Translational Sciences; Burroughs Wellcome Fund; Vanderbilt Diabetes Research and Training Center, Vanderbilt University Medical Center; Chan Zuckerberg Initiative; National Institutes of Health; National Institute of Diabetes and Digestive and Kidney Diseases; United Negro College Fund Special Programs Corporation

bridging (MIB) complex, which is crucial for protein transport, respiratory chain complex assembly, and regulation of cristae integrity. While the role of Sam50 in mitochondrial structure and metabolism in skeletal muscle remains unclear, this study aims to investigate its impact. Serial block-face-scanning electron microscopy and computer-assisted 3D renderings were employed to compare mitochondrial structure and networking in Sam50deficient myotubes from mice and humans with wild-type (WT) myotubes. Furthermore, autophagosome 3D structure was assessed in human myotubes. Mitochondrial metabolic phenotypes were assessed using Gas Chromatography-Mass Spectrometry-based metabolomics to explore differential changes in WT and Sam50-deficient myotubes.

The results revealed increased mitochondrial fragmentation and autophagosome formation in Sam50-deficient myotubes compared to controls. Metabolomic analysis indicated elevated metabolism of propanoate and several amino acids, including ß-Alanine, phenylalanine, and tyrosine, along with increased amino acid and fatty acid metabolism in Sam50-deficient myotubes. Furthermore, impairment of oxidative capacity was observed upon Sam50 ablation in both murine and human myotubes, as measured with the XF24 Seahorse Analyzer. Collectively, these findings support the critical role of Sam50 in establishing and maintaining mitochondrial integrity, cristae structure, and mitochondrial metabolism. By elucidating the impact of Sam50-deficiency, this study enhances our understanding of mitochondrial function in skeletal muscle.

3D reconstruction, MICOS, mitochondria, SAM complex, Sam50

Mitochondria are generally associated with their bioenergetic roles in the cell as the hubs of oxidative phosphorylation (Alston et al., 2017;Duchen & Szabadkai, 2010). Aside from energy production, mitochondria are also integral in other processes such as maintaining calcium homeostasis and apoptosis (Bustos et al., 2017;Duchen & Szabadkai, 2010). Being such a functionally important organelle, mitochondria are also implicated in the pathophysiology of numerous diseases when they become dysfunctional. Understanding the processes involved in mitochondrial dysfunction may provide novel therapeutic treatment options for many common diseases. This is exemplified when considering reactive oxidative species (ROS) signaling which may alter various mitochondrial functions including ATP production, which may be linked to the development of metabolic diseases and structural changes (Rolo & Palmeira, 2006). Mutations or changes in the function of the mitochondrial DNA (mtDNA) can result in altered mitochondrial function; this can have consequences on processes such as insulin production, leading to hypertension and metabolic defects (Dabravolski et al., 2021;Li et al., 2018). Thus, the prospect of understanding how changes in mitochondrial structure may affect metabolism remains pertinent for discovering new therapeutic treatments for diseases with high morbidity and mortality (Bratic & Larsson, 2013).

Critical to the functioning of mitochondria are the structural folds of the inner membrane known as cristae (Cogliati et al., 2016). Cristae provide a means for mitochondria to maximize the amount of oxidative phosphorylation machinery that can be present relative to the internal volume (Cogliati et al., 2016). A key determining factor in mitochondrial ultrastructure is mitochondrial dynamics which consist of continuous cycles of fusion and fission. Mitochondrial dynamics give rise to various mitochondrial phenotypes which range from the typical spherical to stress-states including megamitochondria and donut-shaped (Glancy et al., 2020;Long et al., 2015). GTPases facilitate the dynamic events with Mitofusin-1 (Mfn1) and Mitofusin-2 (Mfn2) fusing the outer membrane and Optic atrophy 1 (Opa1) fusing the inner membrane (Chen et al., 2003;Cogliati et al., 2016;Frezza et al., 2006). Opa1 has been shown to be linked to mitochondrial fragmentation while also affecting cristae dimensions, shapes, and size (Hu et al., 2020). Thus, understanding the mechanisms of how modulators of cristae affect overall mitochondrial structure remains important and is still not completely understood.

Mitochondria can be separated into several distinct compartments (Supporting Information: Figure 1). The inner mitochondrial membrane that invaginates to form crista and at the cristae junctions, the mitochondrial contact site and cristae organizing system (MICOS) complex resides (Friedman et al., 2015;Hu et al., 2020;Kozjak-Pavlovic, 2017). The MICOS complex bonds to the sorting and assembly machinery (SAM) complex (Kozjak-Pavlovic, 2017). The MICOS complex is currently known to be made up of Mic10, Mic13, Mic19, Mic25, Mic26, Mic27, and Mic60 (Kozjak-Pavlovic, 2017).

Together, the interaction between the MICOS and SAM complexes connect the inner and outer mitochondrial membranes to form the mitochondrial intermembrane space bridging (MIB) complex (Kozjak-Pavlovic, 2017). These proteins and their interactions are important for organizing and structuring the mitochondria to allow for maximum and efficient respiration and ATP generation (Rampelt et al., 2017).

The MICOS complex, SAM complex, and OPA1 can all affect the dynamics of the cristae in varying ways (Anand et al., 2021;Ding et al., 2015;Hu et al., 2020;Kozjak-Pavlovic, 2017). The MICOS subunits interact with the SAM complex to assemble β-barrel proteins and maintain contact between the outer and inner membranes (Darshi & Taylor, 2008;Ding et al., 2015;Hu et al., 2020). These all are necessary to maintain the overall shape and distances in mitochondrial networks. Overall, cristae membranes are crucial as they host the electron transport chain machinery and the ATP synthase dimers, which are necessary for the production of ATP through oxidative phosphorylation (Anand et al., 2021;Cogliati et al., 2016). However, given the critical relationship observed between mitochondrial morphology and function, determining the effects of SAM-associated gene knockout is necessary to understand the functional implications concomitant with changes in the mitochondrial ultrastructure and morphology.

While the functions of the SAM complex may easily be considered as a subset of the MIB complex, the SAM complex remains an interesting target distinct from the MICOS complex. The SAM complex is made up of three subunits, Sam35, Sam37, and Sam50 (Kozjak et al., 2003). While both Sam35 and Sam37 are peripheral membrane proteins that are not required for survival, the central component, Sam50, a 7-8 nm diameter β-barrel channel, interacts with the MICOS complex to modulate protein transport and mitochondrial morphology (Darshi and Taylor, 2008;Ding et al., 2015;Kozjak-Pavlovic, 2017;Lionello et al., 2020). Specifically, Sam50 stabilizes the MIB complex for protein transport, respiratory chain complex assembly, and cristae integrity regulation (Ott et al., 2012).

In this way, it is understood that to structurally form and sustain the cristae, proteins of the MICOS complex assemble at the cristae junction and bind directly to Sam50 (Ding et al., 2015;Körner et al., 2012;Rampelt et al., 2017). Within the entire MIB complex, Sam50 is considered to be among the most important proteins, alongside members of the MICOS complex Mitofilin and Chchd3, for proper mitochondria structure and function (Ott et al., 2015). The importance of Sam50 is relevant, as loss of Sam50 can result in diminished cristae count, reduced cristae dynamic events, abnormal cristae, and a lack of cristae junctions, the sites at which cristae are tethered to the mitochondrial inner boundary membrane (Hu et al., 2020). However, the implications of the loss of Sam50 on mitochondrial dynamics in 3D is unknown.

2D microscopy techniques allow for the viewing of typical tubular mitochondria (Lam et al., 2021;Neikirk, Lopez, et al., 2023), however, 3D phenotypes of mitochondria differ between tissue types and disease states (Glancy et al., 2020;Vue, Garza-Lopez, et al., 2023;Vue, Neikirk, et al., 2023). Mitochondrial swelling can occur as a response to dysregulation of ion homeostasis (Javadov et al., 2018), while toroid mitochondria can arise in a calciumdependent manner in stress states (Ahmad et al., 2013). Through fusion and fission dynamics, it is understood that mitochondria can respond to cellular stress to balance factors including volume, which typically allows for greater ATP generation, and surface area, which typically allows for organelle-to-organelle contacts which are crucial for extraneous processes such as calcium homeostasis (Garza-Lopez et al., 2022;Glancy et al., 2020;Vincent et al., 2019). Therefore, 3D microscopy techniques (Garza-Lopez et al., 2022;Marshall, Damo, et al., 2023;Marshall, Krystofiak, et al., 2023;Marshall, Neikirk, et al., 2023) are required for an understanding of many of the diverse phenotypes which mitochondria might display.

This study specifically examined the metabolomics of the knockdown of Sam50 in murine and human myotubes. For the knockout of genes, myotubes, derived from satellite cell fusion, have emerged as robust cellular models which mimic the properties found in skeletal muscle (Smolina et al., 2015). Importantly, human and murine myotubes remain distinct, with separate phenotypes and metabolic progression (Boldrin et al., 2010).

Therefore, human myotubes were ex vivo isolated from skeletal muscle biopsies and differentiated into multinucleated myotubes in culture (Aas et al., 2013). To measure organelle structure, we used a mixed-method approach utilizing principally serial blockface-scanning electron microscopy (SBF-SEM) to allow for manual tracings to be performed on z-direction slices, known as orthos, which were then aligned in analysis software Amira (Garza-Lopez et al., 2022), and reconstructed into 3D structures. SBF-SEM importantly allows for the ability to see mitochondrial complexity, as it offers high x-and y-resolution and range. We compared differences in mitochondrial structure and function based on human myotubes to better elucidate if Sam50 may have a differential role in these models. Beyond structural dynamics, we illustrated how changes in mitochondrial morphology upon Sam50deficinecy may alter mitochondrial function and general biochemical pathways, which offers implications for therapeutical targets involved in Sam50.

All procedures for the care of mice were in accordance with humane and ethical protocols approved by the University of Iowa Animal Care and Use Committee (IACUC) which follows the National Institute of Health (NIH) Guide for the Care and Use of Laboratory Animals, as described previously (Pereira et al., 2017). Experiments utilized wildtype (WT) male C57Bl/6 J mice housed at 22°C on a 12-h light/12-h dark cycle with free access to water and standard chow. Mice were anesthetized with 5% isoflurane/95% oxygen.

Satellite cell isolation was performed as previously described (Pereira et al., 2017). Satellite cells from C57Bl/6 J mice were plated on BD Matrigel-coated dishes and activated to differentiate into myoblasts and myotube. Specifically, for myoblast differentiation, mouse-derived satellite cells with Dulbecco's modified Eagle medium (DMEM)-F12 containing 20% fetal bovine serum (FBS), 40 ng/ml basic fibroblast growth factor, 1 × nonessential amino acids, 0.14 mM βmercaptoethanol, 1 × penicillin/streptomycin, and Fungizone. then maintained with 10 ng/mL basic fibroblast growth factor. When myoblasts were 90% confluent, they were differentiated to myotubes in DMEM-F12 containing 2% FBS and 1 × insulin-transferrin-selenium.

Three days after differentiation, myotubes were infected with 1 µg CRISPR/Cas9 (Santa Cruz Sam50 CRISPR Plasmid; sc-427129) to achieve Sam50 deletion, which was validated by qPCR. Experiments were performed 3-7 days after infection.

The RNA was isolated with an RNeasy kit (Qiagen Inc.) and quantified by measuring absorbance at 260 nm and 280 nm with a NanoDrop 1000 spectrophotometer (NanoDrop products). Isolated RNA (~1 µg) was reverse-transcribed with a High Capacity cDNA Reverse Transcription Kit (Applied Biosciences, Carlsbad CA) and amplified by real-time quantitative PCR (qPCR) with SYBR Green (Life Technologies) (Boudina et al., 2007). Three for each qPCR, t samples (~50 ng DNA each) were placed in a 384-well plate and underwent thermal cycling in an ABI Prism 7900 HT instrument (Applied Biosystems). Thermal cycling conditions were set as follows: 1 cycle at 95°C for 10 min; 40 cycles of 95°C for 15 s, 59°C for 15 s, 72°C for 30 s, and 78°C for 10 s; 1 cycle of 95°C for 15 s; 1 cycle of 60°C for 15 s; and one cycle of 95°C for 15 s. Results were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and presented as relative fold changes.

OCR was measured for Sam50 KD fibroblasts using an XF24 bioanalyzer (Seahorse Bioscience) as previously described (Dranka et al., 2011;Pereira et al., 2017). Briefly, cells were plated at a density of 20 × 10 3 per well and differentiated for 3 days. For Sam50 KO models, 3 days after differentiation, myotubes were transfected with CRISPR/Cas9 to achieve KO as described above. Cells were incubated in XF-DMEM (supplemented with 1 g/L D-Glucose, 0.11 g/L sodium pyruvate, and 4 mM L-Glutamine), without CO 2 for 60 min. Cells were subsequently treated with mitochondrial stress modulators oligomycin (1 μg/mL), carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP; 1 μM), rotenone (1 μM), and antimycin A (10 μM), in that order, while remaining in the XF-DMEM media and assessed using an XF24 Seahorse Analyzer.

After analysis, cells were lysed in 20 μL lysis buffer (10 mM Tris, 0.1% TX-100, pH 7.4) (Dranka et al., 2011), and 480 μL of Bradford reagent. was added to each well as described previously (Boldrin et al., 2010). Total protein concentration was measured at an absorbance at 595 nm and used for normalization. For each sample, three independent experiments were performed for each condition with representative data from the replicates being shown.

2.5 | Segmentation and quantification of 3D SBF-SEM images using amira SBF-SEM orthoslices were 3D reconstructed using contour tracing (manual segmentation) in Amira to perform 3D reconstruction, as described previously (Garza-Lopez et al., 2022;Hinton et al., 2023). 300-400 slices orthoslices were transferred to Amira, and 50-100 serial sections were chosen, stacked, aligned, and visualized. A blinded individual familiar with organelle morphology traced structural features manually on sequential slices of micrograph blocks.

Measurements were taken using Amira, and algorithms were manually entered for measurements not included in Amira, such as sphericity and complexity index. The data are biological replicates.

Samples were extracted for metabolites and prepared as previously designed (Phillips et al., 2022) and performed at University of Iowa High Resolution Mass Spectrometry Facility according to standard procedures. TraceFinder 4.1 was used preloaded with standard verified peaks and retention times to compare metabolite peaks in each sample against an in-house library of standards. to correct for drift over time by using QC samples, we used previously described protocols (Li et al., 2017). All data were normalized to an internal standard to control for extraction, derivatization, and/or loading effects. Metabolomic analysis was performed as described previously (Phillips et al., 2022) using the web service MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/MetaboAnalyst/ModuleView.xhtml, last accessed on September 1, 2022). All tests performed were using built-in comparison, including one-way ANOVA and Fisher's LSD multiple comparison test. The fold enrichment number was calculated by the observed hits by the expected hits, which were determined by MetaboAnalyst 5.0.

In graphs, black bars represent the standard error, and dots represent individual data points. An unpaired t-test was used for data with only two groups while analysis of the variance (ANOVA) with Tukeys' post hoc analysis was used for data with two or more groups. For both types of analyses, GraphPad Prism software package was used.

A minimum threshold of p < 0.05 indicated a significant difference.

Higher degrees of statistical significance (**, ***, ****) were defined as p < 0.01, p < 0.001, and p < 0.0001, respectively.

3.1 | Loss of Sam50 causes decrease in mitochondrial size and alterations in morphology in murine and human myotubes Givfy efficiency (Vue, Neikirk, et al., 2023), we first sought to determine if mitochondrial size changed as a result of Sam50 ablation.

Following CRISPR/Cas9 knockout and validation of Sam50 in myotubes (Figure 1a-a''), SBF-SEM micrographs were obtained (Figure 1b), and we performed manual contour segmentation for 3D reconstruction of the organelles (Figure 1c,d) and subsequent quantification (Figure 1e). For each experimental condition, 10 myotubes were used with sectioning performed on 10 µm by 10 µm slices with a z-directional depth of 300 µm (Figure 1b). Fifty orthoslices (Figure 2a,b) were surveyed and manual contour tracing was performed on each mitochondrion to perform 3D reconstructions (Figure 2a'-b'). This allowed for observation of the complete mitochondrial 3D structure in murine myotubes (Figure 2a''-b'') as well as human myotubes (Figure 2a'''-b''''') (Marshall, Garza-Lopez, et al., 2023). In each experimental condition, three regions of interests (ROI), which represent a distinct cellular location in which mitochondria were measured, are selected. Within each ROI approximately 250 mitochondria were quantified for a total of 750 mitochondria (Supporting Information: Figure 2).

We observed that mitochondrial size was decreased in Sam50deficient murine myotubes compared to control myotubes, as Sam50-deficient myotubes exhibited lower measurements for several measurements quantifying size (Figure 2c-e). The area in 3D, analogous to surface area, decreased upon Sam50 loss (Figure 2c), along with a decrease in volume (Figure 2d). This mimicked a loss in perimeter (Figure 2e), reducing the total mitochondrion surface area.

In considering Sam50-deficient human myotubes (Figure 2c'-e'), we noticed a similar phenotype with loss of mitochondrial volume, area, and perimeter when compared to control myotubes (Figure 2c'-e').

From there, we sought to understand how beyond only size, mitochondria may differentially change in shape in murine and human myotubes following loss of Sam50.

To begin with, we showed 3D-reconstructed mitochondria, from the transverse and longitudinal planes, in control (Figure 3a-a') and Sam50-deficient murine myotubes (Figure 3b-b'), as well as human myotubes (Figure 3a''-b''). Viewing mitochondria from both of these planes showed that at baseline murine myotubes are much more complex than human myotubes. While both murine and human myotubes showed decreased size, potentially representative of fragmentation, following loss of Sam50, there was a more significant loss of structure in murine myotubes. Next, we sought to determine whether mitochondrial sphericity was altered. Sam50-deficiency was associated with an increase in mitochondrial sphericity, which is indicative of reduced complexity in murine (Figure 3c) and human myotubes (Figure 3c'). To further understand mitochondrial complexity, we measured the mitochondrial complexity index which is a measure of the surface area-to-volume ratio. Mitochondria become significantly less complex upon loss of Sam50 (Vincent et al., 2019) in murine (Figure 3d) and human myotubes (Figure 3d'). This was further validated with mito-otyping, an approach that involves organizing mitochondria by volume in a similar manner to the arrangement of chromosomes during karyotyping. With mito-otyping, we observed that while control mitochondria were highly elongated Sam50deficient mitochondria generally become less elongated, resulting in a decrease in overall volume in both murine (Figure 3f) and human (Figure 3f') myotubes. Looking at intra-ROI heterogeneity shows that generally there is consistent intra-data set heterogeneity with moderate between-group differences (Supporting Information:

Figure 2). Overall, loss of Sam50 has been shown to be associated with dysregulation of mitochondrial structure. Considering the dramatic decrease in mitochondrial size and alterations in structure observed in Sam50-deficient myotubes, we also sought to investigate whether these morphological effects influenced the efficiency of mitochondrial functions.

Macroautophagy is the most prevalent cellular mechanism for the clearing of organelles, including mitochondria, principally through autophagosomes (Kamat et al., 2014). Mitophayge, or the targeting of mitochondria in autophagy can serve as a critical cellular response toROS or other forms of dysfunctional mitochondria; as such, mitophagy has been proposed as a key target in limiting the pathological progression of neurodegenerative diseases (Kamat et al., 2014). Sam50 may play a crucial role in the biochemical triggering of mitophagy, given that Sam50 interacts with autophagy receptor p62 (Abudu et al., 2021). Beyond this, past research has also implicated Sam50 in mitophagy as Sam50 deficiency results in autophagic flux, with mtDNA protecting from complete mitochondrial degradation (Jian et al., 2018). However, autophagy remains a dynamic process that may be more effectively studied using 3D reconstruction (Neikirk, Vue, et al., 2023), and the effects of Sam50deficiency on the structure of autophagic organelle machinery are poorly elucidated.

Since both human and murine myotubes reflected mitochondrial structural arrangements with loss of Sam50, we focused on human myotubes. Here, again using 3D reconstruction from 10 µm by 10 µm slices obtained from 10 human myotubes, with a z-directional depth of 300 µm, we surveyed 50 orthoslices (Figure 4a). Following manual segmentation of these 50 orthoslices, 3D renderings of autophagosomes were able to be created for control and Sam50-deficient human myotubes (Figure 4b,b'). We found that autophagosomes, when normalized to the same area ROIs, were much higher in count upon loss of Sam50 in human myotubes (Figure 4c). Beyond only count, the average volume of each autophagosome increased (Figure 4d). These increases in overall size can be illustrated when autophagosomes are organized based on their volume (Figure 4e). This suggests larger autophagosomes that may be more able to bind increased cargo loads, consistent with the observation that Sam50 depletion leads to increased mitophagy (Jian et al., 2018 et al., 2018). Together, these results show that autophagosome formation was increased in Sam50-deficient myotubes compared to control myotubes.

To investigate whether observed structural changes in mitochondria influenced mitochondrial efficiency, we assessed the overall oxidative capacity of control and Sam50-deficient human and murine myotubes.

For these experiments, we used an XF24 Seahorse Analyzer which measures OCR, an indicator of mitochondrial oxidative photophosphorylation, across various drug applications. Before mitochondrial efficiency studies, knockout of Sam50 was confirmed in both murine and human myotubes (Figure 1a',a''). Overall, Sam50deficient murine and human myotubes showed reduced OCR at all time intervals (Figure 5a,a'). To determine basal OCR, cells were assessed under normal oxidative conditions. Basal OCR was impaired in cells lacking Sam50, compared to control cells (Figure 5b,b'). A series of oxidation modulators were applied to further investigate the effects of Sam50 on oxidative processes. First, oligomycin was applied to inhibit ATPase and induce hyperpolarization of the membrane, which allowed us to examine the role of Sam50 in ATP-linked respiration (Figure 5c,c'). Carbonyl cyanide-4 (trifluoromethoxy) phenylhydrazone (FCCP), a mitochondrial oxidative phosphorylation uncoupling agent, was then applied. With FCCP treatment, we show that under conditions of maximal respiration, OCR is reduced with the loss of Sam50 compared to the control (Figure 5d,d'). Finally, myotubes were treated with antimycin-a and rotenone, inhibitors of complexes in the electron transport chain, which allowed for the glycolytic reserve to be determined as a function between maximum OCR and baseline OCR (Figure 5e,e'). Overall, with our mitochondrial functional assay, we demonstrate that oxidative capacity is impaired upon ablation of Sam50 in murine and human myotubes.

Since human and murine myotubes had similar structural changes with the loss of Sam50, we specifically looked at human myotubes to investigate the specific biochemical changes resulting from the loss of Sam50. To investigate the metabolic effects of Sam50 loss, we employed GC-MS-based metabolomics. Analysis using a volcano plot revealed significant downregulation of ß-alanine, gammaaminobutyric acid (GABA), and hypotaurine upon Sam50 loss (Figure 6a). Heatmap analysis further showed downregulation of various metabolites, including amino acids such as glutamine, phenylalanine, glutamate, threonine, and proline, suggesting a decrease in amino acid products due to increased amino acid metabolism. On the other hand, ribose, ß-Hydroxy, indolepropionic, inotisol, and dihydrophenylalanine were upregulated (Figure 6b). Enrichment analysis confirmed these findings, with significant enrichment observed in the metabolism and degradation of amino acids and fatty acids. Additionally, the metabolomic analysis revealed an increase in amino acid metabolism and fatty acid metabolism (Figure 6c). Collectively, these findings suggest that altered biochemical pathways may arise in response to, or as a mechanism of, the disrupted 3D structure of mitochondria upon Sam50 loss. Together, these data suggest Sam50 is critical for the establishment and maintenance of mitochondria, mitochondrial cristae structure, and mitochondrial metabolism. To our knowledge, we are the first group to look at the 3D structure of mitochondria in murine and human myotubes following the genetic knockdown of Sam50.

Previously, transmission electron microscopy has been utilized to observe that Sam50 can be used as a therapeutic target, with overexpression rescuing mitochondrial morphology upon injuryinduced loss of morphology (Yin et al., 2022). Interestingly, our results show that Sam50 may be functionally required for many of the diverse phenotypes previously observed in other tissue types (Glancy et al., 2020). Loss of Sam50, importantly, showed a loss of mitochondrial morphology diversity, which may have functional impacts on functions including calcium homeostasis and mitochondria-organelle contacts.

Previously, studies have shown that Sam50 levels may rise in response to mitochondrial dysfunction, suggesting it acts as a sort of recovery molecule (Yin et al., 2022). This highlights the importance of understanding how Sam50 interacts with the MICOS complex to influence overall mitochondrial structure. While it is well understood that the MICOS structure modulates cristae changes, here we also show a structure loss in mitochondria structure upon loss of the SAM complex. It is possible this occurs due to the interaction of the MICOS complex, which may also affect overall mitochondrial dynamics, potentially through the modulation of fusion and fission dynamics such as DRP1 and OPA1, which also affects cristae structure (Frezza et al., 2006). Future experiments may look at cristae to understand how cristae structural rearrangement alters 3D mitochondrial general phenotype in the loss of the MIB complex.

Therefore, future experiments should consider further utilizing focused ion beam-scanning electron microscopy, which allows for machine-learning outputs of 3D cristae morphology to be constructed (Garza-Lopez et al., 2022;Marshall, Damo, et al., 2023;Vidavsky et al., 2016).

In this study, we use 3D microscopy to dissect changes in of treatments to correct cellular defects (Yin et al., 2022). Specifically, overexpression of Sam50 was shown to rescue ischemia/reperfusion injury-induced changes by Yin and colleagues in a neural model (Yin et al., 2022), highlighting Sam50 may be an important therapeutic in a mitochondrial-dependent manner.

Importantly, our results show the importance of understanding the functional impact of mitochondria decreasing in elongation, as past mechanistic studies remain conflicting. Previous studies have shown that mitochondrial elongation may be indicative of reduced fission dynamics, such as due to the loss of Drp1 (Santos et al., 2015).

Mitochondria dysfunction may also arise as a result of mtDNA mutations (Dabravolski et al., 2021). Interestingly, in a Drp1dependent manner, mtDNA mutations may cause mitochondrial morphological alterations in Parkinson's disease; however, knockdown of Drp1 can rescue mitochondria through elongation (Santos et al., 2015). Elongated mitochondria can also serve a distinct function to distribute membrane potential across a wider area (Glancy et al., 2020). Past studies have shown that loss of the mitochondrial membrane potential alters mitochondria volume (Safiulina et al., 2006). Membrane potential is also regulated in cristae dynamics (Ghochani et al., 2010), suggesting alteration in the MIB complex may affect membrane potential. However, past research has shown that Sam50 downregulation does not drastically alter membrane potential (Ott et al., 2012). Beyond this, it has also been suggested that elongated mitochondria offer increased ATP production and protect against mitophagy, which may aid in explaining the uptick in autophagy observed following Sam50 KO (Putti et al., 2015).

Outside of elongated mitochondria, a key mitochondrial characteristic observed exclusively in WT myotubes is a diversity of phenotypes. Interestingly, there is greater complexity in murine myotubes but reduced elongation when compared to murine myotubes. For both myotubes, following Sam50 KO, many phenotypes have much higher complexity (Figure 3d), as well as higher surface area. This is epitomized by toroid-shaped mitochondria, which offer mitochondria an increased surface area, at the expense of volume for cristae (Glancy et al., 2020). Although calcium-dependent donut mitochondria have been observed in states of ROS stress and disease (Ahmad et al., 2013;Glancy et al., 2020), they may also serve beneficial tissue-dependent roles (Bleck et al., 2018). Their absence in Sam50-deficient models may signify a potential inhibition of calcium.

Beyond this, their increased surface area may allow WT myotubes to have increased contact sites with the endoplasmic reticulum, lipid droplets, and other organelles. This remains relevant as mitochondria endoplasmic reticulum sites serve a multitude of functions including calcium homeostasis (Giacomello & Pellegrini, 2016;Moltedo et al., 2019) and regulating autophagy as autophagosomes form at ER sites (Axe et al., 2008;Neikirk, Vue, et al., 2023). Future studies may consider further exploring this link by understanding how mitochondrial contact sites may change upon loss of Sam50.

Also offering valuable clues about what may be causing mitochondria function to be affected is metabolic changes we observed using GC-MS. Many of the results from the metabolic analysis show that loss of Sam50 may result in metabolic changes that parallel disease states. For example, elevations in levels of 3,4dihydroxyphenylalanine, which we observed increased in Sam50deficient myotubes (Figure 6a), are understood to occur in association with oxidative stress during disease states (Schoneich et al., 2017;Zhang et al., 2010). Past results have also shown that Sam50deficiency is associated with liver disease through increased lipid accumulation due to increased fatty acid oxidation (Li et al., 2021).

We noted a similar effect as beta-oxidation of fatty acids increasingly occurred in Sam50-deficiency. In the human heart, energetic demands are often carried out through the beta-oxidation of fatty acids (Parra et al., 2013). Therefore, these results suggest that it is possible that fatty acid oxidation is increased upon loss of Sam50 to allow for energy production in states when mitochondria lose optimal functioning; however, future studies may further explore if in mitochondrial disease states, fatty acid oxidation changes in a Sam50-dependent manner.

Of note, we saw largely significant reductions in levels of GABA, hypotaurine, and ß-alanine in Sam50-deficient myotubes (Figure 6a), which may have several potential implications. Beyond functioning in embryonic neurogenesis and as a neurotransmitter, GABA acts in a potential mTOR-dependent manner with accumulation resulting in mitophagy and abnormal morphology (Lakhani et al., 2014). Typically GABA is metabolized by a pathway known as the GABA shunt, found in the mitochondrial matrix; decreased frequency of GABA may be linked to upregulated upstream targets in the GABA shunt (Ravasz et al., 2017). It is well understood that the primary role of hypotaurine is as a precursor for taurine when reacted with hydrogen peroxide (Grove & Karpowicz, 2017). Critically, taurine serves numerous functions beyond antioxidant roles, protecting against mitochondrial disease, potentially through posttranslational modifications of mitochondrial tRNAs (Jong et al., 2021;Suzuki, 2002).

Together, these studies suggest a potential antioxidant role of hypotaurine in mitochondria (Ubuka et al., 2008). Similar to hypotaurine, ß-alanine is also a precursor to taurine production.

Studies have suggested that ß-alanine supplement increases skeletal muscle carnosine content, which affects calcium homeostasis and offers antioxidant abilities, which can increase muscle endurance, yet these results remain debated (Bhattacharya et al., 2015;Saunders et al., 2017). Interestingly, however, increased ß-alanine has also been shown to result in mitochondrial fragmentation and oxidative stress (Shetewy et al., 2016). Overall, results across prior studies show varied results suggesting a tissue-dependent role of ß-alanine which requires further elucidation.

These reductions in ß-alanine parallel a larger trend of amino acid metabolism, evidenced by decreased metabolites of many amino acids (Figure 6b). This can have numerous implications on the dependency of Sam50 for overall mitochondrial health during disease states. For example, the catabolism of fatty acids and amino acids is commonly observed as a method to bolster cancer growth (Li & Zhang, 2016). These changes may also be affecting the mTORC pathway (Goberdhan et al., 2016). Recently it was found that the mTORC pathway interacts in an OPA-1 mediated manner and may be associated with the overall mitochondrial structure in stress states (Hinton et al., 2024). Indeed, it has been found that the ATF4mediated integrated stress response (ISR) is responsible for mediating cellular responses to oxidative stress and regulating amino acid metabolism (Harding et al., 2003). This pathway may be activated by ER or mitochondrial stress-dependent eIF2α phosphorylation (Tian et al., 2021). Future studies may aim to understand if metabolite pathways are changing in response to the structural-related declines in respiratory efficiency, or if the inverse is happening.

Alongside metabolic changes, mitochondrial structure may be related to alterations in autophagosome volume and count that we observed. While higher numbers of autophagosomes and increased volume are consistent with an uptick in mitophagy, future work can be done to explicity define the mechanistic effects of autophagosome volume on the induction of mitophagy. Upregulation of autophagy may be targeting ROS in mitochondria, which is often the cause of mitophagy activation (Basit et al., 2017;Chen et al., 2020). Interestingly, past research has found that cytotoxic ROS-dependent molecules enter the mitochondria in a Sam50mediated manner to cause ROS build-up or apoptosis (Chiusolo et al., 2017). Therefore, loss of Sam50 may serve a cytoprotective effect by preventing apoptosis and increasing proliferation in cancer cells (Capala et al., 2016). Beyond this, Sam50 loss also contributes to an excess of ROS species (Capala et al., 2016;Lionello et al., 2020), which contributes to many pathologies. Therefore, restoring Sam50 may act as a mechanism to prevent ROS overproduction (Capala et al., 2016) and restore mitochondrial and cristae structure (Yin et al., 2022).

Together, our results additionally show that Sam50 loss may further contribute to pathology through impairing mitochondrial structures, such as toroid shape and elongated mitochondria, reducing OCR, and altering metabolic pathways. Therefore, future research evaluating the targeting of Sam50 to restore mitochondria 3D structure, function, and metabolism during in vivo disease states may prove a valuable future avenue.