS 2.0. C Conservation analysis of
   multiple sequences of ATP5F1A protein. The black line indicates the
   ConSurf conservation score (based on UniRef90 alignment, Bayesian
   algorithm), with a score of 1–9 (gradient blue-red, 1 = variable,
   9 = highly conserved). Mutation sites are marked with vertical lines

   ATP5F1A deficiency encompasses a spectrum of disease severities. Based
   on clinical severity and prognosis, the phenotypes can be categorized
   into three groups: self-resolving transient neonatal metabolic
   dysfunction, persistent neurodevelopmental disorders accompanied by
   motor abnormalities, and severe early-onset mitochondrial
   encephalopathy that can lead to fatal consequences. Movement disorders,
   particularly dystonia, are the most common phenotype among these
   patients, followed by neurodevelopmental abnormalities and metabolic
   disturbances, including lactic acidosis and hyperammonemia. Functional
   studies on ATP5F1A mutants demonstrated varying degrees of ATPase
   dysfunction. The p.Arg207His mutant, associated with a transient
   neonatal phenotype, showed reduced enzyme activity [[71]14], while the
   p.Tyr321Cys mutant, associated with lethal encephalopathy, showed
   severe ATPase deficiency and mitochondrial DNA depletion [[72]15].
   Recent investigations confirmed that the p.Glu469Serfs*3 frameshift
   mutation resulted in significantly reduced ATP5F1A mRNA levels and
   compromised ATPase complex assembly [[73]13]. This suggests that loss
   of ATP5F1A function could be a fundamental mechanism underlying
   ATP5F1A-related mitochondrial disorders.

The ATP5F1A c.1252G > A mutation results in reduced mRNA and protein
expression without altering mitochondrial localization

   To investigate the effects of the monoallelic missense mutation of
   ATP5F1A, wild-type (WT) and c.1252G > A mutant-type (MT) plasmid were
   constructed with FLAG tags. qPCR and Western blotting analysis
   demonstrated moderately but significantly reduced mRNA and protein
   expression levels in mutant-transfected cells compared to wild-type
   controls (Fig. [74]3A-C), suggesting that the point mutation leads to a
   decrease in protein expression. To clarify whether the ATP5F1A mutant
   protein changes its subcellular localization, immunofluorescence
   analysis showed that exogenously expressed ATP5F1A co-localized with
   the mitochondrial inner membrane protein IMMT in both wild-type and
   mutant transfected HEK 293T cells (Fig. [75]3D). ATP5F1A is localized
   to the mitochondrial inner membrane, while the p. Gly418Arg mutation
   does not change its mitochondrial localization.

Fig. 3.

   [76]Fig. 3
   [77]Open in a new tab

   Effect of p. Gly418Arg mutation on ATP5F1A mRNA and protein expression
   and subcellular localization. A. Relative expression level of ATP5F1A
   mRNA after HEK 293T cells were transfected with empty vector (NC),
   wild-type (WT) or mutant-type (MT, p. Gly418Arg) vector (qPCR
   detection, internal reference gene: GAPDH; data mean ± SD, n=3;
   **P<0.01, ****P<0.0001, one-way ANOVA). B. Immunoblotting analysis of
   ATP5F1A-FLAG protein (anti-FLAG antibody, internal reference: GAPDH).
   Representative images show the protein expression levels after empty
   vector, WT and MT transfection. C. Quantitative analysis of
   ATP5F1A-FLAG protein expression (ImageJ normalized to GAPDH; mean ± SD,
   n=3; **P<0.01, one-way ANOVA). D. Subcellular localization of
   ATP5F1A-FLAG in HEK 293T cells. Wild-type (WT) and mutant-type (MT)
   proteins (green, anti-FLAG antibody), mitochondrial inner membrane
   marker IMMT (red, anti-IMMT antibody) and cell nuclei (blue, DAPI
   staining). Scale bar, 10μm.

ATP5F1A deficiency causes growth retardation in zebrafish

   To investigate the in vivo functional significance of ATP5F1A, we
   employed zebrafish (Danio rerio) as a model system. The zebrafish
   genome contains an orthologous atp5fa1 sharing 79.2% protein sequence
   homology with human ATP5F1A. Developmental expression profiling by
   RT-qPCR revealed atp5fa1 was highly expressed maternally in the zygote,
   decreased from 0 to 12hpf, and then gradually increased during the
   organogenesis period (12-120hpf) (Fig. [78]4A). The atp5fa1 mRNA and
   protein levels were significantly reduced in the zebrafish model after
   MO injections compared to control-MO injected larvae, confirming
   partial knockdown of atp5fa1. While the protein reduction was modest
   (30%), significant developmental phenotypes were observed (Fig. [79]4B,
   C and D), suggesting that even partial Atp5fa1 reduction can impact
   embryonic development. atp5fa1 knockdown in early embryos causes a
   sharp drop in embryo survival within 24hpf, and almost all embryos die
   at 144hpf (Fig. [80]4E). Morphometric analysis at 72hpf and 120hpf
   revealed a significant reduction in body length, accompanied by
   abnormal phenotypes such as body axis bending and pericardial edema
   (Fig. [81]4F and G).

Fig. 4.

   [82]Fig. 4
   [83]Open in a new tab

   atp5fa1 knockdown causes growth and development disorders in zebrafish.
   A Temporal expression profile of atp5fa1 mRNA during zebrafish
   embryonic development (qPCR detection, internal reference gene:
   β-actin). B The atp5fa1 mRNA expression levels in the control
   (control-MO) and atp5fa1-KD (atp5fa1-MO) larvae at 24hpf, 72hpf and
   120hpf (mean ± SD, n = 15 embryos/group; ****P < 0.0001, Student
   t-test). C Immunoblotting analysis of Atp5fa1 protein (anti-ATP5F1A
   antibody, internal reference: Gapdh). Representative results show the
   difference in protein expression between the control group and the
   knockdown group. D Quantitative analysis of Atp5fa1 protein expression
   (ImageJ software normalized to Gapdh; mean ± SD, n = 3 independent
   experiments; ***P < 0.001, Student t-test). E Survival curves of
   control group and atp5fa1-KD group larvae (0-144hpf; total sample size
   n = 300 embryos/group). F Phenotypic comparison of 72hpf and 120hpf
   larvae. Scale bar, 300 μm. G Quantification of body length of 72hpf and
   120hpf larvae (mean ± SD, n = 10; ****P < 0.0001, Student t-test)

ATP5F1A deficiency causes abnormal neuronal development and motor dysfunction
in zebrafish

   To explore the impact of ATP5F1A deficiency on motor function, the
   mnx1:eGFP transgenic strain of zebrafish was selected for research on
   motor neurons and axons. Comparative analysis revealed that the axon
   length of motor neurons in atp5fa1-knockdown zebrafish was shortened,
   with some axons completely absent (Fig. [84]5A). In rescue experiments
   using human wild-type ATP5F1A mRNA, although the distribution of motor
   neuron axons remained disordered, no shortening or absence of axons was
   observed (Fig. [85]5A). These results indicated that reduced atp5fa1
   expression will lead to impaired development of motor nerves.
   Subsequently, behavioral tests on 120 hpf larvae were conducted
   (Fig. [86]5B). The free-swimming experiment showed that the total
   distance traveled and swimming speed of the atp5fa1-MO group were
   significantly lower than those of the control group, while the rescue
   experimental group partially restored motor function (Fig. [87]5C and
   D). In the 10-min light-dark alternation experiment of zebrafish, the
   juvenile fish in the atp5fa1-MO group showed no obvious response to the
   light stimulation, whereas mRNA rescue treatment partially recovered
   the light-dark responsive swimming behavior (Fig. [88]5E).

Fig. 5.

   [89]Fig. 5
   [90]Open in a new tab

   Effects of Atp5fa1 reduction on zebrafish motor function. A
   Representative images of peripheral neurons in transgenic (mnx1:mGFP)
   zebrafish at 72hpf. Scale bar, 200 μm. B Movement trajectories of
   120hpf zebrafish larvae in the control group (control-MO), atp5fa1
   knockdown group (atp5fa1-MO), atp5fa1-MO and ATP5F1A mRNA co-injection
   group (n = 8 embryos/group). The lines indicate movement trajectories.
   C Quantification analysis of total swimming distance. (Mean ± SD, n = 8
   embryos/group; ****P < 0.0001, Student t-test). D Quantification
   analysis of velocity. (Mean ± SD, n = 8 embryos/group; ****P < 0.0001,
   Student t-test). E Curves showing changes in swimming distance over
   time for zebrafish larvae in the control group, atp5fa1-MO group and
   atp5fa1-MO group + ATP5F1A mRNA during the light-dark alternation
   experiment. Swimming distance represents cumulative distance recorded
   every 5 min.

ATP5F1A deficiency causes impaired mitochondrial structure and function

   ATP5F1A is localized in the inner mitochondrial membrane, and it is
   worth investigating whether its deficiency affects the morphological
   structure and function of mitochondria. Therefore, after knocking down
   atp5fa1 using MO, the ultrastructure of mitochondria in the neurons of
   zebrafish larvae was observed through transmission electron microscopy.
   Compared to the control-MO group, the mitochondria in the atp5fa1-MO
   group showed a shallower matrix, and the mitochondrial cristae were
   shorter or even absent (Fig. [91]6A). At the same time, we explored the
   effect of Atp5fa1 reduction on the activity of mitochondrial complex V.
   The results showed that the activity of complex V in the atp5fa1-MO
   group decreased significantly. Compared to the control-MO group, the
   activity of complex V decreased by approximately 30% (Fig. [92]6B).

Fig. 6.

   [93]Fig. 6
   [94]Open in a new tab

   Effects of Atp5fa1 reduction on mitochondrial structure and function in
   zebrafish larvae. A Ultrastructure of mitochondria in the neurons of
   zebrafish larvae of the control group (control-MO) and atp5fa1 gene
   knockdown group (atp5fa1-MO) under transmission electron microscopy
   (The red framed area is a local magnified view, and the arrow indicates
   the disordered cristae structure or vacuolization; scale bar: 2 μm in
   the main image, 500 nm in the magnified image). B Comparison of the
   activity of mitochondrial complex V in zebrafish between the control
   group and atp5fa1-MO group (Mean ± SD; ***P < 0.001, Student t-test)

Differential expression and enrichment analysis of mRNAs

   To determine the molecular mechanisms underlying the phenotypes of
   developmental disorders and motor dysfunction in zebrafish caused by
   ATP5F1A deficiency, this study performed RNA-seq on the whole embryos
   of control-MO and atp5fa1-MO zebrafish, and conducted a comparative
   analysis of the obtained data. Pearson correlation analysis was carried
   out based on the FPKM (Fragments Per Kilobase of exon model per Million
   mapped fragments) distribution in all samples. The results showed that
   the correlation coefficients of the samples within each group were all
   close to 1, indicating a strong correlation among the biological
   replicate samples, which ensured the reliability of the experimental
   data (Fig. [95]7A). Compared with the control group, a total of 2261
   DEGs were identified in the atp5fa1 knockdown group. Among them, 562
   genes were upregulated and 1699 genes were downregulated. The
   differentially expressed genes were visualized using a volcano plot and
   a bar chart (Fig. [96]7B). GO enrichment analysis showed that the
   differentially expressed genes generated by atp5fa1 knockdown were
   enriched in functions related to neurotransmission, including nervous
   system process, anterograde trans-synaptic signaling, chemical synaptic
   transmission, trans-synaptic signaling, regulation of trans-synaptic
   signaling, postsynaptic membrane, synaptic membrane, neurotransmitter
   receptor activity, etc. (Fig. [97]7C). qPCR verification showed that
   the mRNA levels of neural-related genes such as gabra1, grin1a, gad1b,
   syngr1a, gria3a, grm1a, hcn1 and syt16 were significantly decreased
   (Fig. [98]7E), suggesting that atp5fa1 plays an important role in
   maintaining neurotransmission function.

Fig. 7.

   [99]Fig. 7
   [100]Open in a new tab

   RNA-seq analysis of zebrafish model reveals the molecular regulatory
   network of atp5fa1 knockdown. A Transcriptome correlation analysis
   between samples (Pearson correlation coefficient heat map based on FPKM
   value). B Volcano plot of differentially expressed genes between
   control-MO group and atp5fa1-MO group. C Gene Ontology (GO) enrichment
   analysis of differentially expressed genes. D Kyoto Encyclopedia of
   Genes and Genomes (KEGG) enrichment analysis of differentially
   expressed genes between the control-MO group and the atp5fa1-MO group.
   E qPCR validation of neurotransmission-related genes (gabra1, grin1a,
   gad1b, gria3a, grm1a, hcn1, syngr1a, syt16) (mean ± SD, n = 15
   embryos/group; **P < 0.01, *** P < 0.001, ****P < 0.0001, Student
   t-test). F qPCR validation of Apelin-autophagy relative genes apln,
   becn1, and map1lc3b (mean ± SD, n = 15 embryos/group; **P < 0.01,
   ***P < 0.001, ****P < 0.0001, Student t-test). G) Western blot analysis
   of autophagy-related proteins P62, Lc3b-I, and Lc3b-II in control-MO
   and atp5fa1-MO zebrafish larvae at 72hpf. H Quantitative analysis of
   Lc3b-II protein expression normalized to Gapdh (mean ± SD, n = 3;
   **P < 0.01, Student t-test). I Quantitative analysis of P62 protein
   expression normalized to Gapdh (mean ± SED, n = 3; ***P < 0.001,
   Student t-test)

   After microinjecting Morpholino, 31 KEGG pathways were significantly
   enriched, including Neuroactive ligand-receptor interaction, PPAR
   signaling pathway, Apelin signaling pathway, and Calcium signaling
   pathway, etc. (Fig. [101]7D). Transcriptome sequencing showed that
   compared with the control-MO group, the differentially expressed genes
   in the atp5fa1-MO group were mainly enriched in the Apelin signaling
   pathway. To investigate the effect of Atp5fa1 reduction on the Apelin
   signaling pathway, qPCR analysis revealed that the mRNA expression
   level of apln was decreased in the atp5fa1-MO zebrafish model
   (Fig. [102]7F). Meanwhile, we found that the decrease in apln induced a
   reduction in the mRNA expression levels of becn1 and map1lc3b in the
   atp5fa1-MO zebrafish model (Fig. [103]7F). Western blot analysis was
   performed to detect the expression levels of autophagy-related proteins
   in zebrafish models (Fig. [104]7G). The results showed that the
   expression level of autophagy substrate protein P62 was significantly
   upregulated in the atp5fa1-MO group (Fig. [105]7H), while the
   expression of lipidated LC3-II was significantly downregulated
   (Fig. [106]7I). These changes indicate that the knockdown of atp5fa1
   impairs the initiation of autophagy, while also reducing the efficiency
   of autophagic clearance.

Discussion

   This study provides further evidence that the de novo ATP5F1A
   c.1252G > A (p.Gly418Arg) variant causes mitochondrial complex V
   deficiency, enhancing our understanding of its pathogenic mechanisms.
   The disease phenotype and underlying mechanisms of ATP5F1A deficiency
   were investigated using in vitro cell experiments and Atp5f1a-deficient
   zebrafish models.

   Mutations in the ATP5F1A gene result in disease phenotypes with
   significant clinical heterogeneity. Developmental delay,
   hyperlactatemia, and motor dysfunction were the most common features of
   ATP5F1A deficiency. Recently, Harrer et al. reported three patients
   from two families with hereditary spastic paraplegia (HSP) carrying the
   ATP5F1A c.1252G > A (p.Gly418Arg) variant, and for the first time,
   described the clinical features using protein structural modeling
   [[107]13]. This study independently identified the same variant in a
   Han Chinese patient, further confirming its pathogenicity and expanding
   the ethnic spectrum of affected individuals. Notably, while Harrer et
   al. primarily observed HSP, previously reported patients with ATP5F1A
   mutations—including this case—exhibited multisystem involvement, such
   as growth retardation, movement disorders, and metabolic abnormalities.
   This heterogeneity may arise from three factors: (i) genetic modifiers,
   which significantly influence disease severity through background
   genomic variation; (ii) ethnic-specific differences, reflecting the
   impact of genetic background on phenotypic expression; (iii) epigenetic
   regulation, such as DNA methylation, which may modulate phenotype
   expression [[108]16]. Although the exact mechanisms of ATP5F1A
   modification remain unclear, similar genetic screening models have
   demonstrated that modifier factors can shape mutant phenotypes,
   offering important insights into phenotypic variability.

   Zebrafish is a valuable model organism for studying human genetic
   diseases due to its optical transparency in embryos and larvae, ease of
   genetic manipulation, and high genetic homology with humans [[109]17].
   In recent years, zebrafish models have successfully replicated key
   phenotypes of various mitochondrial complex deficiencies. Knockout of
   the ptpmt1 gene causes growth retardation, craniofacial abnormalities,
   and swim bladder loss, mimicking complex IV deficiency–related diseases
   [[110]18]. The chchd10 gene knockout results in motor coordination loss
   and disrupted neuromuscular synapses, reflecting primary mitochondrial
   disease features [[111]19]. The ndufb7-MO model exhibits ventricular
   dilation and lactic acidosis, indicating metabolic compensation for
   complex I dysfunction [[112]20]. In this study, the atp5fa1-MO
   zebrafish model showed growth retardation and impaired motor function,
   mirroring phenotypes observed in patients with ATP5F1A mutations. This
   model provides an effective tool to explore the molecular mechanisms
   and potential treatments for complex V deficiency. To rule out
   off-target effects of Morpholino technology, we performed rescue
   experiments by injecting human wild-type ATP5F1A mRNA. Partial recovery
   of abnormal motor neuron morphology confirmed a causal link between the
   gene defect and observed phenotypes.

   Mitochondrial function is essential for neural development [[113]21].
   During neurogenesis, metabolism shifts from glycolysis to OXPHOS. In
   mature neurons, energy production relies mainly on OXPHOS to support
   complex neural activity [[114]22, [115]23]. Previous studies using
   mitochondrial gene-deficient models have shown that impaired OXPHOS
   leads to neurodevelopmental disorders (NDDs) [[116]24] and
   neurodegenerative diseases [[117]25]. NDUFB7 deficiency causes complex
   I dysfunction and reduced neuronal volume in the zebrafish midbrain and
   hindbrain [[118]20]. In FARS2-deficient mice, disrupted mitochondrial
   translation impairs OXPHOS, blocking ectoderm development and
   neurogenesis [[119]26]. In this study, atp5fa1 knockdown in zebrafish
   was used to investigate nervous system pathology under mitochondrial
   complex V deficiency. Results showed disorganized mitochondrial cristae
   and reduced complex V activity in neurons, along with abnormal neural
   development and motor impairment. These phenotypes closely match
   neurological symptoms in ATP5F1A-deficient patients. Transcriptome
   analysis revealed that atp5fa1 downregulation altered the expression of
   29 mitochondrial-related genes, including those involved in ATP
   synthase, the electron transport chain, and carrier proteins. These
   findings suggest that mitochondrial dysfunction underlies the observed
   neuronal abnormalities. Therefore, reduced ATP5F1A expression may cause
   neurodevelopmental defects by disrupting mitochondrial structure and
   function.

   Transcriptomic analysis revealed a significant downregulation of the
   Apelin signaling pathway in atp5fa1 knockdown zebrafish models. Apelin
   (APLN), an endogenous ligand of the G protein-coupled receptor APJ, has
   a complex and context-dependent relationship with autophagy in
   neurodegenerative diseases. The precursor is enzymatically cleaved into
   several bioactive forms, including apelin-36, apelin-26, apelin-19,
   apelin-17, apelin-13, Pyr-apelin-13, and apelin-12 [[120]27, [121]28].
   In Alzheimer’s disease models, Apelin-13 inhibits autophagy by
   activating the PI3K/Akt/mTOR pathway, reducing Beclin-1 and LC3B-II
   expression and exerting neuroprotective effects [[122]29, [123]30].
   Conversely, in Parkinson’s disease models, Apelin activates autophagy
   via the AMPK/mTOR/ULK pathway to clear pathological protein aggregates
   [[124]31]. qPCR and Western blot analyses in this study showed that
   atp5fa1 knockdown significantly reduced apln and becn1 mRNA levels,
   decreased Lc3b-II protein expression, and increased P62 accumulation.
   Therefore, this coordinated pattern suggests that atp5fa1 knockdown may
   disrupt Apelin signaling and impair autophagy progression.

   Based on the observed mitochondrial dysfunction and autophagy
   dysregulation, we propose a comprehensive therapeutic strategy for
   ATP5F1A-related diseases. Modulating autophagy is a promising approach,
   and enhancers like rapamycin warrant systematic evaluation.
   Additionally, mitochondrial bioenergetics enhancers like coenzyme Q10,
   creatine, and nicotinamide riboside may help compensate for ATP
   deficiency [[125]32, [126]33, [127]34]. Drug development should
   prioritize patient-derived iPSC neuronal models for personalized
   screening [[128]35]. Furthermore, integrating CRISPR-Cas9 gene editing
   offers opportunities for mechanistic studies and potential curative
   therapies by correcting pathogenic mutations [[129]36].

   This study has several limitations. Functional analysis was conducted
   in HEK 293T cells, which do not fully replicate affected tissue
   environments; patient-derived cells would provide more relevant models.
   While MO-mediated knockdown in zebrafish effectively simulates the
   heterozygous phenotype, it carries risks of off-target effects and
   transient action. Validation using stable CRISPR/Cas9 knockout lines
   would strengthen these findings. Single-cell multi-omics approaches are
   essential to unravel molecular complexity, as integrating
   transcriptomics, proteomics, and metabolomics can identify cell
   type-specific vulnerabilities and novel therapeutic targets within
   dysregulated networks [[130]37] thereby facilitating translation from
   mechanistic insight to precision therapy.

   In summary, this study preliminarily elucidates the impact of ATP5F1A
   deficiency on clinical phenotype, biological function, and molecular
   mechanisms through clinical data, genetic analysis, and experiments
   using cell and animal models. It confirms p.Gly418Arg as a recurrent
   pathogenic ATP5F1A mutation and establishes a link between reduced
   ATP5F1A expression and neurological phenotypes. Our findings suggest
   that the apelin–autophagy signaling axis may serve as a therapeutic
   target in ATP5F1A-related disorders, offering new mechanistic insights
   for targeted treatment development.

Acknowledgements