Label-Free Quantitative Proteomics in a Methylmalonyl-CoA Mutase-Silenced Neuroblastoma Cell Line

Methylmalonic acidemias (MMAs) are inborn errors of metabolism due to the deficient activity of methylmalonyl-CoA mutase (MUT). MUT catalyzes the formation of succinyl-CoA from methylmalonyl-CoA, produced from propionyl-CoA catabolism and derived from odd chain fatty acids β-oxidation, cholesterol, and branched-chain amino acids degradation. Increased methylmalonyl-CoA levels allow for the presymptomatic diagnosis of the disease, even though no approved therapies exist. MMA patients show hyperammonemia, ketoacidosis, lethargy, respiratory distress, cognitive impairment, and hepatomegaly. The long-term consequences concern neurologic damage and terminal kidney failure, with little chance of survival. The cellular pathways affected by MUT deficiency were investigated using a quantitative proteomics approach on a cellular model of MUT knockdown. Currently, a consistent reduction of the MUT protein expression was obtained in the neuroblastoma cell line (SH-SY5Y) by using small-interfering RNA (siRNA) directed against an MUT transcript (MUT siRNA). The MUT absence did not affect the cell viability and apoptotic process in SH-SY5Y. In the present study, we evaluate and quantify the alterations in the protein expression profile as a consequence of MUT-silencing by a mass spectrometry-based label-free quantitative analysis, using two different quantitative strategies. Both quantitative methods allowed us to observe that the expression of the proteins involved in mitochondrial oxido-reductive homeostasis balance was affected by MUT deficiency. The alterated functional mitochondrial activity was observed in siRNA_MUT cells cultured with a propionate-supplemented medium. Finally, alterations in the levels of proteins involved in the metabolic pathways, like carbohydrate metabolism and lipid metabolism, were found.


Introduction
Hereditary methylmalonic acidemias (MMAs) are severe autosomal recessive inborn errors of intermediary metabolism caused by the deficient activity of methylmalonyl-CoA mutase (MUT) or defects in the synthesis of 5-deoxyadenosyl cobalamin, the active form of vitamin B12 and the essential cofactor of MUT. MUT converts methylmalonyl-CoA into succinyl-CoA, a Krebs cycle intermediate. Methylmalonyl-CoA is produced from the catabolism of propionyl-CoA, derived from the degradation of cholesterol, branched-chain amino acids (valine, isoleucine, methionine, and threonine), and odd chain fatty acids β-oxidation [1]. The defect in the MUT protein causes an increase in the level of methylmalonyl-CoA, which is then converted into methylmalonic acid (MMA). Hereditary MMAs are included in newborn screening panels in several countries [2][3][4][5], allowing for a presymptomatic diagnosis of the disease. To this aim, targeted mass spectrometry-based metabolomics is a powerful tool to profile amino acids and acylcarnitines in a quantitative manner [6][7][8].
Isolated MMAs are caused by a complete (mut 0 ) or partial (mut − ) loss of MUT activity [9]. At the moment, no approved therapies exist for isolated MMAs. Patients are treated with dietary protein restriction and cofactor supplementation. Despite these treatments, the mortality carries on being around 20%, and the disease progression is characterized by acute metabolic instability, especially for mut 0 patients [10]. Methylmalonic acidemia patients show hyperammonemia, ketoacidosis, lethargy, respiratory distress, cognitive impairment, and hepatomegaly. The long-term consequences concern neurologic damage and terminal kidney failure, with little chance of survival. The metabolic instability, disability, and death rate are reduced as a result of liver and/or combined liver/kidney transplantation [11,12]. However, transplantation utility is limited by the low number of liver donors, significant surgery risks, and high procedural costs [13]. Thus, efforts are devoted to develop new therapies as an alternative to transplantation [14].
Despite the elevated levels of MMA, biological fluids are used as a hallmark of the pathology, and its accumulation may account for multisystemic pathological dysfunction (especially at a neuronal, hepatic, and renal level), the molecular mechanisms underlying the damage induced by MMA are not fully detailed.
In order to investigate the cellular pathways altered downstream by MUT deficiency, we used small-interfering RNA to knockdown the MUT protein expression in a neuroblastoma cell line, (namely SH-SY5Y). Label-free quantitative proteomics [15,16] were used to identify the proteins whose levels were found to be deregulated after MUT enzyme knockdown. We found deregulation in the levels of mitochondrial proteins, such as electron transfer flavoprotein subunit alpha and 2-oxoglutarate carrier, crucially involved in the mitochondrial oxido-reductive homeostasis balance. In addition, we also observed significant differences in the level of proteins enrolled in the metabolic pathways, such as carbohydrate metabolism (Gamma-enolase and fructose-bisphosphate aldolase C) and lipid metabolism (sphingomyelin phosphodiesterase 4 and sulfatase-modifying factor 2). The cellular pathways altered upon the MUT enzyme reduction may represent putative therapeutic targets, which should be possibly taken into account for the design of new therapies to alleviate patients' clinical manifestations.

MUT Silencing
The human SH-SY5Y cell line was transfected using a specific siRNA to reduce the expression of MUT protein (siRNA_MUT). A siRNA molecule unable to target known cellular transcripts was used as the negative control (Scramble). The silencing was evaluated 24 and 48 h after siRNA transfection. As shown in Figure 1, the MUT protein expression was reduced by about 50% after 24 h and by about 70% (p < 0.001) after 48 h. The 48-h time point was chosen for the following experiments. MUT silencing was evaluated 24 and 48 h after siRNA transfection by Western blot analysis using MUT specific antibodies. The silencing was carried out in three independent experiments at 24 and 48 h (A). The MUT optical density was measured and normalized by β-actin protein signal pixels (B). The results are reported as the mean ± standard deviation (SD). Statistical significance was calculated by one-way two tail paired t-test. p-values are indicated as follows: NS = non significant = p > 0.05; *** = p < 0.005.

Cell Survival and Apoptosis
The apoptosis rate was measured in MUT silenced SH-SY5Y (siRNA_MUT) 48 h after transient transfection. The cells transfected with scramble siRNA (Scramble) and untransfected cells have been used as the controls. The cells were stained with Annexin V and propidium iodide (PI), and analyzed by flow cytometry in order to evaluate the possible differences in the apoptotic rates. Indeed, the results revealed a very low percentage of cells with a high Annexin V signal and low PI signal (cells in early apoptosis), with no significant difference between the siRNA_MUT and Scramble cells (Figure 2). The percentage of healthy cells with both low (Annexin V and propidium iodide) signals was unaffected by the MUT silencing, as well as the percentage of cells with both high signals (representing cells in necrotic or late apoptotic state), which similarly showed no significant variation. A very low percentage of cells with high Annexin V and low PI signal (cells in early apoptosis) was present in all of the samples. This latter observation may indicate that the cells with both high signals were probably necrotic with the absence of apoptotic processes. In the examined temporal window, MUT silencing slightly affected cell viability without modifying the apoptotic rate, if compared with the Scramble siRNA transfection. In order to provide a quantitative estimation of the number of viable cells in the culture, a neutral-red uptake assay [17] was performed (Supplemental Figure S1) 48 h after transfection. Differences in the Scramble and siRNA_MUT cell viabilities were not observed. Moreover, the cell viability was comparable to the control untransfected cells. MUT silencing was evaluated 24 and 48 h after siRNA transfection by Western blot analysis using MUT specific antibodies. The silencing was carried out in three independent experiments at 24 and 48 h (A). The MUT optical density was measured and normalized by β-actin protein signal pixels (B). The results are reported as the mean ± standard deviation (SD). Statistical significance was calculated by one-way two tail paired t-test. p-values are indicated as follows: NS = non significant = p > 0.05; *** = p < 0.005.

Cell Survival and Apoptosis
The apoptosis rate was measured in MUT silenced SH-SY5Y (siRNA_MUT) 48 h after transient transfection. The cells transfected with scramble siRNA (Scramble) and untransfected cells have been used as the controls. The cells were stained with Annexin V and propidium iodide (PI), and analyzed by flow cytometry in order to evaluate the possible differences in the apoptotic rates. Indeed, the results revealed a very low percentage of cells with a high Annexin V signal and low PI signal (cells in early apoptosis), with no significant difference between the siRNA_MUT and Scramble cells ( Figure 2). The percentage of healthy cells with both low (Annexin V and propidium iodide) signals was unaffected by the MUT silencing, as well as the percentage of cells with both high signals (representing cells in necrotic or late apoptotic state), which similarly showed no significant variation. A very low percentage of cells with high Annexin V and low PI signal (cells in early apoptosis) was present in all of the samples. This latter observation may indicate that the cells with both high signals were probably necrotic with the absence of apoptotic processes. In the examined temporal window, MUT silencing slightly affected cell viability without modifying the apoptotic rate, if compared with the Scramble siRNA transfection. In order to provide a quantitative estimation of the number of viable cells in the culture, a neutral-red uptake assay [17] was performed (Supplemental Figure S1) 48 h after transfection. Differences in the Scramble and siRNA_MUT cell viabilities were not observed. Moreover, the cell viability was comparable to the control untransfected cells.

Proteomic Profiles
A quantitative proteomic analysis was performed using the human SH-SY5Y cell line, in which the MUT expression was reduced 48 h after transient transfection with siRNA against MUT. The cells transfected with scramble siRNA and harvested at the same time point (48 h) have been chosen the as proteomic experiment control. Cellular proteomes were resolved on a 10% Sodium Dodecyl Sulphate (SDS)-polyacrylamide gel ( Figure 3). Each gel lane was fractionated in order to obtain 40 fractions, which were cut and properly processed for protein identification by nanoLC-MS/MS [18,19]. The protein species

Proteomic Profiles
A quantitative proteomic analysis was performed using the human SH-SY5Y cell line, in which the MUT expression was reduced 48 h after transient transfection with siRNA against MUT. The cells transfected with scramble siRNA and harvested at the same time point (48 h) have been chosen the as proteomic experiment control. Cellular proteomes were resolved on a 10% Sodium Dodecyl Sulphate (SDS)-polyacrylamide gel ( Figure 3). Each gel lane was fractionated in order to obtain 40 fractions, which were cut and properly processed for protein identification by nanoLC-MS/MS [18,19]. The protein species identified by more than three peptides were taken into account and included in our proteomic dataset. The resulted proteomic dataset was constituted by about 1000 proteins in both siRNA_MUT and Scramble cells. The details of the protein identification are reported in Supplemental Table S1.
Label-free proteomic analysis was performed to estimate the relative abundance of each protein by two spectral counting parameters, R SC and Fold NSAF [20]. The spectral counting parameters, R SC and Fold NSAF , were correlated by the Pearson correlation coefficient r = 0.9788, R 2 = 0.9581, p < 0.0001 (Supplemental Figure S2 and Table S2). identified by more than three peptides were taken into account and included in our proteomic dataset. The resulted proteomic dataset was constituted by about 1000 proteins in both siRNA_MUT and Scramble cells. The details of the protein identification are reported in Supplemental Table S1. Label-free proteomic analysis was performed to estimate the relative abundance of each protein by two spectral counting parameters, RSC and FoldNSAF [20]. The spectral counting parameters, RSC and FoldNSAF, were correlated by the Pearson correlation coefficient r = 0.9788, R 2 = 0.9581, p < 0.0001 (Supplemental Figure S2 and Table S2). According to the RSC and FoldNSAF values, the proteins were accepted to be deregulated when both of the following conditions occurred: RSC > +3.5 or <−3.5; FoldNSAF > +3.5 or <−3.5. RSC and FoldNSAF values referring to the differentially expressed proteins are reported in Table 1. The two different quantitative analytical indices identified 57 more abundant and 56 less abundant proteins in the silenced cells (siRNA_MUT).  According to the R SC and Fold NSAF values, the proteins were accepted to be deregulated when both of the following conditions occurred: R SC > +3.5 or <−3.5; Fold NSAF > +3.5 or <−3.5. R SC and Fold NSAF values referring to the differentially expressed proteins are reported in Table 1. The two different quantitative analytical indices identified 57 more abundant and 56 less abundant proteins in the silenced cells (siRNA_MUT).

Functional and Biological Annotation
In order to elucidate the implications of the differentially expressed proteins in the cellular processes subsequent to MUT silencing, we analyzed the identified proteins using the Protein Analysis Through Evolutionary Relationship (PANTHER) and Reactome databases. The PANTHER enrichment analysis allowed to cluster 92/118 deregulated hits according to their molecular function, as reported in Figure 4A. The following three main categories were identified: 41 hits as binding proteins (34.7-44.6%), 29 hits as proteins involved in catalytic activity (24.6-31.5%), and 13 hits as the proteins involved in the structural molecular activity (11.0-14.1%). We focused on the proteins involved in catalytic activity. This class was further subdivided into eight enzymatic categories, as reported in Figure 4B and Supplemental Table S3. Within the oxidoreductase activity category, the most interesting proteins are represented by electron transfer flavoprotein subunit alpha, mitochondrial (P13804, ETFA, Rsc = −3.68, and Fold NSAF = −7.04), and Peroxiredoxin-6 (P30041, PRDX6, Rsc = −3.68, and Fold NSAF = −3.73), both found to be less abundant in the siRNA_MUT sample. ETFA is a crucial enzyme involved in mitochondrial fatty acid β-oxidation, shuttling electrons from flavoprotein dehydrogenases, and the membrane-bound ubiquinone oxidoreductase. The impairment of ETFA-mediated processes affects the intracellular acidity. The type II glutaric aciduria is an example of the ETFA defect, characterized by glutaric, lactic, ethylmalonic, butyric, isobutyric, 2-methyl-butyric, and isovaleric acid accumulation. Moreover, PRDX6 is involved in the cell redox homeostasis, playing a protective role against oxidative stress. The PRDX6 down-regulation could affect the short chain fatty acid and phospholipid hydroperoxides reduction. Recently, it was reported that PRDX6 is involved in liver damage induced by oxidative stress [21]. Indeed, a deep linkage exists between methylmalonic acidemia and oxidative metabolism dysfunction [22][23][24]. Previous studies have already shown that increased MMA levels affect mitochondrial morphology and cytochrome c oxidase activity in patients [25,26].
The impairment of ETFA-mediated processes affects the intracellular acidity. The type II glutaric aciduria is an example of the ETFA defect, characterized by glutaric, lactic, ethylmalonic, butyric, isobutyric, 2-methyl-butyric, and isovaleric acid accumulation. Moreover, PRDX6 is involved in the cell redox homeostasis, playing a protective role against oxidative stress. The PRDX6 down-regulation could affect the short chain fatty acid and phospholipid hydroperoxides reduction. Recently, it was reported that PRDX6 is involved in liver damage induced by oxidative stress [21]. Indeed, a deep linkage exists between methylmalonic acidemia and oxidative metabolism dysfunction [22][23][24]. Previous studies have already shown that increased MMA levels affect mitochondrial morphology and cytochrome c oxidase activity in patients [25,26].  The PANTHER enrichment analysis was also performed to define the main biological processes involved in the identified protein dataset ( Figure 5A). The following two main categories were identified: 59 hits as cellular process (29.8-50.0%) and 50 hits as metabolic process (25.3-42.4%). In the metabolic process, seven different subcategories were found and reported in Figure 5B and Supplemental Table S4. Gamma-enolase (P09104, ENOG, Rsc = −3.68, Fold NSAF = −6.16) and fructose bisphosphate aldolase C (P09972, ALDOC, Rsc = −3.68, Fold NSAF = −6.85), involved in energetic metabolism, were under-represented in the siRNA_MUT cells. In our previous proteomic investigation about the proteomic profiles of patients' transplanted liver tissues [27], we also observed a decreased cellular level of proteins involved in the energy metabolism, gluconeogenesis, and Krebs cycle anaplerosis. The deregulation of the glucose metabolism, referred to "glycolysis" (p-value 1.31 × 10 −5 , False Discovery Rate (FDR) 4.46 × 10 −4 ) and "gluconeogenesis" (p-value 6.57 × 10 −3 , FDR 1.31 × 10 −2 ), was also confirmed by the Reactome database (Supplemental Table S5). In particular, M2OM belongs to the mitochondrial carrier protein family, controls the transport of 2-oxoglutarate across the inner mitochondrial membrane, and regulates the malate-aspartate and oxoglutarate-isocitrate shuttles. It also takes part in the nitrogen metabolism. Moreover, M2OM is responsible for glutathione uptake [28], and the glutathione deficiency is a complication of methylmalonic acidemia. Low levels of glutathione affect the cellular oxidative stress. As already mentioned above, evidences show once again a strong connection between methylmalonic acidemia and oxidative metabolism dysfunction [22][23][24]. These results are consistent with our previous studies showing disturbances in the glutathione metabolism in the lymphocytes of patients with cblC defect-associated MMAs [29].
(p-value 1.31 × 10 −5 , False Discovery Rate (FDR) 4.46 × 10 −4 ) and "gluconeogenesis" (p-value 6.57 × 10 −3 , FDR 1.31 × 10 −2 ), was also confirmed by the Reactome database (Supplemental Table S5). In particular, M2OM belongs to the mitochondrial carrier protein family, controls the transport of 2-oxoglutarate across the inner mitochondrial membrane, and regulates the malate-aspartate and oxoglutarate-isocitrate shuttles. It also takes part in the nitrogen metabolism. Moreover, M2OM is responsible for glutathione uptake [28], and the glutathione deficiency is a complication of methylmalonic acidemia. Low levels of glutathione affect the cellular oxidative stress. As already mentioned above, evidences show once again a strong connection between methylmalonic acidemia and oxidative metabolism dysfunction [22][23][24]. These results are consistent with our previous studies showing disturbances in the glutathione metabolism in the lymphocytes of patients with cblC defect-associated MMAs [29].  Reactome also mapped five proteins involved in the "Cellular response to Hypoxia" (five hits; p-value 1.56 × 10 −3 , FDR 4.76 × 10 −3 ) (Supplemental Table S5). Hypoxia has been recently associated with metabolic acidosis and with methylmalonic acidemia [30]. Moreover, the Reactome tool showed four hits for the involvement of proteins in the "lipid metabolism" (four hits; p-value 1.68 × 10 −1 , FDR 1.68 × 10 −1 ) (Supplemental Table S5). Lipid metabolism involves SMPD4/NSMA3 and SUMF2, which belong to the glycosphingolipidic metabolism; SMPD4/NSMA3 is over-represented in the cells not-expressing MUT; it could be a protein of relevant interest, because it catalyzes the hydrolysis of membrane sphingomyelin to form phosphorylcholine and ceramide. A significant myelin content reduction was observed in the cerebrum from rat brains after the administration of MMA [31]. The MMA content may be related to the delayed myelination/cerebral atrophy and neurological dysfunction found in methylmalonic acidemia children. Conversely, the SUMF2 results are down-represented in cells not-expressing MUT; it heterodimerizes with another member of the same protein family, which is characterized by enzymatic activity that is able to generate C-alpha-formylglycine and activate sulfatases after the oxidation of cysteine residues to C-alpha-formylglycine. Multiple sulfatase deficiency (MSD; OMIM 272200) is a rare autosomal recessive inborn error of metabolism caused by mutations in the sulfatase modifying factor 1 gene, resulting in tissue accumulation of sulfatides, sulphated glycosaminoglycans, sphingolipids, and steroid sulfates [32].

MUT Silencing Decreases Cell Viability and Mitochondrial Functionality in Propionate-Enriched Culture Medium
Studies on cells carrying defects of MUT have been performed elsewhere after addition, in the culture medium, of metabolic precursors (e.g., propionate) of methylmalonil-CoA, in order to make more evident pathway unbalances [33]. A propionate-enriched culture medium was used to investigate whether the MUT knock-down could affect the cell viability and mitochondria functionality, using Scramble cells as the control. The siRNA_MUT cells cultured in a propionate-supplemented medium showed a slightly reduced cell viability, but a significant decreased mithocondrial functionality ( Figure 6). The variation between the fold-changes of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and neutral-red assays was calculated as ∆ = 0.26. The reduced mitochondrial functionality, showed by a mitochondrial succinate dehydrogenase-based assay, supports the proteomic results showing the deregulation of a group of mitochondrial proteins. myelin content reduction was observed in the cerebrum from rat brains after the administration of MMA [31]. The MMA content may be related to the delayed myelination/cerebral atrophy and neurological dysfunction found in methylmalonic acidemia children. Conversely, the SUMF2 results are down-represented in cells not-expressing MUT; it heterodimerizes with another member of the same protein family, which is characterized by enzymatic activity that is able to generate C-alpha-formylglycine and activate sulfatases after the oxidation of cysteine residues to C-alpha-formylglycine. Multiple sulfatase deficiency (MSD; OMIM 272200) is a rare autosomal recessive inborn error of metabolism caused by mutations in the sulfatase modifying factor 1 gene, resulting in tissue accumulation of sulfatides, sulphated glycosaminoglycans, sphingolipids, and steroid sulfates [32].

MUT Silencing Decreases Cell Viability and Mitochondrial Functionality in Propionate-Enriched Culture Medium
Studies on cells carrying defects of MUT have been performed elsewhere after addition, in the culture medium, of metabolic precursors (e.g., propionate) of methylmalonil-CoA, in order to make more evident pathway unbalances [33]. A propionate-enriched culture medium was used to investigate whether the MUT knock-down could affect the cell viability and mitochondria functionality, using Scramble cells as the control. The siRNA_MUT cells cultured in a propionate-supplemented medium showed a slightly reduced cell viability, but a significant decreased mithocondrial functionality ( Figure  6). The variation between the fold-changes of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and neutral-red assays was calculated as ∆ = 0.26. The reduced mitochondrial functionality, showed by a mitochondrial succinate dehydrogenase-based assay, supports the proteomic results showing the deregulation of a group of mitochondrial proteins. Figure 6. Cell viability and mitochondrial functionality assays of siRNA_MUT cells in sodium propionate-medium. Neutral-red and MTT assays were performed using 4.2 × 10 2 cells/mm 2 , cultured in a medium containing 25 mM sodium propionate, 48 h after transfection. The Scramble and siRNA_MUT cells were compared. The signals were expressed as relative units (R.U.). The variation between the MTT and neutral-red fold-changes was reported (∆ = 0.26). Experiments were performed in three independent replicates. The results are reported as the mean ± SD. A one-way two-tail paired t-test was used to calculate the statistical significance (p-value); *** = p < 0.005. Figure 6. Cell viability and mitochondrial functionality assays of siRNA_MUT cells in sodium propionate-medium. Neutral-red and MTT assays were performed using 4.2 × 10 2 cells/mm 2 , cultured in a medium containing 25 mM sodium propionate, 48 h after transfection. The Scramble and siRNA_MUT cells were compared. The signals were expressed as relative units (R.U.). The variation between the MTT and neutral-red fold-changes was reported (∆ = 0.26). Experiments were performed in three independent replicates. The results are reported as the mean ± SD. A one-way two-tail paired t-test was used to calculate the statistical significance (p-value); *** = p < 0.005. with a negative control siRNA (scramble siRNA) (sc-37007, Santa Cruz Biotechnology, Dallas, TX, USA), following identical procedures adopted for MUT siRNA, were used as the control.

Cellular Lysis and Western Blotting Analysis
The cells were washed twice with an ice-cold PBS buffer and were mechanically removed from the plates by scraping in presence of a PBS buffer. They were centrifuged at 250 Relative Centrifugal Force (RCF) for 10 min at 4 • C. The supernatant of PBS was removed and the cellular pellets were . Mouse monoclonal anti-β-actin (ab8226, Abcam, Cambridge, UK) was used as the internal control for immunoblotting at a dilution of 1:5000. Immunoblot detections were carried out using horseradish peroxidase-conjugated anti-mouse antibodies and enhanced chemiluminescence (GE Healthcare, Piscataway, NJ, USA). The signals were visualized by X-ray film exposure. The images were acquired by a GS-800 calibrated densitometer scan (Biorad, Hercules, CA, USA). The MUT protein pixels were quantified and normalized by β-actin protein signal pixels [34,35].

Apoptosis Assay by Flow Cytometry
The cells were transfected in 60 mm-diameter plates by using Scramble siRNA and MUT siRNA, respectively. After 48 h, the cells were detached from the plate by incubation for 2 min at 37 • C with Trypsin-EDTA (Sigma Aldrich, St. Louis, MO, USA), 1 mL of PBS was added to the plate, and the cellular suspension was recovered from the plate and centrifuged for 10 min at 250 RCF at 4 • C. The cell pellets were resuspended in 2 mL of ice-cold PBS and centrifuged again for 10 min at 250 RCF at 4 • C. The cell pellets were resuspended in 100 µL of 1X Binding Buffer (0.1 M HEPES pH = 7.4; 1.4 M NaCl; 25 mM CaCl 2 ) containing 5 µL fluorescein isothiocyanate (FITC)-conjugated Annexin V (BD Biosciences, San Jose, CA, USA) and 5 µL propidium iodide (PI), for 15 min at room temperature in the dark. Subsequently, 400 µL of 1X Binding Buffer was added to each sample and the cells were analyzed, within 10 min using a FACSCanto II flow cytometer (BD Biosciences, San Jose, CA, USA). The cells with high Annexin V and low PI signals were considered to be in the early stages of the apoptotic process. The cells with high Annexin and PI signals were considered to be in late apoptosis or necrosis. The cell transfections and subsequent analyses were performed in three independent replicates.

Neutral-Red and MTT Assays
Both assays were performed as elsewhere reported [36]. Briefly, 48 h after propionate treatment (72 h after transfection), the cells were washed with PBS, and the culture medium was replaced with a fresh medium containing 0.5 mg/mL of MTT (Sigma-Aldrich, St. Louis, MO, USA) or 0.33 mg/mL neutral-red (Sigma-Aldrich, St. Louis, MO, USA). The cells were incubated for 2 h at 37 • C and then washed with PBS, which was completely removed. Then, for the MTT, a solution of 1 N hydrogen chloride-isopropanol (1:24, v:v) was pipetted to each well, and mixed to dissolve the dark-blue formazan crystals formed. After a few minutes of gentle agitation on a rocking platform at room temperature, the absorbance of each sample was read at 570 nm in a Perkin Elmer Enspire microplate reader. For the neutral-red assay, a solution of acetic acid-water-ethanol (1:49:49, v:v:v) was pipetted to each well to solubilize the dye, and after a few minutes of gentle agitation, the absorbance of each sample was read at 540 nm in the plate reader. For neutral-red and MTT assays, 4.2 × 10 2 cells/mm 2 were seeded into the wells of a 24-well microplate (Costar, Corning Inc., Corning, NY, USA). The transfections were performed as described above. 24 h after transfection, the culture medium was replaced with a fresh medium containing 25 mM sodium propionate. A neutral-red assay was performed onto untrasfected, Scramble, and siRNA_MUT cells. An MTT assay was performed onto Scramble and siRNA_MUT cells. The values were normalized versus untreated untransfected cells in the neutral-red assay and versus Scramble cells in the MTT assay, and expressed as relative units (R.U.).
Experiments were performed in three independent replicates and the averages and standard deviations were reported on to the graphs. One-way two-tail paired t-test was used to calculate the statistical significance (p-value).

Proteomic Analysis
Aliquots (100 µg) of the protein extracts from three replicates for each cellular condition were fractionated on a preparative by 10% SDS-PAGE, 16 × 20 cm. The protein electrophoretic patterns were stained using Gel Code Blue Stain Reagent (Thermo Fisher Scientific, Waltham, MA, USA). Each gel lane was cut into 5 mm slices and these later were excised from gel. An in situ trypsin digestion of the slices was carried out [37][38][39][40]. Peptide mixtures were resuspended in 0.2% HCOOH and an MS analysis was performed using a LTQ-Orbitrap XL (Thermo Scientific, Bremen, Germany) coupled with nanoEASY II, Nanoseparations chromatographic system (75 µm-L 20 cm, column, Thermo Scientific, Bremen, Germany). The peptide mixture was concentrated and desalted onto a 2 cm trapping column (C18, ID 100 µm, 5 µm) and then fractionated onto 20 cm C18 reverse phase silica capillary column (ID 75 µm, 5 µm) (Nanoseparations). The peptides were eluted by a nonlinear gradient-4% B solvent (A eluent: 0.1% formic acid; B eluent: 80% acetonitrile, 0.08% formic acid) during 5 min, from 4 to 40% B in 45 min, and from 40 to 90% B in 1 min at flow rate of 250 nL/min [17]. An MS analysis was performed with a resolution set to 30000, and mass range from m/z 400 to 1800 Da. The three most intense doubly, triply, and fourthly charged ions were selected and fragmented using Collision Induced Dissociation (CID) fragmentation. A proteomic analysis was performed using a Proteome Discoverer™ platform (version 1.3.0.339; Thermo Scientific, Bremen, Germany), interfaced with an in-house Mascot server (version 2.3, Matrix Science, London, UK) for protein identifications. All of the peak lists were processed using the following parameters: ( [20].

Quantitative Label-Free Comparative Analysis
The spectral counting (SpC) approach [15,16] was used to compare the protein expression profiles of the siRNA_MUT cells with those of the negative control (Scamble). In order to perform a quantitative analysis, the abundances of the proteins present in each proteomes were estimated by means of the spectral counting, whereas the protein fold changes were expressed as R SC , calculated according to the following formula: R SC is the log ratio of abundance between samples 1 (Scramble) and 2 (siRNA_MUT); n1 and n2 are the SpCs for the given protein in sample groups 1 and 2, respectively; t1 and t2 are the total numbers of spectra over all of the proteins in the two sample groups; f is a correction factor set to 0.5 and used to eliminate discontinuity due to SpC = 0 [20]. The normalized spectral abundance factor (NSAF) for a given protein was calculated as the ratio of its spectral abundance factor SAF (SpC divided by protein length) and the sum of all SAFs for the proteins identified within that run. The NSAF values allow for comparing the relative abundance of proteins both between and within the samples. To measure the relative abundance for each protein identified in the dataset, Fold NSAF was calculated as log 2 (NSAF1/NSAF2), where NSAF1 is referred to siRNA_MUT, and NSAF2 to the Scramble conditions, respectively. Within the obtained datasets, proteins showing R SC > +3.5 or R SC <−3.5 and Fold NSAF > +3.5 or <−3.5, were considered as differentially expressed between the analyzed groups. A statistical analysis was performed using the GraphPad Prism Version 5.03 (La Jolla, CA, USA). The R SC and Fold NSAF correlation was evaluated using the Pearson's coefficient test.

Bioinformatic Analysis
To investigate the potential cellular processes affected by the MUT knockdown, we analyzed the identified proteomic dataset using PANTHER (Protein Analysis Through Evolutionary Relationship) database (Available online: http://www.pantherdb.org) [41,42]. Moreover, the Reactome database (Available online: https://www.reactome.org), an open-source and open access pathway database summarizing diverse pathway model collection, was also used to support the enrichment analysis and combine the pathway analysis with a functional classification of the differentially expressed proteins [43][44][45].

Conclusions
In summary, the proteomic characterization of a methylmalonyl-CoA mutase-silenced neuroblastoma cell line allowed us to define a dataset of deregulated proteins and relative alterated cellular pathways that may be investigated to highlight the unknown molecular mechanism underlying MMA damage. The identification of deregulated mitochondrial proteins is a key result of the definition of the molecular mechanisms involved in MMA pathophysiology. In fact, although it is clear that the increased methylmalonic acid levels affect the progression of the disease, the role of the mitochondria in this process has not been detailed yet.

Conflicts of Interest:
The authors declare no conflict of interest.