YME1L1 Dysfunction Associated With 3‐Methylglutaconic Aciduria

ABSTRACT

3‐methylglutaconic aciduria (3‐MGCA) is a biochemical finding in a diverse group of inherited metabolic disorders. Conditions manifesting 3‐MGCA are classified into two major categories, primary and secondary. Primary 3‐MGCAs involve two inherited enzymatic deficiencies affecting leucine catabolism, whereas secondary 3‐MGCAs comprise a larger heterogeneous group of conditions that have in common compromised mitochondrial energy metabolism. Here, we report 3‐MGCA in two siblings presenting with sensorineural hearing loss and neurological abnormalities associated with a novel, homozygous missense variant (c.1999C>G, p.Leu667Val) in the YME1L1 gene which encodes a mitochondrial ATP‐dependent metalloprotease. We show that the identified variant results in compromised YME1L1 function, as evidenced by abnormal proteolytic processing of substrate proteins, such as OPA1 and PRELID1. Consistent with the aberrant processing of the mitochondrial fusion protein OPA1, we demonstrate enhanced mitochondrial fission and fragmentation of the mitochondrial network in patient‐derived fibroblasts. Furthermore, our results indicate that YME1L1^L667V is associated with attenuated activity of rate‐limiting Krebs cycle enzymes and reduced mitochondrial respiration, which may explain the build‐up of 3‐methylglutaconic and 3‐methylglutaric acid due to the diversion of acetyl‐CoA, not efficiently processed in the Krebs cycle, towards the formation of 3‐methylglutaconyl‐CoA, the precursor of these metabolites. In summary, our findings classify YME1L1 deficiency as a new type of secondary 3‐MGCA, thus expanding the genetic landscape and facilitating the diagnosis of inherited metabolic disorders featuring this biochemical phenotype.

YME1L1 dysfunction is a new type of secondary 3‐methylglutaconic aciduria, observed in two siblings with a novel homozygous YME1L1 variant.

1. Introduction

3‐methylglutaconic aciduria (3‐MGCA) is a characteristic biochemical feature of a growing number of clinically and genetically heterogeneous inherited metabolic disorders. It is defined by abnormally elevated urinary excretion of 3‐methylglutaconic (3‐MGC) and 3‐methylglutaric (3‐MG) acid, which in most cases reflects compromised mitochondrial function of varying degree and severity [1, 2]. Based on the underlying biochemical defect, inborn errors of metabolism associated with increased 3‐MGC acid excretion are classified into primary and secondary 3‐MGCAs [3].

Primary 3‐MGCAs include autosomal recessive disorders caused by defects in the leucine catabolism pathway due to mutations in either the AUH (OMIM #250950) or HMGCL gene (OMIM #246450), which encode 3‐methylglutaconyl‐CoA hydratase (AUH) and 3‐hydroxy‐3‐methylglutaryl‐CoA (HMG‐CoA) lyase, respectively. The metabolic block caused by deficiencies in either of the above enzymes results in the build‐up of intermediates of the leucine catabolism pathway, including 3‐methylglutaconyl‐CoA (3‐MGC‐CoA), which is considered the precursor of 3‐MGC and 3‐MG acid excreted in urine [4, 5, 6].

Secondary 3‐MGCAs constitute a more heterogeneous group of disorders, which are not caused due to impaired leucine catabolism, but result from gene defects affecting proteins with diverse function. Unlike primary 3‐MGCAs, in secondary 3‐MGCAs, the metabolic etiology of increased 3‐MGC and 3‐MG acid excretion remains unclear. Over the past three decades, several hypotheses have been proposed to explain the biochemical basis of secondary 3‐MGCA, as it is found in several, ostensibly different inherited metabolic disorders all of which have in common compromised mitochondrial function. The prevailing notion suggests that secondary 3‐MGCA arises from defects that either directly or indirectly affect mitochondrial oxidative energy metabolism, particularly oxidative phosphorylation [1, 2, 4]. Compromised aerobic respiration in non‐hepatic tissues prevents NADH and FADH₂ produced by oxidative metabolism from being efficiently oxidized, leading to their accumulation in the mitochondrial matrix [4, 5]. This build‐up diminishes the activity of key Krebs cycle enzymes, resulting in increased levels of acetyl‐CoA within mitochondria, part of which is shunted into the “acetyl‐CoA diversion pathway” towards the formation of 3‐MGC‐CoA, which leads to the production of 3‐MGC and 3‐MG acid [4, 5].

To date, defects in more than 20 genes have been linked to inborn errors of metabolism characterized by secondary 3‐MGCA as a discriminative feature [1, 7]. Notably, in almost all conditions featuring secondary 3‐MGCA, the affected gene encodes a protein with an established role in mitochondrial function [1, 7, 8]. Therefore, although not specific to a particular disease, 3‐MGCA is widely regarded as a biochemical finding associated with mitochondrial dysfunction.

Here, we report a novel homozygous variant in the YME1L1 gene identified in two siblings presenting with sensorineural hearing loss and neurological findings. We provide evidence that the identified variant leads to functional impairment of YME1L1 and is linked with 3‐MGCA. This study is the second report of patients with a mitochondriopathy associated with YME1L1 dysfunction. As opposed to the herein reported siblings, the previously described YME1L1 patients displayed a more severe, infantile‐onset phenotype, with prominent neurological features including global developmental delay, muscle weakness, leukoencephalopathy, and optic nerve atrophy [9]. Our study establishes for the first time an association between YME1L1 dysfunction and 3‐MGCA, thus expanding the spectrum of inherited metabolic diseases linked to this biochemical phenotype.

2. Materials and Methods

2.1. Qualitative and Quantitative Urine Organic Acid Analysis

Urine organic acid analysis was performed by gas chromatography–mass spectrometry (GC–MS). Qualitative urine organic acid analysis was performed on an Agilent 6890 N/5973 system, essentially as described elsewhere [10]. Briefly, following oxidation with hydroxylamine hydrochloride, organic acids were extracted from urine samples using liquid–liquid (ethylacetate and ether) extraction, converted to their corresponding trimethylsilyl esters with BSTFA + 1% TMCS (Sigma‐Aldrich, 15 238) and subjected to chromatographic separation. The identity of organic acids was determined by comparing the generated mass spectra to a custom library. Quantitative urine organic acid analysis was performed on an Agilent 7890/5875C GC–MS system. Extraction was carried out with ethylacetate, while oxidation and derivatization with BSTFA/TMCS were performed as described above. Quantification of organic acids was achieved using linear calibration curves generated for each analyte by calculating the peak area ratios (analyte/internal standard). Concentrations were normalized to creatinine and expressed as mmol/mol creatinine.

2.2. Quantification of Blood Lactate Levels

Lactate concentration in blood was determined spectrophotometrically using the lactate reagent set (Instruchemie, 2864), according to the manufacturer's instructions.

2.3. Genetic Analysis

Genomic DNA was extracted from peripheral blood samples using the QIAmp DNA Blood Midi Kit (Qiagen, 51 185), and subjected to whole exome sequencing. DNA libraries were prepared using the TruSeq DNA exome library preparation kit (Illumina, 20 020 615), according to the manufacturer's instructions. Paired‐end sequencing of the pooled libraries was performed on a NextSeq 500 system (Illumina) using the High Output Kit v2.5 (300 Cycles), (Illumina, 20 024 908), according to the manufacturer's guidelines. Demultiplexing and adapter trimming was performed automatically using BaseSpace Sequencing Hub Apps (Illumina). Bioinformatics processing, analysis, annotation, and interpretation was performed by the VarSome Clinical platform (Version 11.4) using the human reference genome build hg19. The presence of pathogenic or likely pathogenic variants or variants of uncertain significance in known genes associated with 3‐MGCA or mitochondrial disorders was assessed by in silico panel analysis [1, 7]. The identified variant was confirmed by conventional polymerase chain reaction (PCR) followed by automated sequencing of exon 18 of the YME1L1 gene. PCR amplification was carried out using the Amplitaq Gold DNA polymerase (Applied Biosystems, N808‐0240) and the following primers: Forward: 5′‐TTGTACTTCTTTCTTGTCCCCCA‐3′, Reverse: 5′‐CAGAGGGCCTTCAGCTGTTAT‐3′. Amplification products were purified using the ExoSAP‐IT PCR Product Cleanup Reagent (Applied Biosystems, 78 200) and bidirectional Sanger sequencing was performed using the BigDye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, 4 337 449) and the above‐mentioned primers. The sequencing reactions were subjected to clean‐up using the Performa DTR Gel Filtration Cartridges (EdgeBio, 42 453) and capillary electrophoresis was performed on an Applied Biosystems 3500xL Genetic Analyzer. The generated data were compared to the reference YME1L1 transcript, as listed in the GenBank database (NM_014263.4).

2.4. Human Skin Fibroblasts

Control skin fibroblasts derived from apparently healthy individuals were purchased from the Coriell Institute for Medical Research (GM00969, GM08333,) and Lonza (CC‐2509) and are designated as Control A, Control B, and Control C, respectively. Primary human fibroblasts were established from skin punch biopsy derived from Patient 1, following collagenase digestion (Sigma‐Aldrich, C5138) at 37°C for 45 min and centrifugation at 580 × g for 7 min to pellet cells. Dermal fibroblasts were cultured in Amnioprime medium (Capricorn Scientific, APR‐B) supplemented with 10% active fetal bovine serum (FBS, Gibco, ThermoFischer Scientific, 10 270 106) and 2 mM L‐Glutamine (Biosera, XC‐T1715). Patient‐derived and control fibroblasts were maintained at 37°C in a humidified incubator with 5% CO_2, in Minimum Essential Medium (MEM, Gibco, ThermoFischer Scientific, 10 370 021) supplemented with 15% active FBS, 2 mM L‐Glutamine, and 1x Penicillin/Streptomycin (Biosera, XC‐A4122), unless otherwise stated.

2.5. Immunofluorescence Staining and Microscopy

Cells grown on glass coverslips were fixed with 4% paraformaldehyde (PFA) in Phosphate Buffered Saline (PBS) for 10 min at room temperature (RT), followed by permeabilization with 0.1% Triton X‐100 in PBS for 10 min at RT. Cells were then blocked with 5% normal goat serum (Biosera, GO‐605) in PBS containing 0.05% Tween 20 (PBST) for 1 h at RT and incubated with primary antibodies diluted in 5% goat serum in PBST overnight (O/N) at 4°C. The next day, cells were incubated with the appropriate secondary fluorescently labeled antibodies for 1 h at RT. Antibodies used for immunofluorescence staining are listed in Table S1. Nuclei were counterstained with DAPI. Images were acquired with a Leica TCS‐SP2 DMIRE2 confocal microscope, using a 63× objective lens and 1.5× digital zoom.

2.6. Analysis of Mitochondrial Morphology

Quantification of parameters related to the morphology of the mitochondrial network in control and patient‐derived fibroblasts was performed in ImageJ using the Mitochondria Analyzer plugin, as previously described [11]. For this, image deconvolution was first conducted on confocal z‐stacks of TOMM20 fluorescently labeled mitochondria with the Iterative Deconvolution 3D plugin using a theoretical point spread function (PSF) generated with the Diffraction‐PSF‐3D plugin in ImageJ [12].

2.7. Immunoblot Analysis

Cell pellets were incubated in ice‐cold lysis buffer (150 mM NaCl, 1% Triton X‐100, 0.1% SDS, 50 mM Tris–HCl pH 8.0) supplemented with 1× protease inhibitor cocktail (Sigma‐Aldrich, S8820) for 30 min on ice. Following brief sonication, cell lysates were centrifuged at 13000 × g for 15 min at 4°C, and supernatants were collected. Protein concentration was determined using the Bicinchoninic Acid (BCA) assay [13]. Protein samples were mixed with an equal volume of 2× SDS sample buffer (100 mM Tris–HCl pH 6.8, 4% SDS, 20% glycerol, 0.01% bromophenol blue, 200 mM DTT), incubated at 95°C or 37°C (for OXPHOS antibody cocktail) for 10 min, resolved by SDS‐PAGE, and transferred onto PVDF membranes (BIO‐RAD, 1620177). Blocking was performed with 5% non‐fat milk or 5% BSA (for DRP1) in TBST (10 mM Tris–HCl pH 8.0, 150 mM NaCl, 0.05% Tween‐20) for 3 h at RT. Membranes were sequentially incubated with primary and HRP‐conjugated secondary antibodies diluted in blocking buffer O/N at 4°C and 1 h at RT, respectively. Proteins were visualized using a chemiluminescent substrate reagent (BIO‐RAD, 170–5061) and images were obtained with the FUSION Solo X (Vilber) chemiluminescence imaging system. GAPDH served as a loading control. Densitometry was performed using ImageJ, and protein levels were normalized against GAPDH. Values are expressed as arbitrary units (AU) and represent mean ± SEM of three independent experiments. Antibodies used for immunoblotting are listed in Table S1.

2.8. RNA Isolation, Reverse Transcription, and Quantitative PCR

Total RNA was isolated from fibroblasts using RNAzol RT (Sigma‐Aldrich, R4533) and cDNA synthesis was performed with the PrimeScript RT reagent Kit (TaKaRa Bio, RR037A), according to the manufacturer's instructions. Quantitative PCR was performed on an Applied Biosystems QuantStudio 7 Flex real‐time PCR system using the SYBR Green Universal Master Mix (Applied Biosystems, 4 309 155). Values were normalized against the housekeeping gene TATA‐box binding protein (TBP). Data were analyzed using LinRegPCR version 2020.0 [14]. The following primers were used: PRELID1 Forward: 5′‐CTCCAGAGCTGTCCAGGAAT‐3′, Reverse: 5′‐ACAAGTGTTTTGGAAGGGGC‐3′; TBP Forward: 5′‐TGCACAGGAGCCAAGAGTGAA‐3′, Reverse: 5′‐CACATCACAGCTCCCCACCA‐3′.

2.9. Expression Constructs

The expression vector of human YME1L1 in frame with a C‐terminal myc‐FLAG epitope was purchased from Origene (RC201088). The YME1L1 c.1999C>G variant was introduced by means of site‐directed mutagenesis from wild‐type YME1L1‐myc‐FLAG using the Q5 site‐directed mutagenesis kit (NEB, E0554S) and the following primers: Forward: 5′‐AATAAGAATCGTTCTAAGGGACTC‐3′, Reverse: 5′‐TCTTGTTCGATGGCAG‐3′, according to the manufacturer's instructions.

2.10. Culture and Transfection of HeLa Cells

HeLa cells were cultured in Dulbecco's Modified Eagle Medium (DMEM, Biosera, LM‐D1109) high glucose supplemented with 10% heat inactivated FBS (Gibco, ThermoFischer Scientific, 10 500 064), 2 mM L‐Glutamine, and 1 Penicillin/Streptomycin at 37°C under 5% CO₂. Transient transfections were performed using Lipofectamine 3000 Transfection Reagent (Invitrogen, ThermoFischer Scientific, L3000008), according to the instructions of the manufacturer.

2.11. Assessment of the Activity of Krebs Cycle Enzymes

Control and patient‐derived fibroblasts were cultured in DMEM (PAN Biotech, P04–04510) high glucose supplemented with 10% heat inactivated FBS (iFBS, GE Healthcare Hyclone, SV30160.03) and 1× Antibiotic‐Antimycotic solution (Biosera, XC‐A4110) at 37°C in a humidified 5% CO₂ atmosphere. Cell pellets were resuspended in ice‐cold PBS and the activities of citrate synthase (CS), succinate dehydrogenase (SQR), fumarate hydratase (FH), and α‐ketoglutarate dehydrogenase (α‐KGDH) were measured essentially as described elsewhere [15, 16, 17]. For CS, FH, and α‐KGDH, cell suspension was solubilized in 1.5% lauryl maltoside (N‐dodecyl‐β‐D‐maltoside), centrifuged, and the supernatant was used for the enzyme assays. CS activity was determined using acetyl‐CoA and oxaloacetate as substrates and the chromogenic compound 5,5′‐dithiobis(2‐nitrobenzoic acid) (DTNB) by monitoring the increase in absorbance at 412 nm. FH activity was determined using L‐malic acid by following the increase in absorbance at 250 nm. For α‐KGDH, the reduction of NAD⁺ (0.5 mM) in the presence of α‐ketoglutaric acid was monitored at 340 nm. For the SQR assay, cell suspension in PBS was employed and enzyme activity was calculated based on the reduction of 2,6‐dichlorophenolindophenol (DCPIP) monitored by a decrease in absorbance at 600 nm. All assays were performed in triplicates using a temperature‐controlled spectrophotometer (Shimadzu UV2600, Japan) at 37°C. Activities were expressed as nmol of substrate converted per minute and normalized to the protein content in the reaction (nmol.min⁻¹.mg prot⁻¹). Total protein amount was determined by the Lowry protein assay [18]. All chemicals were purchased from Sigma‐Aldrich.

2.12. Seahorse Analysis of Oxygen Consumption Rate

Oxygen consumption rate (OCR) was measured using a Seahorse XFe24 Analyzer (Agilent Technologies). One day before the assay, cells were seeded onto 24‐well plates (Agilent Technologies) coated with 0.01% poly‐L‐lysine at a density of 30 000 cells per well, in DMEM (PAN Biotech, P04–04510) high glucose supplemented with 10% heat inactivated FBS (iFBS, GE Healthcare Hyclone, SV30160.03) and 1× Antibiotic‐Antimycotic solution (Biosera, XC‐A4110). The next day, the culture medium was replaced with 450 μL of assay medium [DMEM without glucose and sodium pyruvate, pH 7.4 (Sigma‐Aldrich, D5030), supplemented with 2 mM L‐Glutamine (Biosera, XC‐T1715)] and three initial OCR measurements were taken. Next, sequential measurements were obtained following the addition of (1) 10 mM glucose (three measurements), (2) 2 μM oligomycin (two measurements), (3) 0.77 μM FCCP (two measurements) and (4) 2 μM rotenone + 1 μM antimycin A + 100 mM 2‐deoxy‐D‐glucose (2‐DG), pH 7.4 + 40 μM Hoechst 33342 (five measurements). OCR was normalized to the cell number using Seahorse XF Imaging and Normalization System (Agilent Technologies) with BioTek Cytation 5 Cell Imaging Multimode Reader (Agilent Technologies) and the obtained values were used to calculate basal respiration, ATP‐linked respiration, and maximal respiration.

2.13. Statistical Analysis

Statistical analysis was performed using GraphPad Prism 9. One‐way analysis of variance (ANOVA) with Tukey's post hoc test or two‐tailed Mann Whitney test was used to determine statistical significance. p values ≤ 0.05 were considered to be statistically significant.

3. Results

3.1. Clinical Presentation and Laboratory Results

Two siblings, the only offspring of healthy unrelated parents from a closed Cypriot community, were referred to the genetics clinic. The first patient (Patient 1), a male currently at the age of 9 years, was born at term via caesarian section due to preeclampsia with a birth weight of 3010 g. He failed the neonatal hearing test with initial reported sensorineural hearing loss of 45%–50%, which later progressed to 80%. He subsequently developed speech delay and dysarthria, which greatly improved following the placement of cochlear implants at the age of 4.5 and 5.5 years. His weight, height, and head circumference were normal. He exhibited a mild delay in motor milestones and was able to walk independently at the age of 19 months. His speech difficulties were attributed to his hearing impairment. He suffered several episodes of severe neutropenia after viral infections. Extensive investigations, including bone marrow aspiration and myelogram were unrevealing.

Patient 2, currently at the age of 7 years, is a female patient who was also born at term via caesarian section due to preeclampsia with a birth weight of 3482 g. Like her older sibling, she had an abnormal hearing test with an 80% hearing loss. Cochlear implants were placed earlier in her case, at 3 and 3.5 years of age. Unlike her brother, she has never presented with neutropenia. She was also reported to achieve developmental milestones earlier than her brother and to have better language and learning skills. Her growth and head circumference were normal, and she was able to walk independently at the age of 15 months.

Because of their hearing loss, both siblings underwent cytogenetic investigation, which revealed a normal karyotype. A duplication of approximately 0.6 Mb on the long arm of Chromosome 11 was identified in Patient 1 by array CGH. The same duplication was found in the patient's father but not in Patient 2. Urine organic acid analysis performed at the age of 6 months for Patient 1 and 12 months for Patient 2, as part of their metabolic investigation, revealed increased excretion of 3‐methylglutaconic (3‐MGC) (4.8 and 4.4 times above the upper limit of normal, respectively) and 3‐methylglutaric (3‐MG) acid, which persisted during follow‐up (Figure 1A,B). Blood lactate levels were mostly and repeatedly found to be above the normal reference range (Figure 1B). Together, the above findings were suggestive of a mitochondriopathy. A brain MRI scan performed at the age of 3.5 years (Patient 1) and 21 months (Patient 2) showed bilateral symmetric hyperintense changes in the globus pallidus on T2‐weighted axial images (Figure 1C), further supporting the possibility of an underlying mitochondrial disorder.

FIGURE 1.

Metabolic and brain abnormalities of patients. (A) Urine organic acid profiles obtained by gas chromatography–mass spectrometry (GC–MS). A marked elevation in the amount of 3‐methylglutaric (3‐MG) and 3‐methylglutaconic (3‐MGC) acid (cis and trans forms) is evident in both patients. A chromatogram from an unrelated, age‐matched control individual is shown for comparison purposes. (B) Urinary concentrations of 3‐MGC and 3‐MG acid determined by GC–MS in both patients are 4.4–4.8 and 14–15 times above the upper normal limit, respectively. Blood lactate levels, lowest and highest values of repeated measurements during patient follow‐up. Mean values and number of repeated measurements are shown in parentheses. (C) T2‐weighted axial brain MRI images of Patients 1 and 2 showing bilateral symmetric hyperintense changes in both globi pallidi (arrows). IS: internal standard (margaric acid).

3.2. Identification of a Novel YME1L1 Variant

To determine the genetic cause of 3‐MGC/3‐MG aciduria, both siblings were referred for whole exome sequencing analysis. A novel variant, c.1999C>G p.(Leu677Val), in exon 18 of the YME1L1 gene (NM_014263.4; NC_000010.10:g.27403458G>C; GRCh37) was identified in homozygosity in the patients, while both parents were accordingly heterozygous carriers (Figure 2A). The YME1L1 gene encodes a mitochondrial ATP‐dependent zinc metalloprotease (YME1L1), which is embedded in the inner mitochondrial membrane (IMM). YME1L1 together with the scaffold protein SLP2 and the rhomboid protease PARL form a proteolytic complex in the IMM, known as the SPY complex [20], implicated in the regulation of mitochondrial dynamics and quality control [21, 22]. Specifically, YME1L1 is implicated together with OMA1 in the proteolytic processing of the dynamin‐like GTPase OPA1, which is a key player in IMM fusion [23]. YME1L1 is synthesized as a precursor containing an N‐terminal mitochondrial targeting sequence (MTS) of 150 amino acids [9, 24]. Proteolytic cleavage of the MTS by the mitochondrial processing peptidase (MPP) gives rise to the mature form of YME1L1, which assembles into homo‐oligomeric complexes possessing proteolytic activity within the IMM [25, 26, 27]. The variant identified in the two patients results in a leucine to valine amino acid substitution at position 667 within the C‐terminal metalloprotease domain of the protein (Figure 2B). The affected amino acid is highly conserved across different species (Figure 2C).

FIGURE 2.

Identification of a novel YME1L1 variant. (A) Electropherograms showing the presence of the YME1L1 variant in homozygosity in Patient 1 (same for Patient 2) and heterozygosity in both parents. (B) Domain organization of YME1L1 (NP_055078.1), indicating the position of the identified variant p.(Leu667Val). (C) Multiple alignment of YME1L1 amino acid sequences (generated by Clustal Omega [19]) from different species. The affected amino acid is indicated in red color. MTS: Mitochondrial targeting sequence; TM: Transmembrane; HESGH: Motif of metalloprotease (zinc) activity.

3.3. Effects of the c.1999C > G Variant on YME1L1 Abundance and Processing

We first sought to address possible consequences of the identified variant on the protein's abundance or processing. For this, we performed western immunoblot in patient‐derived fibroblasts using an antibody against the C‐terminal part (amino acids 348‐716) of YME1L1 (YME1L1_C). We found no significant changes in the abundance of either the precursor or mature form of YME1L1^L667V, as compared to the wild‐type protein in control fibroblasts (Figure 3A). This finding suggests that the variant does not influence the levels or processing of YME1L1. Intriguingly, in addition to the mature and precursor forms of YME1L1, a prominent, faster‐migrating band corresponding to a protein of ~40 kDa was detected only in the patient's fibroblasts (Figure 3A). To further investigate the nature and origin of this truncated protein fragment, we employed an antibody (YME1L1_N) directed against the N‐terminal 240 amino acids of the YME1L1 precursor. As with the YME1L1_C antibody, the abundance of the mature protein appeared similar in patient and control fibroblasts; however, unlike YME1L1_C, the YME1L1_N antibody did not detect the 40 kDa fragment, suggesting that this does not encompass the N‐terminal part of the protein (Figure 3B).

FIGURE 3.

Effects of the c.1999C>G variant on the levels and processing of YME1L1. (A) Representative western blot and densitometry for the evaluation of YME1L1 protein levels (precursor and mature forms), using an antibody against the C‐terminal part of YME1L1 (YME1L1_C, Proteintech, 11 510‐1‐AP). Densitometry analysis is presented as a ratio to GAPDH (n = 3). (B) Detection of YME1L1 using an antibody against the N‐terminal segment of the protein (YME1L1_N, ABclonal, A10402). GAPDH served as a loading control. (C, D) Western blot of cell lysates from HeLa cells transiently transfected with either wild‐type YME1L1‐myc‐FLAG or YME1L1^L667V‐myc‐FLAG probed with a myc (C) or YME1L1_C (D) antibody. Non‐transfected HeLa cells (C, D) and patient‐derived fibroblasts (D) served as controls. GAPDH was used as a loading control. The arrow indicates a truncated protein fragment of ~40 kDa detected only in patient's fibroblasts. Data are presented as mean ± SEM, ns, not statistically significant (one‐way ANOVA). m: Mature, p: Precursor, u: Unspecific.

The above findings may suggest that the truncated YME1L1 protein detected in patient‐derived fibroblasts could result from either aberrant splicing or proteolytic cleavage of the YME1L1^L667V protein. To address the above, we transiently expressed either wild‐type YME1L1 or YME1L1^L667V, fused to a C‐terminal myc‐FLAG epitope in HeLa cells. Expression of both wild‐type YME1L1 and YME1L1^L667V was verified in transiently transfected HeLa cells using an antibody against the myc epitope (Figure 3C). This detected both the mature and precursor forms of YME1L1 in each case, but no additional shorter fragment specific to cells expressing the YME1L1^L667V variant. Furthermore, western blotting with the YME1L1_C antibody failed to detect the 40 kDa fragment present in the patient's fibroblasts, in YME1L1^L667V‐transfected HeLa cells (Figure 3D). The above results most likely exclude aberrant proteolysis as the process giving rise to the 40 kDa YME1L1 fragment and may instead indicate a partial splicing defect.

3.4. Patient‐Derived Fibroblasts Display Fragmentation of the Mitochondrial Network

Given the well‐established role of YME1L1 in the regulation of mitochondrial fusion [28, 29], we next evaluated the morphology of the mitochondrial network in the patient's fibroblasts. As revealed by TOMM20 staining, the mitochondrial network in the patient's fibroblasts appeared less tubular, as compared to controls (Figure 4A). Quantification of parameters related to mitochondrial morphology and network complexity unveiled statistically significant differences between the patient's and control fibroblasts. In particular, mitochondria in patient's fibroblasts appeared smaller, as evidenced by decreased area and perimeter, and displayed reduced aspect ratios and form factors, demonstrating that YME1L1^L667V is associated with more globular and less elongated mitochondria (Figure 4B). Furthermore, the mitochondrial tubules in the patient's fibroblasts exhibited shorter branch lengths, indicative of a less interconnected mitochondrial network (Figure 4B). Taken together, the above findings suggest that YME1L1^L667V is associated with enhanced mitochondrial fragmentation, which may indicate impaired YME1L1 function.

FIGURE 4.

YME1L1^L667V is associated with mitochondrial fragmentation. (A) Representative confocal images of control and patient‐derived fibroblasts immunostained for the mitochondrial marker TOMM20. Inserts correspond to boxed areas at higher magnification. (B) Quantitative assessment of mitochondrial morphology and network complexity parameters in patient's and control fibroblasts, using the Mitochondria Analyzer plugin of ImageJ (Mean area, Control A, n = 36; Control B, n = 40; Patient, n = 49; Mean perimeter, Control A, n = 37; Control B, n = 41; Patient, n = 49; Mean aspect ratio, Control A, n = 37; Control B, n = 40; Patient, n = 49; Mean form factor, Control A, n = 38; Control B, n = 41; Patient, n = 48; Mean branch length, Control A, n = 36, Control B, n = 42; Patient, n = 49). Data are presented as mean ± SEM, ****p < 0.0001; ***p < 0.001, ns, not statistically significant (one‐way ANOVA). Scale bar: 10 μm.

3.5. Leu667Val Is Associated With Impaired Proteolysis of YME1L1 Substrates

It has been previously reported that YME1L1 loss‐of‐function impairs the proteolytic processing of OPA1, resulting in mitochondrial fragmentation in cells and tissues [28, 29, 30]. Furthermore, a homozygous YME1L1 variant, previously identified in a consanguineous family, was reported to cause mitochondriopathy associated with neurological dysfunction and enhanced mitochondrial fission [9]. Given the above and the observed fragmentation of the mitochondrial network in the patient's fibroblasts, we sought to assess the effects of the YME1L1^L667V variant on OPA1 processing. OPA1 undergoes proteolytic cleavage by OMA1 and YME1L1 in the IMM, at the S1 and S2 cleavage sites, respectively, resulting in the balanced accumulation of IMM‐anchored long (L‐OPA1) and soluble short (S‐OPA1) forms. IMM fusion requires L‐OPA1, whereas S‐OPA1 forms are associated with mitochondrial fission [23, 31]. Immunoblot analysis revealed a distorted cleavage pattern of OPA1 in patient's fibroblasts, as compared to controls, resulting in the accumulation of short forms and concomitant reduction of long forms of OPA1 (Figure 5A). We further evaluated the lipid transfer protein PRELID1, the turnover of which is mediated by YME1L1 [32]. PRELID1 protein levels were found to be significantly increased in the patient's fibroblasts, as compared to controls (Figure 5B). To exclude a transcriptional origin for PRELID1 accumulation in the patient's cells, we evaluated PRELID1 mRNA levels by quantitative PCR. These were found to be comparable to those of control fibroblasts, suggesting that elevated PRELID1 protein levels in the patient's fibroblasts are not due to transcriptional upregulation, but rather result from impaired proteolysis (Figure 5C).

FIGURE 5.

YME1L1^L667V variant is associated with impaired proteolytic activity. (A) Representative western blot and densitometry (ratio to GAPDH, n = 4) for assessing OPA1 proteolytic processing by YME1L1 in control and patient‐derived fibroblasts. An aberrant cleavage pattern of OPA1 resulting in the reduction of long (L‐OPA1) and accumulation of short (S‐OPA1) forms is observed in patient's cells. (B) Representative western blot and densitometry (ratio to GAPDH, n = 3) for the assessment of PRELID1 protein levels in cell lysates from patient's fibroblasts and healthy controls. (C) Assessment of PRELID1 mRNA expression levels by quantitative PCR. Values were normalized to mRNA levels of the TBP housekeeping gene (n = 3). (D) Representative western blot and densitometry (ratio to GAPDH, n = 3) for the evaluation of the fusion and fission regulators MFN2 and DRP1, respectively (n = 3). Data are presented as mean ± SEM, ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05; ns, not statistically significant (one‐way ANOVA).

To exclude secondary defects that could be associated with mitochondrial fragmentation, we also assessed the levels of MFN2 and DRP1, which are key regulators of mitochondrial fusion and fission, respectively. Comparable levels of MFN2 and DRP1 were detected in fibroblasts from both the patient and healthy controls (Figure 5D). Together, the above findings indicate that the Leu667Val variant compromises YME1L1 function and likely underlies the biochemical and cellular abnormalities, as well as the clinical manifestations seen in our patients.

3.6. YME1L1^L667V Is Associated With Impaired Krebs Cycle Enzyme Activity and Mitochondrial Respiratory Function

Several lines of evidence link secondary 3‐MGCA with mitochondrial dysfunction and particularly with attenuated activity of the Krebs cycle, which is thought to result in the diversion of part of the acetyl‐CoA towards the formation of 3‐MGC‐CoA, the precursor of 3‐MGC and 3‐MG acid [1]. Given our finding that YME1L1 dysfunction is associated with 3‐MGCA, we aimed to further explore the potential cause of 3‐MGCA in these patients. Towards this end, we first evaluated the activity of key Krebs cycle enzymes in the patient's and control fibroblasts. We found that the activities of citrate synthase (CS) and α‐ketoglutarate dehydrogenase (α‐KGDH) were significantly reduced in the patient's cells, as compared to controls, while the activities of succinate dehydrogenase (SQR) and fumarate dehydrogenase (FH) displayed no statistically significant difference (Figure 6A). We further investigated the functional consequences of YME1L1 Leu667Val variant on mitochondrial respiration by means of oxygen consumption rate (OCR) measurements using a Seahorse Extracellular Flux analyzer. We found that basal, ATP‐linked, and maximal respiration were significantly reduced in the patient's fibroblasts, as compared to controls, suggesting an attenuated mitochondrial respiratory capacity (Figure 6B).

FIGURE 6.

YME1L1^L667V is associated with Krebs cycle dysfunction and attenuated mitochondrial respiration. (A) Assessment of the enzymatic activities of the Krebs cycle enzymes CS, αKGDH, SQR, and FH in patient's and control fibroblasts (n = 3 for αKGDH and FH; n = 4 for CS and SQR). (B) Seahorse oxygen consumption rate (OCR) measurements and quantification of basal, ATP‐linked and maximal respiration in control and patient‐derived fibroblasts. The control group represents combined measurements from dermal fibroblasts derived from three unrelated healthy individuals. (C) Representative western blot and densitometry for the evaluation of the abundance of respiratory chain enzyme complexes (CI–CV) in protein extracts from control and patient's fibroblasts (n = 3). GAPDH served as loading control, and densitometry is expressed as a ratio to GAPDH. Data are presented as mean ± SEM, ****p < 0.0001; ***p < 0.001; **p < 0.01; *p < 0.05, ns, not statistically significant (one‐way ANOVA for A, C; two‐tailed Mann Whitney for B). α‐KGDH: α‐ketoglutarate dehydrogenase, CS: Citrate synthase, FH: Fumarate hydratase, SQR: Succinate dehydrogenase.

Assessment of components of respiratory chain enzyme complexes did not reveal any statistically significant changes in their abundance in the patient's fibroblasts (Figure 6C). The above findings indicate that a diminished capacity of the Krebs cycle and mitochondrial respiration may underlie 3‐MGCA in the herein reported patients.

4. Discussion

We report, for the first time, that YME1L1 dysfunction, caused by a newly identified homozygous missense variant [c.1999C>G p.(Leu667Val)], is associated with 3‐MGCA. To our knowledge, this is the second report of patients in the literature featuring YME1L1‐associated mitochondriopathy. The first link between a pathogenic YME1L1 variant, p.(Arg149Trp), and mitochondrial disease was demonstrated in four children from a single consanguineous pedigree of Saudi Arabian descent [9]. Whether these patients also exhibited 3‐MGCA remains unknown, since this was not assessed in the above study. Moreover, evaluation of these patients for 3‐MGCA for the purposes of this study was not feasible, since they have been lost to follow‐up. An association between YME1L1 dysfunction and 3‐MGCA is also indirectly supported by a recent study reporting that impairment of the IMM disaggregase CLPB disrupts the integrity and function of the SPY complex, of which one of the components is YME1L1 [33]. Pathogenic variants in CLPB are known to be associated with 3‐MGCA and neutropenia (OMIM #619835) [34]. Consistent with the above, one of the patients reported in this study exhibited, in addition to 3‐MGCA, severe neutropenia, which may underscore the reported functional link between CLPB and components of the SPY complex.

The herein described patients exhibit features of a mitochondrial disorder, including motor delay, ataxia, sensorineural hearing loss, mild developmental delay, and learning challenges, while their brain MRI revealed symmetrical high signal in the basal ganglia. These clinical manifestations partially overlap with those observed in the previously reported patients [9] and also in cases of 3‐MGCA caused by mutations in other genes [1, 7]. In general, the patients described in the initial report had a more severe clinical presentation, as compared to the herein reported cases, characterized by motor, developmental, and speech delay, intellectual disability, hearing impairment, leukoencephalopathy, and optic nerve atrophy. Additional features, such as short stature, low weight, and microcephaly or macrocephaly, were inconsistently present [9]. As opposed to these patients [9], the siblings described in this study have normal growth and head circumference and show no evidence of optic nerve atrophy so far, as assessed by ophthalmological examination performed once a year since diagnosis. The milder clinical manifestations could be related to the nature of the amino acid substitution and its position within the YME1L1 protein in each case. In our patients, the identified variant results in a conservative amino acid substitution of leucine to valine, p.(Leu667Val). The affected amino acid lies within a highly conserved region of the C‐terminal protease domain of YME1L1, and according to our findings, does not appear to interfere with the maturation of the protein. On the other hand, the previously identified Arg149Trp variant substitutes a positively charged, polar arginine residue with a non‐polar, bulky aromatic amino acid (tryptophan) within the MTS of YME1L1. The above amino acid change was shown to impair MPP‐mediated processing and maturation of YME1L1, resulting in the subsequent degradation of the accumulated precursor [9], which could be related to a more severe clinical presentation.

It is noteworthy that for both siblings reported in this study, maternal preeclampsia was present during the final stages of pregnancy. Recent studies revealed a strong correlation between impaired mitochondrial function and preeclampsia, supporting that mitochondrial dysfunction contributes to its pathophysiology [35]. Additionally, a shift in mitochondrial dynamics towards fission was reported in preeclamptic placenta [36]. The effects of deleterious YME1L1 variants on mitochondrial morphology and function, as described by us and others, may thus be related to an increased risk for maternal preeclampsia.

Our results indicate the presence of a ~ 40 kDa truncated YME1L1 protein, in addition to the full‐length mature protein, in the patient's fibroblasts. This shorter protein fragment was not detected in HeLa cells transiently expressing YME1L1^L667V, suggesting that it likely results from a partial splicing defect. This could be related to the position of the C>G transition, which lies in the proximity of the exon 18/intron junction. As the origin and nature of this truncated YME1L1 fragment are currently elusive, its potential role in disease pathogenesis remains unknown.

We provide evidence that the Leu667Val variant is associated with impaired proteolytic activity of YME1L1, supporting its pathogenicity. This is demonstrated by the increased protein levels of PRELID1 in the patient's fibroblasts. PRELID1 is a lipid transfer protein which is constitutively degraded by YME1L1, as a mechanism to regulate cardiolipin levels and resistance to apoptosis [32]. Moreover, we demonstrate aberrant proteolytic processing of OPA1, giving rise to the accumulation of short forms of OPA1 and increased mitochondrial fission in patient‐derived fibroblasts. Aligned with our results, the previously reported YME1L1 Arg149Trp variant was also reported to induce accumulation of PRELID1 and fragmentation of the mitochondrial network associated with abnormal OPA1 processing [9]. Additionally, YME1L1 loss‐of‐function in the heart or the nervous system of mice, resulted in compromised proteolytic cleavage of OPA1, leading to mitochondrial fragmentation [29, 30], Similar to YME1L1, defects in other genes associated with 3‐MGCAs have been shown to disturb mitochondrial morphology or OPA1 processing. Biallelic pathogenic variants in the HTRA2 gene (OMIM #617248) which encodes a mitochondrial serine protease impair OPA1 proteolytic processing, resulting in the accumulation of S‐OPA1 forms. However, no impairment of mitochondrial morphology was observed, suggesting an indirect effect of HTRA2 on OPA1, which may result in transient fusion events influenced by OPA1 levels [37, 38]. Furthermore, fibroblasts from patients with Costeff optic‐atrophy syndrome (OMIM #258501) caused by compound heterozygous variants in OPA3, exhibited morphological alterations in the mitochondrial network, compatible with fragmented mitochondria [39]. Our findings further demonstrate that the YME1L1 Leu667Val variant does not interfere with the protein's maturation, suggesting that the MPP‐mediated processing of the protein in the IMM is not disrupted. Conversely, and despite the fact that the previously reported YME1L1^R149W mutant was correctly targeted to mitochondria, its proteolytic maturation by MPP was compromised [9].

We found reduced activities of CS and α‐KGDH in patient‐derived fibroblasts, without significant changes in FH or SQR activity. These findings may indicate a diminished capacity of the Krebs cycle. CS catalyzes the condensation of oxaloacetate with acetyl‐CoA to generate citrate, an irreversible reaction representing the first rate‐limiting step of the Krebs cycle and a crucial determinant of the energy‐generating process through mitochondrial respiration [40, 41]. A reduction in CS activity could restrict the availability of downstream intermediates, thus affecting mitochondrial energy production. Additionally, the α‐KGDH complex catalyzes the second rate‐limiting step of the Krebs cycle, the irreversible conversion of α‐ketoglutarate to succinyl‐CoA in the presence of NAD⁺ and coenzyme A (CoA). This enzyme complex is tightly regulated by several factors such as CoA and NAD⁺, and is particularly sensitive to feedback inhibition by increased succinyl‐CoA/CoA and NADH/NAD⁺ ratios [42]. Consistent with the above, mitochondrial respiratory capacity was found to be reduced in patient‐derived fibroblasts. Together, the attenuated activity of Krebs cycle enzymes and compromised mitochondrial respiration may provide a mechanistic link to the 3‐MGCA observed in our patients. In particular, our findings are in support of the acetyl‐CoA diversion pathway model, which suggests the formation of 3‐MGC‐CoA, the precursor of 3‐MGC and 3‐MG acid, from acetyl‐CoA, not processed in the Krebs cycle [4]. According to this hypothesis, compromised energy production leads to the elevation of the NADH/NAD⁺ ratio within mitochondria resulting in the inhibition of Krebs cycle enzymes [4, 43]. Ultimately, this metabolic obstruction results partly in the redirection of acetyl‐CoA towards the production of 3‐MGC and 3‐MG acid, through their precursor 3‐MGC‐CoA [4].

In conclusion, our findings link YME1L1 dysfunction with 3‐MGCA. YME1L1 should thus be considered a potential candidate gene in the diagnostic evaluation of patients with 3‐MGCA presenting with sensorineural hearing impairment and neurological manifestations.

Author Contributions

Petros P. Petrou and Olga Grafakou conceptualized and designed the study. Anthi Demetriadou performed the experiments, analyzed the data, and wrote the manuscript. Petros P. Petrou evaluated and interpreted data, wrote the manuscript, coordinated, and supervised the study. Olga Grafakou performed clinical evaluation of the patients, wrote the manuscript, and provided skin biopsy. Theodoros Georgiou performed whole‐exome sequencing and interpreted biochemical results. Daniela Burska performed and analyzed Krebs cycle activity experiments. Anna Malekkou performed Sanger sequencing, family study, and variant classification. Jana Krizova performed and analyzed Seahorse experiments. Efstathia Paramera, Gavriella Mavrikiou, and Maria Dionysiou performed biochemical tests. Athina Theodosiou analyzed whole‐exome sequencing data. Carolina Sismani interpreted genetic data. Violetta Anastasiadou and Ioannis Ioannou performed clinical evaluation of the patients. Evangelos Papakonstantinou and Anthi Drousiotou interpreted biochemical findings. Hana Hansikova interpreted Krebs cycle enzyme activity and Seahorse measurements. All authors reviewed, edited, and approved the final manuscript.

Ethics Statement

The study was conducted in accordance with the Helsinki Declaration (https://www.wma.net/policies‐post/wma‐declaration‐of‐helsinki/) and was approved by the Cyprus National Bioethics Committee. Informed consent was obtained from the patients' parents.

Consent

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. Informed consent was obtained from all patients for being included in the study.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1.

Data Availability Statement

The variant reported in this study, NM_014263.4(YME1L1):c.1999C>G (p.Leu667Val) has been classified as likely pathogenic according to the joint consensus recommendation of the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen) [44] and submitted to ClinVar with the accession number SCV005373488. The data that support the findings of this study will be made available by the corresponding author upon reasonable request.