Strategic validation of variants of uncertain significance inECHS1genetic testing

1 ECHS1 is the causative gene for mitochondrial short -chain enoyl -CoA hydratase 1 2 deficiency. While genetic analysis studies have diagnosed numerous cases with ECHS1 3 variants, the increasing number of variants of uncertain significance ( VUS) in genetic 4 diagnosis is a major problem. Therefore, we constructed an assay system to verify VUS 5 function. A high-throughput assay using ECHS1 knockout cells was performed to index 6 these phenotypes by expressing cDNAs containing VUS. The functional validation of 7 VUS identified novel variants causing loss of ECHS1 function . Moreover, we identified 8 cases with functional ECHS1 defects through multi -omics analysis. We identified a 9 synonymous substitution , p.P163=, and candidate patho genic variants in the above 10 validation experiments. In summary, this study uncovered new ECHS1 cases based on 11 VUS validation and omics analysis ; these analys es are applicable to functional 12 evaluation of other genes associated with mitochondrial disease. 13 14

15 Variants of uncertain significance, Mitochondrial disease, RNA -seq, H igh-throughput 16 assay, Multi-omics 17 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint

1 ECHS1 encodes short-chain enoyl -CoA hydratase 1 (SCEH) , responsible for the 2 degradation of branched chain amino acids and fatty acids (Sharpe & McKenzie, 2018). 3 Abnormalities in valine metabolism particularly impact the pathogenesis of SCEH 4 deficiency (MIM #616277) ; accumulation of valine metabolites such as S -(2-carbox-5 ypropyl) cysteine (SCPC) and S -(2-carboxypropyl) cysteamine (SCPCM) derived from 6 methacrylyl-CoA, and S -(2-carboxyethyl) cysteine (SCEC), S -(2-carboxyethyl) 7 cysteamine ( SCECM), and 2 -methyl-2,3-dihydroxybutyric acid (MDHB), derived from 8 acryloyl-CoA, is often observed (Peters et al, 2014). Mutations in ECHS1 mainly cause 9 Leigh encephalopathy with patients presenting elevated plasma lactate and brain 10 magnetic resonance imaging (MRI) abnormalities (Peters et al , 2015; Yamada et al , 11 2015; Haack et al , 2015a; Sakai et al , 2015; Tetreault et al , 2015) . In addition, 12 mitochondrial respiratory chain complex abnormalities have also been reported to cause 13 mitochondrial dysfunction in cases with ECHS1 mutations (Sakai et al, 2015; Haack et 14 al, 2015a; Tetreault et al, 2015). Numerous Japanese cases with ECHS1 mutations have 15 also been reported (Sakai et al, 2015; Haack et al, 2015a; Yamada et al, 2015; Ogawa 16 et al , 2017, 2020) . Numerous ECHS1 variants have been reported, among which, 17 pathogenic variants exist in various genetic regions. A variant frequently reported in 18 Asians is Asn59Ser. The expansion in the clinic of genetic testing has resulted in the 19 rapid accumulation of variants of uncertain significance (VUS). Mitochondrial diseases 20 are no exception, and the VUS number in ECHS1 is increasing. On the other hand, 21 recent studies have also shown that a valine -restricted diet is effective for case s with 22 ECHS1 mutations (Yang & Yu, 2020; Sato -Shirai et al , 2021) . Quick diagnosis is 23 important for early treatment of SCEH deficiency. 24 Recently, various approaches have been tried to solve VUS. Functional analysis using 25 cultured cells and model organisms is a powerful validation method, providing strong 26 evidence of pathogenicity according to the American College of Medical Genetics and 27 Genomics (ACMG) guidelines (Richards et al, 2015). Especially, high-throughput assays 28 and assay methods combining CRISPR/Cas9 and genome-sequencing technologies are 29 being used for VUS verification in cancer-causing genes (Findlay et al, 2018; Kweon et 30 al, 2020) . Although VUS have been validated for IVD, ACADVL, and ACAD9, all 31 causative genes of inborn errors of metabolism (D’Annibale et al, 2021; Xia et al, 2021; 32 D’Annibale et al, 2022), there is little research on VUS verification in rare diseases. 33 In this study, we focused on VUS in ECHS1, common in many cases of mitochondrial 34 diseases (Kohda et al , 2016; Ogawa et al , 2017, 2020) . In Japan, Tohoku Medical 35 Megabank has a genome project for healthy subjects (Tadaka et al , 2019, 2021). In 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint recessive diseases, the causative variant is often present at a certain frequency within a 1 given race. Therefore, in addition to the VUS found in previous our studies of patients 2 with mitochondrial diseases, we conducted a variant validation of rare variants registered 3 in the Japanese Multi-Omics Reference Panel (jMorp). 4 Here, we constructed an assay system with ATP measurement using cells deficient in 5 the ECHS1 gene for systematic VUS verification and validation of heterozygotic variants. 6 Furthermore, we found a novel ECHS1 variant by multi-omics analysis. Coincidentally, 7 the newly found variants were verified by the assay system as being caused by ECHS1. 8 9

10 Cases with ECHS1 variants 11 In previous genomic studies, gene panel sequencing and whole exome sequencing led 12 to the discovery of 15 variants in ECHS1 cases (Table 1). In Japanese people, the most 13 frequently identified variant was Asn59Ser, followed by Ala2Val (Haack et al , 2015b; 14 Sakai et al, 2015; Ogawa et al, 2020) (Table 1 and Fig. 1). These variants are common 15 in other Asian cases; the allele frequencies of Asn59Ser and Ala2Val are 0.0005769 and 16 0.0001927 in the gnomAD v3.1.2 East Asian, respectively. Several of these cases also 17 had VUS in ClinVar (as of 20220715, and the same afterhere ), such as Met1Val and 18 Leu8Pro (Ogawa et al , 2017; Uchino et al , 2019) . In addition, we also found 19 heterozygous variant s in ECHS1 with conflicting interpretations of pathogenicity 20 (Thr266Ala and Ala278Thr), VUS (Arg272Gln), or unreported (Ala268Thr) . All th ese 21 variants are rare and may be disease -causing, but no experimental verification of the 22 variants has been performed so far . Variants with low allele frequency have been 23 identified, and a number of them have been designated as VUS. Accurate and quick 24 interpretation of these variants is essential for improved diagnosis. 25 26 VUS validation in ECHS1 27 To perform VUS validation of ECHS1, we first generated cells deficient in the ECHS1 28 gene by targeting ECHS1 using the CRISPR/Cas9 system in HEK293FT cells. The 29 ECHS1 KO cell line had an in-frame deletion of 18 bases in exon 2 of the ECHS1 gene 30 (Fig. 2A). Th is in-frame deletion resulted in loss of ECHS1 mRNA and protein, as 31 confirmed by qRT-PCR and western blotting (Fig. 2B). We also detected abnormalities 32 in the mitochondrial function in ECHS1 KO cells. In galactose medium, ATP production 33 is largely dependent on mitochondrial respiration (Robinson et al, 1992). Accordingly, we 34 measured ATP levels of wild-type (WT) and ECHS1 KO cells after incubation in glucose 35 and galactose media. The ratio of ATP levels under galactose and glucose conditions 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint was stable, and no significant difference was observed between ECHS1 KO and WT 1 cells under these conditions (Fig. 2C). The cellular toxicity of ECHS1 deficiency is due 2 to accumulation of intermediates in valine metabolism. That explains why a v aline-3 restricted diet is recommended to patients harboring ECHS1 pathogenic variants (Sato-4 Shirai et al, 2021). Consistently, valine addition to the galactose medium lead to further 5 reduced ATP level in ECHS1 KO cells, in a dose -dependent manner (Fig. 2D). These 6

suggest that the abnormal mitochondrial function in ECHS1 KO cells results from 7 the abnormal valine metabolism, as shown in human patients, making it a suitable model 8 to validate VUS of ECHS1. Accordingly, we performed a functional verification of VUS 9 using ECHS1 KO cells expressing VUS. 10 First, variants identified from previous genomic studies and from jMorp were selected for 11 functional validation. The targets were 15 variants identified by our genome analysis and 12 six rare variants registered in jMorp (4.7KJPN released on 20190902; Fig. 1 and Table 13 S1). The Leu8Pro and Asn59Ser variants were found in our patients as well as registered 14 in jMorp (4.7KJPN). Other four varian ts were not identified in our genome analysis of 15 mitochondrial disease patients . ECHS1 is a mitochondria -localized protein with a 16 mitochondrial targeting sequence (MTS) at the N-terminal. Four variants were located at 17 the translation initiation site and the MTS. We compared ECHS1 KO cells transfected 18 with an empty vector and ECHS1 WT cDNA to ECHS1 KO cells transfected with an 19 ECHS1 gene including VUS. 20 Exogenously expressed ECHS1 shows uncleaved and cleaved forms (Fig. 3A, single 21 and double asterisks, respectively). Some variants were dominantly expressed in an 22 uncleaved form. Further, we found a significant decrease in the expression level of 23 variants in the MTS and at leucine 145 (Fig. 3A). 24 Next, we examined VUS functions in the mitochondria. ECHS1 WT expression restored 25 ATP levels in ECHS1 KO cells (Fig. 3B). The ATP expression of KO cells transfected with 26 each variant was compared with that for KO cells transfected with empty vector or 27 transfected with WT vector; changes in ATP levels were evaluated by one-way ANOVA 28 followed by Dunnet’s test (Fig. 3C). In addition to Ala2Val and Asn59S, reported as 29 pathogenic variants, the two VUS with altered start codons failed to restore ATP levels 30 probably due to significantly reduced protein expression (Fig. 3B). A marked down -31 regulation in protein levels was observed, with Leu8Pro showing a significant reduction 32 in the lower band considered to correspond to mature ECHS1 (Fig. 3A). His119Gln, 33 Pro163Leu, Ala268Thr, and Ala278Thr variants had higher expression levels, but failed 34 to fully improve ATP levels equivalent to WT, suggesting that these variants were 35 deleterious (Fig. 3B). 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Although Leu8Pro and Phe33Ser were reported as pathogenic variant s (Kohda et al, 1 2016; Uchino et al , 2019) , ATP assays demonstrate a mild functional decline . Since 2 Leu8Pro and Phe33Ser were identified as a compound heterozygote with Asn59Ser, a 3 nonfunctional variant (Fig. 3B), we hypothesized that Leu8Pro and Phe33Ser might be 4 mild deleterious variants capable of restoring KO cells but not KO cells expressing highly 5 toxic variant s. To test this hypothesis, we reproduce d compound hetero zygous 6 genotypes by transfecting ECHS1 KO cells with two different variants (Fig. 4A). The ATP 7 assay demonstrated that WT restored KO cells upon co -transfection with As n59Ser, 8 whereas, as expected, all three variants tested did not show recovery (Fig. 4B and C). 9 Our system is valuable to validate the pathogenic nature of the compound heterozygous 10 state. 11 12 Validation of ECHS1 variants from omics analysis 13 In case 1, a T hr266Pro variant was identified; in case 2, a n Ala268Thr variant was 14 identified; in cases 3 and 4, a Ala278Thr variant was identified; in case 5, an Ala278Val 15 variant was identified (Table 2, Fig. S2 ). In addition, these cases also had a P ro163= 16 variant, identified in a Samoan family and a very frequent variant (0.01156) in gnomAD 17 v3.1.2 East Asian populations, and suggested to exhibit splicing abnormalities (Simon et 18 al, 2021). In the present study, RNA-seq analysis was performed on case 1 and case 4. 19 Then, w e identified sequence reads showing exon skipping as consequence of the 20 Pro163= variant (Fig. 5A). To examine whether exon skipping actually increased in cases 21 with Pro163=, we plotted RNA -seq counts of the ECHS1 gene and the number of 22 detected reads showing exon skipping (Fig. 5B). We analyzed 26 RNA -seq data , 23 including case 1 and case 4, as well as o ne case with heterozygous Pro163= ; exon 24 skipping increased in these samples. In addition, after counting the allele numbers at the 25 Thr266Pro and Ala278Thr variant positions showed a bias in allele expression (Fig. 5C). 26 This indicates that the significantly reduced expression of alleles with Pro163= in case 1 27 and case 4 fibroblasts. Furthermore, proteome analysis in case 1 fibroblasts confirmed 28 a significant decrease in ECHS1 expression (Fig. 6A). In addition, protein expression in 29 case 1 fibroblasts was confirmed by sodium dodecyl sulfate polyacrylamide gel 30 electrophoresis (SDS-PAGE) followed by Western blot (WB), revealing a marked decrease 31 in protein expression (Fig. 6B). Furthermore, previous s tudies experimentally 32 demonstrated that case 4 leads to decreased ECHS1 expression and enzymatic activity, 33 as well as accumulation of intermediate products of valine metabolism (Kuwajima et al, 34 2021). We here concluded that and non -synonymous substitution combinations are 35 disease-causing in these cases. 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint 1

2 ECHS1 is one of the most frequent genes found in patients with mitochondrial diseases. 3 In addition to our genetic study (Haack et al, 2015b; Ogawa et al, 2017, 2020), ECHS1 4 has been reported as a frequent cause of mitochondrial disease in genetic studies of 5 Leigh encephalopathy in Asian countries such as China (Stenton et al , 2022 ) and 6 Korea(Lee et al, 2020). The length of the coding region is 873 bp . About 250 variants 7 have been reported in ClinVar, of which about 40 are pathogenic/likely pathogenic and 8 nearly 80 have been registered as VUS or with conflicting interpretations of pathogenicity. 9 However, the molecular mechanism by which ECHS1 variants cause disease is poorly 10 understood. In this study, we established a system to functionally validate ECHS1 11 variants, finding some potential pathogenic variants. 12 Resolving VUS is a major challenge in various diseases , because VUS continue to 13 accumulate under the circumstances where genome analysis is becoming more 14 common. Genome sequence projects in healthy individuals have revealed allele 15 frequencies in various races. The accumulation of genome analysis information is 16 expected to lead to the discovery of novel pathog enic variants, since rare variants also 17 found in healthy individuals can be pathogenic. Against this background, it may be 18 possible to make a proactive evaluation of variants not yet been associated with disease. 19 Based on the above, we could identify novel pathogenic variants in this study. 20 Furthermore, the combination of RNA -seq, proteomics and conventional genomic 21 analysis enabled a reliable diagnosis. 22 ECHS1 is a nuclear -encoded gene transported to mitochondria by the MTS (Burgin & 23 McKenzie, 2020) . MTS locates at the N-terminus and is cleaved in mitochondria by 24 mitochondrial proteases(Vaca Jacome et al, 2015). There are three points involved in 25 pathogenicity: 1) expression, 2) location, and 3) function. It was reported that ECHS1 is 26 degraded through ubiquitin-proteasome pathway in cancer cells (PMID: 34615856). As 27 shown in Fig. 2A, Met1Ala, Met1Thr, and Ala2Val were weakly expressed. Consistently, 28 these variants failed to restore ATP decrease in ECHS1 KO cells. Sinc e inhibition of 29 proteasome degradation by MG132 had no effect on the protein level of all ECHS1 30 variants tested in this study (Fig. S1), these substitutions may alter the expression level 31 at the transcriptional or translational level. The Leu8Pro variant was detected dominantly 32 in the uncleaved form (Fig. 3A), suggesting mis -targeting to mitochondria. Leu145Pro 33 was also weakly expressed, but increased expression with higher amount of transfection 34 did not restore. As with Asn54Ser, Leu145Pro is a functionally disrupted variant. 35 Interestingly, Leu145Pro, a variant reported only in jMorp and not previously identified in 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint patients with the disease, had a high molecular weight . It is expected to be reported in 1 Japanese case with this variant. We propose that those who identify variants in ECHS1 2 by genetic test will refer to our variant evaluation study as evidence and that this will lead 3 to a solution to the cause of the disease in the future. 4 Leu8Pro and Phe33Ser, considered “Likely benign” varia nts from VUS validation 5 experiments (Fig. 3B and C), were validated by co-expressing with the Asn59Ser variant 6 (Fig. 4B and C). By validating the assumption of complex heterozygosity in combination 7 with a pathogenic variant, we could obtain data accurately reflect ing the functional 8 evaluation of hypomorphic variants. The assay was effective for variants in the borders, 9 for which ve ry subtle validation results were obtained. Despite its higher experimental 10 complexity, the assay can be performed while maintaining the conventional throughput, 11 and the validation targets can be expanded. In our reported ECHS1 mutant cases, the 12 most common combination of is Asn59Ser and Ala2Val. However, since Asn59Ser has a 13 higher allele frequency than Ala2Val, cases homozygous for Asn59Ser should be more 14 frequent; however, no patients homozygous for Asn59Ser have been found. This 15 suggests that Asn59Ser may be so harmful that in homozygosity it may result in severe 16 developmental abnormalities in the prenatal period. The high number of Asn59Ser and 17 Ala2Val combinations may be due to Ala2Val having a smaller functional loss than 18 Asn59Ser, as shown experimentally in this study. Asn59Ser is only viable in a compound 19 heterozygous state with less toxic variants. The combination with milder variants such 20 as Leu8Pro and Phe33Ser, verified in this study, might allow normal development until 21 birth. In other words, this combined assay could first accurately indicate functional 22 abnormalities for variants with abnormalities intermediate between normal and 23 completely defective, being a very effective and essential method for variant evaluation. 24 Mostly, in silico predictions and experimental validation of most variants are comparable. 25 On the other hand, silico predictions for some variants differ from our validation 26 experiments; therefore, validation experiments are important for such variants. For 27 example, Ala2Val, already been reported as pathogenic, was not highly damaging in in 28 silico predictions (Table S1). However, the validation results suggest that although the 29 experimental data indicat e an effect on protein gene expression itself, the functional 30 effect may not be as significant. A similar trend was observed for Leu8Pro, expected to 31 have a significant effect on protein localization but relatively little effect on protein function. 32 His119Gln was also rated mostly tolerant or benign in silico, but experimental validation 33 suggested that the variant affects gene function. The major achievement of this study is 34 providing quantitative experimental validation data for each variant on the s ame 35 functional platform. We are expanding the validation of other genes ; in particular, VUS 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint validation of genes involved in respiratory chain complex I confirmed this to be a 1 reproducible and efficient analysis. 2 In this validation, we focused on variants specific to Japanese and Asian populations. 3 Naturally, this should be further expanded to include variants from other ethnicities. Our 4 assay system is based on a very simple method and can easily be expanded to other 5 VUS, by including gnomAD and ClinVar variants. In the present study, VUS was verified 6 for jMorp (4.7KJPN) variants, but since then, the jMorp data has been updated and the 7 number of registrations ha s increased. While planning our validation experiment, the 8 variant data of 4.7K JPN were registered, from which we extracted variants with very low 9 allele frequencies (Table S2) . However, there are now 38K JPN, with 20 more 10 registrations from our validated variants. The data is being updated at an accelerated 11 pace, with 14KJPN in 2021 and 38 KJPN in 2022. Given the constant updates to the 12 database, it is important to verify these variants in the future. 13 ECHS1 protein loss is considered to have a threshold of 30 –40% for disease(Simon et 14 al, 2021). This is shown by the analysis with Pro163= and Ala278Thr from the study of 15 Simon et al. Those homozygous for Pro163= did not develop the disease, despite 16 showing a protein expression around 40%. On the other hand, the combination of 17 Pro163= and Ala278Thr showed <30% protein expression and developed the disease. 18 Ala278Thr is considered highly toxic, as it was found to be deleterious in the VUS 19 validation experiment. Although Ala278Val has not been validated, we think it likely to be 20 as deleterious as Ala278Thr since many scores in silico also indicated damaging effects. 21 For Thr266Pro, no significant decrease in ATP levels was observed in VUS validation. 22 However, abnormalities at the protein level were evident from the proteome and 23 immunoblotting. VUS expression experiments with the Thr266Pro variant also showed 24 reduced ECHS1 expression (Fig. 3A). Considering these results, Thr266Pro might have 25 less functional loss and more impact on protein expression. In addition, the forced 26 expression system showed a protein decrease not expected to significantly reduce ATP 27 levels. In protein expression validation (Fig. S1) , the amount of mature ECHS1 28 synthesized from Thr266Pro was lower than that of i mmature ECHS1, suggesting that 29 this variant has a significant effect on protein expression and maturation. For Ala268Thr, 30 a mild decrease in ATP levels was observed. Since the value of Ala268Thr was similar 31 to that of Pro163Leu, reported as likely pathogenic in ClinVar, Ala268Thr was considered 32 to have an effect on protein function. Case 2 with the Ala268Thr variant had later onset 33 and milder symptoms than other ECHS1 cases, suggesting that the variant itself has a 34 milder effect. Pro163=, despite its high allele frequency [0.01156 in gnomAD v3.1.2 East 35 Asian, 0.00789 in the jMorp (38KJPN)] , is a possible causa tive variant; its combination 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint with other pathogenic variants leads to disease development. We found four cases with 1 Pro163= and other rare variant in this study. There have been very few reports of such 2 a high frequency variant in studies of mitochondrial diseases; this variant might be 3 responsible for the increased frequency of cases with ECHS1 mutations. However, 4 variants without amino acid substitutions have been overlooked as benign in previous 5 genetic diagnosis. Given the high potential number of patie nts with Pro163=, it is 6 expected that the number of cases with Pro163= will further increase because of our 7 findings. 8 Our VUS verification system has some limitations. Even patho genic variants may be 9 missed in this experimental system. In fact, Pro163Leu, registered as likely pathogenic 10 in ClinVar, showed a mild score. Thus, variants with intermediate ratings could be 11 pathogenic. In such cases, it would be necessary to validate them using an assay system 12 that assumes compound heterozygosity. Moreover, since this is a forced expression 13 system, variants that cause protein stability or splicing abnormalities may be missed. To 14 extract these variants, other assay systems and functional experiments using patient 15 specimens are required. 16 17

18 Cell culture and knockout cell generation 19 Cells were cultured at 37°C and 5% CO2 in Dulbecco's modified Eagle’s medium (DMEM 20 with 4.5 g/L glucose; Nacalai Tesque) supplemented with 10% fetal bovine serum and 21 1% penicillin–streptomycin. 22 Single guide RNAs (sgRNAs) were designed using CRISPRdirect software (Naito et al, 23 2015).he target sequence was as follows: 5’-GGGCCTTGGGGCGGTTCAGT-3’. gRNA 24 oligonucleotides were inserted into a pSpCas9(BB) -2A-Puro (PX459) V2.0 (Addgene 25 62988) plasmid as previously described (Ran et al , 2013) . HEK293FT cells were 26 transfected with PX459 including ECHS1 targeted sgRNA. Cells were selected using 2 27 µg/mL puromycin and single cells were isolated. Genomic DNA was extracted from 28 isolated cells, and sgRNA target sites were amplified using KOD FX Neo (Toyobo). 29 Primer sequences are as follows: 5’ -CCCATGACCGTCTTCACTCG-3’ and 5’ -30 ACATCCCTTCCCCCACTCTC-3’. PCR products were purified and directly sequenced. 31 32 Vector construction 33 cDNA of ECHS1 (NM_00492) WT and VUS were synthetized and inserted into 34 pCDNA3.1 Lifect -EGFP (Addgene 67303) with HindIII and XbaI sites by GENEWIZ/ 35 Azenta. 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint 1 ATP assay 2 HEK293FT WT or ECHS1 knockout cells were seeded in a collagen-coated 96-well plate 3 (354650, Corning, AZ, USA) at 1 × 104 cells/well with growth medium containing 25 mM 4 glucose. For VUS validation, ECHS1 knockout cells were transfected with 20 ng of 5 expression vectors encoding WT or ECHS1 variants. One day after plating or 6 transfection, the medium was replaced with 25 mM glucose or 10 mM galactose medium 7 supplemented with dialyzed 10% fetal bovine serum (04 -011-1A, Biological industries, 8 KibbutzBeit-Haemek, Israel) and L -valine (13046-62, Nacalai Tesque, Kyoto, Japan). 9 Four days after culture in galactose or glucose medium, the ATP content was measured 10 using the CellTiter-Glo® Luminescent Cell Viability Assay kit (Promega) with a VICTOR 11 Nivo multimode microplate reader (PerkinElmer, MA, USA). 12 13 RNA-seq 14 RNA was purified from fibroblasts by the Maxwell RSC simplyRNA Cells Kit and a 15 Maxwell RSC Instrument (Promega). After quality check by Agilent 2100 and Qubit 2.0, 16 the mRNA was enriched using oligo(dT) beads and rRNA removed using the Ribo-Zero 17 kit. The mRNA was fragmented randomly by adding fragmentation buffer; then, cDNA 18 was synthesized by using the mRNA template and random hexamers primers, followed 19 by addition of a custom second-strand synthesis buffer (Illumina), dNTPs, RNase H and 20 DNA polymerase I to initiate second-strand synthesis. Second, after a terminal repair, A 21 ligation, and sequencing adaptor ligation, the double -stranded cDNA library was 22 completed through size selection and PCR enrichment. Sequencing was performed 23 using 150-bp paired-end reads on a NovaSeq6000 (Illumina). Fastq files were aligned to 24 the GRCh38/hg38 genome by STAR. Gene read counts were quantified by STAR 25 quantMode GeneCounts function. The aligned BAM files were loaded into the Integrated 26 Genomics Viewer and visualized using a Sashimi plot for mRNA splicing analysis. 27 28 qRT-PCR 29 RNA was isolated from culture fibroblasts using FastGene RNA Basic Kit. The isolated 30 RNA was reverse -transcribed to cDNA using ReverTra Ace® qPCR RT Master Mix 31 (TOYOBO) according to manufacturer’s instructions. The synthesized cDNA was used 32 as a template for qRT -PCR in a 7500 Fast Real -Time PCR System (Thermo Fisher 33 Scientific) using THUNDERBIRD® Probe One -step qRT-PCR Kit (TOYOBO). Primer 34 sequences are as follows: ECHS1 -F, 5’-GTCTTCAGGGCCTGGTTGAG-3’, ECHS1-R, 35 5’-CTGTGCAAACTGGGCCTTCT-3’, ACTB -F, 5’ -GCGAGAAGATGACCCAGATC-3’, 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint ACTB-R, 5’-GGATAGCACAGCCTGGATAG-3’. 1 2 Sanger sequencing 3 The ECHS1 variants of patients and family members were sequenced by Sanger 4 sequencing. PCR products were direc tly sequenced using BigDye v3.1 Terminators 5 (Applied Biosystem, ThemoFisher Scientific) or SupreDye v3.1 reagent (Edge 6 BioSystems) and ABI 3130XL (Applied Biosystems, ThemoFisher Scientific). 7 8 Proteome 9 Samples from fibroblasts were prepared as described previously (Borna et al, 2019). The 10 samples were measured in both data -dependent and data -independent modes 11 performed on the Q -Exactive Plus mass spectrometer (Thermo Fisher Scientific) as 12 previously described (Borna et al, 2019). Fin ally, 5,979 proteins were detected in 16 13 samples including two healthy controls and 14 mitochondrial disease patients. Then, 14 outlier protein expression analysis was performed using OUTRIDER (Brechtmann et al, 15 2018). 16 17 Immunoblotting analysis 18 SDS–PAGE and we stern blot were performed as previously described (Kohda et al , 19 2016). To isolate mitochondria, cell pellets were suspended in mitochondria isolation 20 buffer A (220 mM mannitol, 20 mM HEPES, 70 mM sucrose, 1 mM EDTA, pH 7.4, 2 21 mg/mL bovine serum albumin, 1× protease inhibitor cocktail) and homogenized with 20 22 strokes on ice. Homogenates were separated into cytosolic and nuclear fractions after 23 centrifugation at 700 g for 5 min at 4°C. The supernatants were centrifuged at 10,000 g 24 for 10 min at 4°C. Mitochondrial pellets were rinsed twice with mitochondria isola tion 25 buffer B (220 mM mannitol, 20 mM HEPES, 70 mM sucrose, 1 mM EDTA, pH 7.4, 1× 26 protease inhibitor cocktail). Then, mitochondrial protein levels were determined using a 27 bicinchoninic acid assay. For SDS –PAGE analyses, enriched mitochondria were 28 solubilized in RIPA buffer (Nacalai Tesque, Japan) and denatured for 5 min at 95°C. 29 Prepared samples were separated by electrophoresis on 10% or 15% SDS–PAGE gels, 30 depending on the size of the detected protein. Each antibody was obtained as follows; 31 ECHS1 (11305-1-AP , Proteintech, IL, USA), GAPDH (G9545, Sigma-Aldrich), beta-actin 32 (A5441, Sigma-Aldrich). 33 34 Ethics statement 35 The studies were approved by the regional Ethics Committees at Juntendo University, 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Saitama Medical University, and Chiba Children’s Hospital, Kin dai University. We 1 obtained written informed consent from the parents. All methods were performed in 2 accordance with relevant guidelines and regulations. 3 4 Statistics 5 Data are expressed as the mean  ±  SEM. The statistical significance of differences was 6 determined by one -way ANOVA followed by Dunnet’s test using Prism 9 ( GraphPad 7 Software Inc., CA, USA). 8 9 10 Data availability 11 Raw data are available from the corresponding author upon reasonable request. ECHS1 12 knockout cells can also be distributed. Some genomic information that could be used to 13 identify individuals cannot be shared due to ethical restrictions. 14 15 Acknowledgments 16 We thank the family for their participation in the research presented here and the 17 Laboratory of Molecular and Biochemical Research, Biomedical Research Core 18 Facilities, Juntendo University Graduate School of Medicine, and Kasumi Kanai for 19 technical assistance. This work was supported by a grant for the Practical Research 20 Project for Rare/Intractable Diseases from AMED to H.O., K.M., Y .O. and A.O. (Fund ID: 21 JP21im0210625, JP21ek0109511, JP22ek0109485, JP22ek0109468, JP22gk0110038, 22 JP19ek0109273), Program for Promoting Platform of Genomics based Drug Discovery 23 to Y.O. (Fund ID: JP22kk0305015), the Research Center Network for Realization of 24 Regenerative Medicine (The Acceleration Program for Intractable Diseases Research 25 utilizing Disease-specific iPS cells, JP21bm0804018), and JSPS KAKENHI JP19H03624 26 to Y .O. and JP20H03648 to H.O. 27 28 Author contributions 29 YK, AS and YO wrote the manuscript. YK, AS, TK, TE, TM, MS, NI, YN, HN, YY and YW 30 performed the experiments. YK, AS, TK, TE, TM, MS, NI, YN, HN, YY , AIO, YW and YO 31 analyzed the data. TK, TE, TM, MS, TF, HO, AO and KM acquired clinical information. 32 YK, AS, TF, KRN and AIO did bioinformatics and statistical analysis. YK, AS and AO 33 supervised the study. All authors discussed the results an d commented on the 34 manuscript. 35 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Disclosure and competing interests statement 1 The authors declare no competing interests. 2 3 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint

1 Borna NN, Kishita Y, Kohda M, Lim SC, Shimura M, Wu Y, Mogushi K, Yatsuka Y, 2 Harashima H, Hisatomi Y, et al (2019) Mitochondrial ribosomal protein PTCD3 3 mutations cause oxidative phosphorylation defects with Leigh syndrome. 4 Neurogenetics 20: 9–25 5 Brechtmann F, Mertes C, Matusevičiūtė A, Yépez VA, Avsec Ž, Herzog M, Bader DM, 6 Prokisch H & Gagneur J (2018) OUTRIDER: A Statistical Method for Detecting 7 Aberrantly Expressed Genes in RNA Sequencing Data. Am J Hum Genet 103: 8 907–917 9 Burgin HJ & McKenzie M (2020) Understanding the role of OXPHOS dysfunction in the 10 pathogenesis of ECHS1 deficiency. FEBS Lett 594: 590–610 11 D’Annibale OM, Koppes EA, Alodaib AN, Kochersperger C, Karunanidhi A, Mohsen 12 AW & Vockley J (2021) Characterization of variants of uncertain significance in 13 isovaleryl-CoA dehydrogenase identified through newborn screening: An 14 approach for faster analysis. Mol Genet Metab 134: 29–36 15 D’Annibale OM, Koppes EA, Sethuraman M, Bloom K, Mohsen AW & Vockley J (2022) 16 Characterization of exonic variants of uncertain significance in very long-chain 17 acyl-CoA dehydrogenase identified through newborn screening. J Inherit Metab 18 Dis: 1–12 19 Findlay GM, Daza RM, Martin B, Zhang MD, Leith AP, Gasperini M, Janizek JD, Huang 20 X, Starita LM & Shendure J (2018) Accurate classification of BRCA1 variants with 21 saturation genome editing. Nature 562: 217–222 22 Haack TB, Jackson CB, Murayama K, Kremer LS, Schaller A, Kotzaeridou U, de Vries 23 MC, Schottmann G, Santra S, Büchner B, et al (2015a) Deficiency of ECHS1 24 causes mitochondrial encephalopathy with cardiac involvement. Ann Clin Transl 25 Neurol: n/a-n/a 26 Haack TB, Jackson CB, Murayama K, Kremer LS, Schaller A, Kotzaeridou U, de Vries 27 MC, Schottmann G, Santra S, Büchner B, et al (2015b) Deficiency of ECHS1 28 causes mitochondrial encephalopathy with cardiac involvement. Ann Clin Transl 29 Neurol: 492–509 30 Kohda M, Tokuzawa Y, Kishita Y, Nyuzuki H, Moriyama Y, Mizuno Y, Hirata T, Yatsuka 31 Y, Yamashita-Sugahara Y, Nakachi Y, et al (2016) A Comprehensive Genomic 32 Analysis Reveals the Genetic Landscape of Mitochondrial Respiratory Chain 33 Complex Deficiencies. PLoS Genet 12: 1–31 34 Kuwajima M, Kojima K, Osaka H, Hamada Y, Jimbo E, Watanabe M, Aoki S, Sato-35 Shirai I, Ichimoto K, Fushimi T, et al (2021) Valine metabolites analysis in ECHS1 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint deficiency. Mol Genet Metab Reports 29 1 Kweon J, Jang AH, Shin HR, See JE, Lee W, Lee JW, Chang S, Kim K & Kim Y (2020) 2 A CRISPR-based base-editing screen for the functional assessment of BRCA1 3 variants. Oncogene 39: 30–35 4 Lee JS, Yoo T, Lee M, Lee Y, Jeon E, Kim SY, Lim BC, Kim KJ, Choi M & Chae JH 5 (2020) Genetic heterogeneity in Leigh syndrome: Highlighting treatable and novel 6 genetic causes. Clin Genet 97: 586–594 7 Naito Y, Hino K, Bono H & Ui-Tei K (2015) CRISPRdirect: Software for designing 8 CRISPR/Cas guide RNA with reduced off-target sites. Bioinformatics 31: 1120–9 1123 10 Ogawa E, Fushimi T, Ogawa-Tominaga M, Shimura M, Tajika M, Ichimoto K, 11 Matsunaga A, Tsuruoka T, Ishige M, Fuchigami T, et al (2020) Mortality of 12 Japanese patients with Leigh syndrome: Effects of age at onset and genetic 13 diagnosis. J Inherit Metab Dis 43: 819–826 14 Ogawa E, Shimura M, Fushimi T, Tajika M, Ichimoto K, Matsunaga A, Tsuruoka T, 15 Ishige M, Fuchigami T, Yamazaki T, et al (2017) Clinical validity of biochemical 16 and molecular analysis in diagnosing Leigh syndrome: a study of 106 Japanese 17 patients. J Inherit Metab Dis 40: 685–693 18 Peters H, Buck N, Wanders R, Ruiter J, Waterham H, Koster J, Yaplito-Lee J, 19 Ferdinandusse S & Pitt J (2014) ECHS1 mutations in Leigh disease: A new inborn 20 error of metabolism affecting valine metabolism. Brain 137: 2903–2908 21 Peters H, Ferdinandusse S, Ruiter JP, Wanders RJA, Boneh A & Pitt J (2015) 22 Metabolite studies in HIBCH and ECHS1 defects: Implications for screening. Mol 23 Genet Metab 115: 168–173 24 Ran FA, Hsu PD, Wright J, Agarwala V, Scott DA & Zhang F (2013) Genome 25 engineering using the CRISPR-Cas9 system. Nat Protoc 8: 2281–2308 26 Richards S, Aziz N, Bale S, Bick D, Das S, Gastier-Foster J, Grody WW, Hegde M, 27 Lyon E, Spector E, et al (2015) Standards and guidelines for the interpretation of 28 sequence variants: A joint consensus recommendation of the American College of 29 Medical Genetics and Genomics and the Association for Molecular Pathology. 30 Genet Med 17: 405–424 31 Robinson BH, Petrova-Benedict R, Buncic JR & Wallace DC (1992) Nonviability of cells 32 with oxidative defects in galactose medium: A screening test for affected patient 33 fibroblasts. Biochem Med Metab Biol 48: 122–126 34 Sakai C, Yamaguchi S, Sasaki M, Miyamoto Y, Matsushima Y & Goto Y ichi (2015) 35 ECHS1 mutations cause combined respiratory chain deficiency resulting in leigh 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint syndrome. Hum Mutat 36: 232–239 1 Sato-Shirai I, Ogawa E, Arisaka A, Osaka H, Murayama K, Kuwajima M, Watanabe M, 2 Ichimoto K, Ohtake A & Kumada S (2021) Valine-restricted diet for patients with 3 ECHS1 deficiency: Divergent clinical outcomes in two Japanese siblings. Brain 4 Dev 43: 308–313 5 Sharpe AJ & McKenzie M (2018) Mitochondrial fatty acid oxidation disorders 6 associated with short-chain enoyl-CoA hydratase (ECHS1) deficiency. Cells 7: 1–7 13 8 Simon MT, Eftekharian SS, Ferdinandusse S, Tang S, Naseri T, Reupena MS, 9 McGarvey ST, Minster RL, Weeks DE, Nguyen DD, et al (2021) ECHS1 disease 10 in two unrelated families of Samoan descent: Common variant - rare disorder. Am 11 J Med Genet Part A 185: 157–167 12 Stenton SL, Zou Y, Cheng H, Liu Z, Wang J, Shen D, Jin H, Ding C, Tang X, Sun S, et 13 al (2022) Leigh Syndrome: A Study of 209 Patients at the Beijing Children’s 14 Hospital 15 Tadaka S, Hishinuma E, Komaki S, Motoike IN, Kawashima J, Saigusa D, Inoue J, 16 Takayama J, Okamura Y, Aoki Y, et al (2021) jMorp updates in 2020: Large 17 enhancement of multi-omics data resources on the general Japanese population. 18 Nucleic Acids Res 49: D536–D544 19 Tadaka S, Katsuoka F, Ueki M, Kojima K, Makino S, Saito S, Otsuki A, Gocho C, 20 Sakurai-Yageta M, Danjoh I, et al (2019) 3.5KJPNv2: an allele frequency panel of 21 3552 Japanese individuals including the X chromosome. Hum Genome Var 6 22 Tetreault M, Fahiminiya S, Antonicka H, Mitchell GA, Geraghty MT, Lines M, Boycott 23 KM, Shoubridge EA, Mitchell JJ, Michaud JL, et al (2015) Whole-exome 24 sequencing identifies novel ECHS1 mutations in Leigh syndrome. Hum Genet 25 134: 981–991 26 Uchino S, Iida A, Sato A, Ishikawa K, Mimaki M, Nishino I & Goto Y ichi (2019) A novel 27 compound heterozygous variant of ECHS1 identified in a Japanese patient with 28 Leigh syndrome. Hum Genome Var 6: 6–9 29 Vaca Jacome AS, Rabilloud T, Schaeffer-Reiss C, Rompais M, Ayoub D, Lane L, 30 Bairoch A, Van Dorsselaer A & Carapito C (2015) N-terminome analysis of the 31 human mitochondrial proteome. Proteomics 15: 2519–2524 32 Xia C, Lou B, Fu Z, Mohsen AW, Shen AL, Vockley J & Kim JJP (2021) Molecular 33 mechanism of interactions between ACAD9 and binding partners in mitochondrial 34 respiratory complex I assembly. iScience 24: 103153 35 Yamada K, Aiba K, Kitaura Y, Kondo Y, Nomura N, Nakamura Y, Fukushi D, 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Murayama K, Shimomura Y, Pitt J, et al (2015) Clinical, biochemical and 1 metabolic characterisation of a mild form of human short-chain enoyl-CoA 2 hydratase deficiency: Significance of increased n-acetyl-s-(2-3 carboxypropyl)cysteine excretion. J Med Genet 52: 691–698 4 Yang H & Yu D (2020) Clinical, biochemical and metabolic characterization of patients 5 with short-chain enoyl-CoA hydratase(ECHS1) deficiency: Two case reports and 6 the review of the literature. BMC Pediatr 20: 1–10 7 8 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Figures 1 2 Figure 1. Summary of ECHS1 variants 3 Gene structure (top) and corresponding amino acids (bottom) of ECHS1. Variants 4 registered r as pathogenic (red) and likely pathogenic (yellow) in ClinVa r, and jMorp 5 (square) are shown. The two underlined variants were identified from our genomic 6 analysis study; no experimental VUS verification was performed. 7 8 Figure 2. Characterization of ECHS1 KO cells. 9 A. Genomic analysis of ECHS1 KO cells by Sanger sequencing. 10 B. EHCS1 expression levels in WT and ECHS1 KO HEK293FT cells quantified by qRT-11 PCR. Whole cell extracts from WT and ECHS1 KO HEK293FT cells analyzed by 12 immunoblotting using the indicated antibodies. Single and double asterisks indicate 13 uncleaved and cleaved forms, respectively. 14 C. ATP assay of WT and ECHS1 KO HEK293FT cells cultured in glucose or galactose 15 medium. The graph shows the ATP level in galactose medium divided by galactose. 16 D. ATP assay of WT and ECHS1 KO HEK293FT cells treated with the indicated L-valine 17 concentrations. Bar graphs represent the average ATP level in each condition from three 18 biological independent experiments. Error bars, ±SEM. Statistical analysis was 19 performed using ANOVA followed by Dunnet’s test. RLU, relative luciferase unit. 20 21 Figure 3. Validation of ECHS1 VUS. 22 WT and ECHS1 KO HEK293FT cells transfected with the indicated expression vectors 23 subjected to immunoblotting analysis (A) and ATP assay (B). 24 A. Whole cell extracts were analyzed by immunoblotting using the indicated antibodies. 25 Single and double asterisks indicate uncleaved and cleaved forms, respectively. 26 B. ATP assay 4 days after treatment with L -valine (0.8 mM). Bar graphs represent the 27 average ATP level in each condition from three biological independent experiments. Error 28 bars, ±SEM. Red bars, pathogenic variants. Gray bars, VUS. RLU, relative luciferase 29 unit. 30 C. Statistical analysis of Figure B using ANOVA followed by Dunnet’s test. The color 31 scale shows the P value compared with vector (vs Vec, statistically different red to blue) 32 and WT ECHS1 (vs WT, statistically different blue to red). 33 34 Figure 4. Validation of ECHS1 VUS compound hetero. 35 A. Whole cell extracts from WT and ECHS1 KO HEK293FT cells transfected with the 36 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint indicated expression vectors analyzed by immunoblotting using the indicated antibodies. 1 B. ATP a ssay of ECHS1 KO HEK293FT cells expressing the indicated expression 2 vectors treated with additional L-valine (0.8 mM) for four days. Bar graphs represent the 3 average ATP level in each condition from three biological independent experiments. Error 4 bars, SEM. RLU, relative luciferase unit. 5 C. Statistical analysis of Figure B using ANOVA followed by Dunnet’s test. The color 6 scale shows the P value compared with vector (vs Vec, statistically different red to blue) 7 and WT ECHS1 (vs WT, statistically different blue to red). 8 9 Figure 5. Abnormal splicing and allele-biased gene expression 10 A. RNA-seq data from two patients with c.489G>A (p.P163=) showing reads suggestive 11 of splicing abnormalities in exons with c.489G>A. 12 B. Read counts of the ECHS1 gene are plotted on the horizontal axis and the number of 13 detected exon skipping is plotted on the vertical axis. Gene counts were calculated by 14 STAR, and the number of exon skipping was extracted from the Sashimi plot data. 15 C. Ratio of c.489G>A and c.796A>C and c.832G>A var iants on the IGV viewer. In two 16 cases, the allele expression with c.489G>A was decreased. 17 18 Figure 6. ECHS1 protein status in the patient with Pro163= and Ala278Thr 19 A. OUTRIDER analysis illustrating protein expression in a volcano plot. ECHS1 was 20 detected as a protein with a large decrease in expression. 21 B. Western blotting for ECHS1 in patients with ECHS1 mutations and controls. 22 Compared to previously reported ECHS1 cases, case 1 showed significantly reduced 23 ECHS1 expression. β-actin was detected as a loading control. 24 25 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Tables 1 2 Patient ID DNA (NM_004092.4) Protein(NP_004083.3) 0207 c.5C>T p.Ala2Val 0207 c.176A>G p.Asn59Ser 0255 c.176A>G p.Asn59Ser 0255 c.413C>T p.Ala138Val 0346ES c.176A>G p.Asn59Ser 0346ES c.476A>G p.Gln159Arg 0346YS c.176A>G p.Asn59Ser 0346YS c.476A>G p.Gln159Arg 0376 c.98T>C p.Phe33Ser 0376 c.176A>G p.Asn59Ser 0536 c.1A>G p.Met1Val 0536 c.5C>T p.Ala2Val 0775 c.5C>T p.Ala2Val 0775 c.88+2T>C 1038ES c.5C>T p.Ala2Val 1038ES c.176A>G p.Asn59Ser 1038YS c.5C>T p.Ala2Val 1038YS c.176A>G p.Asn59Ser 1135ES c.5C>T p.Ala2Val 1135ES c.176A>G p.Asn59Ser 1135YS c.5C>T p.Ala2Val 1135YS c.176A>G p.Asn59Ser 1553 c.5C>T p.Ala2Val 1553 c.176A>G p.Asn59Ser 2521 c.23T>C p.Leu8Pro 2521 c.176A>G p.Asn59Ser 2637 c.5C>T p.Ala2Val 2637 c.176A>G p.Asn59Ser 2816 c.23T>C p.Leu8Pro 2816 c.176A>G p.Asn59Ser Table 1. ECHS1 variants identified from genomic analysis 3 4 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Case Sex Onset Symptoms Complex deficiency OCR Var 1 Var 2 1 female > 1 y poor suckling, metabolic acidosis CIV 26.1% (Fb) 68% (Fb) c.796A>C (p.Thr266Pro) c.489G>A (p.Pro163=) 2 male ≤ 1 y gait disorder, nystagmus, MR, MRI abnormality Normal (Fb) N.T. c.802G>A (p.Ala268Thr) c.489G>A (p.Pro163=) 3 male ≤ 1 y listlessness, mental retardation, regression after exanthema subitum, MRI abnormality, deafness, fatigue, hypotonia Normal (Fb) 36% (Fb) c.832G>A (p.Ala278Thr) c.489G>A (p.Pro163=) 4 male > 1 y developmental delay, convulsion, MRI abnormalities, regression Normal (Fb) 54% (Fb) c.832G>A (p.Ala278Thr) c.489G>A (p.Pro163=) 5 male > 1 y Leigh syndrome, increase of 2-methyl- 2,3-dihydroxybutyric acid Normal (Fb) 58% (Fb) c.833C>T (p.Ala278Val) c.489G>A (p.Pro163=) 1 Table 2. Patient summary of cases with ECHS1 variants 2 Fb: Fibroblast, CIV: Mitochondrial respiratory chain complex IV, OCR: oxygen 3 consumption rates. Complex enzyme activity was defined by <40% decrease. For OCR, 4 a value <71.6% was used as diagnostic criterion. 5 6 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint c.1A>G c.5C>T c.98T>C p.Met1Val p.Ala2Val p.Phe33Ser c.23T>C p.Leu8Pro c.88+2T>C c.176A>G p.Asn59Ser c.413C>T p.Ala138Val c.476A>G p.Gln159Arg c.488C>T p.Pro163Leu c.740-3T>A p.Met1Thr c.2T>C p.Asp63Asn c.187G>A p.Lys74Glu c.220A>G c.357C>G p.His119Gln c.434T>C p.Leu145Pro c.802G>A p.Ala268Thr p.Arg272Gln c.815G>A c.832G>A c.796A>C p.Thr266Pro p.Ala278Thr jMorpClinVar Pathogenic ClinVar Likely pathogenic Figure 1 Not verified in this study c.489G>A p.Pro163= . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint KO WT - 40 - 35 - 30 ECHS1 GAPDH A B 1.0 0.5 0 RLU relative to UT ns ns ns ns ns ns WT KO Additional Val [mM] - 0.2 0.4 0.8 - 0.2 0.4 0.8 Glucose RLU relative to UT WT KO Additional Val [mM] - 0.2 0.4 0.8 - 0.2 0.4 0.8 1.0 0.5 0 ns ns ns ns p=0.0197 p=0.0017 Galactose D GAACCGCCCCAAGGCCCT NRPKAL ECHS1 NM_004092.3 NP_004083.3 RLU Gal/Glc 0 0.8 0.6 0.4 0.2 p=0.1246 KO WT C c.156_173del p.Arg54_Asn59del WT KO 0.4 0.2 0 0.6 0.8 1.0 1.2ECHS1/ACTB Figure 2 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint A B C Deleterious Benign Vec WT M1V M1T A2V L8P F33S N59S D63N K74E H119Q L145P P163L T266P A268T R272Q A278T vs Vec - 0.0015 0.6909 0.9888 0.9008 0.2205 0.0089 0.9997 0.0246 <0.0001 0.6735 0.9309 0.4297 0.0077 0.4514 0.0583 0.9953 vs WT 0.0015 - 0.0762 0.0158 0.0351 0.3391 0.999 0.0027 0.956 0.4807 0.0803 <0.0001 0.1651 0.9991 0.1547 0.7744 0.0002 WT Cell ECHS1 KO cells Vec M1V M1T A2V L8P N59S D63N K74E H119Q L145P A268T R272Q A278T WT F33S P163L T266P ECHS1 b-actin -35 -30 -55 -45 ***Vec M1V M1T A2V L8P N59S D63N K74E H119Q L145P A268T R272Q A278T WT F33S P163L T266P 1.0 0.5 0 ATP level RLU relative to WT Figure 3 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint WT Cell Vec WT ECHS1 KO cellsA2V L8P N59S F33S Vec WT A2V L8P F33S N59S ECHS1 b-actin -35 -30 -55 -45 Vec WT ECHS1 KO cells A2V L8P N59S F33S Vec WT A2V L8P F33S N59S A B C Vec WT A2V L8P F33S N59S N59S+ Vec N59S+ WT N59S+ A2V N59S+ L8P N59S+ F33S vs Vec 0.0001 0.2601 0.0031 0.0001 0.6689 0.9541 0.0003 0.3122 0.1256 0.0639 vs WT 0.0001 0.0008 0.0916 0.9895 0.0002 0.0001 0.4773 0.0006 0.0021 0.0047 ATP level RLU relative to WT 1.0 0.5 0 Deleterious Benign Figure 4 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint Case 1 Case 4 Other patient A C Case 1 Case 4 c.832G>A(g.133,362,909C>T) c.796A>C(q.133,364,669T>G) c.489G>A(g.133,368,948C>T) Figure 5 c.489G>A(g.133,368,948C>T) 0 2 4 6 8 10 12 14 1500 2500 3500 4500 Comp het Het No variant ECHS1 counts ECHS1 exon skipping B Case 1 Case 4 . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint A B Control-1 ECHS1 β-actin Control-2 1 2 Case 1 ECHS1 Comp het ECHS1 het ECHS1 Figure 6 Case 2 -log10(p-value) zScore . CC-BY-NC-ND 4.0 International licenseIt is made available under a perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in(which was not certified by peer review)preprint The copyright holder for thisthis version posted October 13, 2022. ; https://doi.org/10.1101/2022.10.09.22280834doi: medRxiv preprint