Strain-specific differences in muscle Ca²⁺ transport and mitochondrial electron transport chain proteins between FVB/N and C57BL/6J mice Free

Skeletal muscle constitutes more than 40% of body mass and contributes significantly to whole-body energy homeostasis. Interestingly, muscle has the capacity to increase its energy expenditure severalfold during prolonged cold exposure and/or strenuous exercise (Periasamy et al., 2017). Hence, it is proposed to be an ideal target for increasing energy expenditure to control obesity (Palmer and Clegg, 2017; van den Berg et al., 2011). Studies have shown that sarco-endoplasmic reticulum calcium ATPase (SERCA) activity is a major determinant of muscle ATP consumption and energy expenditure (Gamu et al., 2020). Data show that SERCA-mediated ATP utilization could account for more than 40% of total energy consumption of resting mouse muscle (Smith et al., 2013). Recent investigations suggest that the SERCA pump is regulated by sarcolipin (SLN) and other newly identified micropeptides in muscle (Primeau et al., 2018; Sahoo et al., 2015; Stammers et al., 2015). SLN binding to SERCA uncouples SERCA-based Ca²⁺ uptake from ATP hydrolysis, promoting increased ATP hydrolysis and heat production, contributing to muscle thermogenesis (Maurya and Periasamy, 2015; Sahoo et al., 2013; Smith et al., 2002). In adult mice, SLN is expressed abundantly in slow-twitch and fast oxidative muscles, whereas it is very low (undetectable) in fast-twitch glycolytic muscles (Babu et al., 2007; Pant et al., 2015; Sopariwala et al., 2015). Studies in genetically altered mouse models showed that loss of SLN affected both thermogenesis and metabolism; mice failed to increase fatty acid oxidation in the skeletal muscle during cold adaptation and became obese when fed on a high fat diet (HFD) (Bal et al., 2012; Maurya et al., 2015; Rotter et al., 2018). Existing data suggest that SLN plays dual roles in muscle energy metabolism. By uncoupling SERCA activity, it increases SERCA-based ATP utilization, creating energy demand, and at the same time it activates local Ca²⁺ signaling pathways to boost mitochondrial oxidative metabolism (Gamu et al., 2014; Maurya et al., 2018; Maurya and Periasamy, 2015; Pant et al., 2016).

Studies defining the function of Ca²⁺-handling proteins including SERCA and SLN using genetic manipulation have heavily relied on two mouse strains C57BL/6J and FVB/N. Some studies conducted in (FVB/N) mice questioned the importance of SLN in energy metabolism in mice, because an obese phenotype was not observed when the mice were fed a HFD (Butler et al., 2015; Campbell and Dicke, 2018). It is well known that different mouse strains show a variable obesity phenotype when challenged with HFD: C57BL/6 mice are highly susceptible to HFD-induced obesity, and are widely used in obesity studies; FVB/N mice are moderately resistant, whereas other strains such as 129S and Bal/c are highly resistant to diet-induced obesity (Li et al., 2020; Montgomery et al., 2013; Smoczek et al., 2020; Surwit et al., 1998). In addition, thermogenic phenotype varies between mice strains (Fromme et al., 2019). The effect of deletion of uncoupling protein (UCP)1, considered to be a major player in brown fat thermogenesis, is not uniform in different backgrounds; some UCP1-KO strains show better cold tolerance than others (Hofmann et al., 2001). These observations suggest that not all mice strains are created equal and their response to adaptive thermogenesis may differ. Therefore it is not surprising that the discrepancy reported for SLN ablation may be due to the choice of mouse strain. Here, we studied whether there are fundamental differences in the expression profile of proteins involved in sarcoplasmic reticulum (SR) Ca²⁺ cycling and mitochondrial Ca²⁺ transport, and electron transport chain (ETC) proteins regulating oxidative metabolism in the skeletal muscles of two mouse strains: FVB/N and C57BL/6. The data presented here show marked differences in the expression of SLN and proteins of mitochondrial metabolism.

Wild-type C57BL/6J (n=8) and FVB/N (n=7) male mice (12 weeks old) were used for this study following a protocol approved by the Ohio State University Institutional Animal Care. All animal procedures were carried out in an American Association for Accreditation of Laboratory Animal Care (AAALAC)-accredited animal facility. The animals were housed in a controlled environment at a temperature of 22±1.0°C with a 12 h:12 h light:dark cycle and relative humidity of about 50%. Mice were provided with diet (2014, Harlan Teklad, Madison, WI, USA) and water ad libitum. The animals were killed using the approved procedure of CO₂ inhalation and cervical dislocation.

Selected muscle samples (diaphragm, quadriceps and soleus) were harvested and immediately frozen in liquid nitrogen and stored at −80°C until use. For protein sample preparation, minced muscle samples were homogenized in ice-cold homogenization buffer (20 mmol l⁻¹ Tris pH 7.8, 137 mmol l⁻¹ NaCl, 2.7 mmol l⁻¹ KCl, 1 mmol l⁻¹ MgCl₂, 5 mmol l⁻¹ NaOP, 10 mmol l⁻¹ NaF, 1% NP-40, 10% glycerol, 1 mmol l⁻¹ EDTA, 0.2 mmol l⁻¹ PMSF, 0.5 mmol l⁻¹ NaVO₄ and protease inhibitor mixture from Thermo Scientific). The homogenate was collected into 1.5 ml pre-chilled Eppendorf tubes, then centrifuged at 12,000 rpm for 30 min in a refrigerated centrifuge at 4°C. The clear protein lysate was aliquoted into pre-chilled Eppendorf tubes and protein concentration was determined by the standard Bradford protocol. Then, the protein preparations were flash frozen in liquid nitrogen and stored at −80°C until further use.

Protein levels were quantified by western blotting following previously published protocols. All proteins were analyzed on 10% Tris-glycine SDS-PAGE gels, while for SLN, 16% Tris-tricine SDS-PAGE gels were used. After electrophoresis, proteins were transferred to nitrocellulose membranes with 0.45 μm pore size, whereas SLN was transferred to 0.25 μm pore size nitrocellulose membranes. After transfer, membranes were blocked for non-specific binding with Odyssey^® blocking buffer (TBS, LI-COR Biosciences) for 1 h at room temperature. The primary antibodies for SERCA1a (1:1000 dilution), SERCA2a (1:1000 dilution), CASQ 2 (1:1000 dilution) and SLN (1:1000 dilution) were generated in the laboratory of M.P. Other antibodies and their dilutions were: mitochondrial oxidative phosphorylation antibody mixture (1:1000 dilution; catalog no. MS604, MitoSciences); anti-VDAC 1 (1:1000 dilution; catalog no. AVC-001) and anti-MCU (1:1000 dilution; catalog no. ACC-328, Alomone Labs); and anti-mNCX (1:1000 dilution; catalog no. PA5-51259, Thermo Fisher Scientific). After overnight treatment with primary antibody at 4°C, the blots were washed 3 times with 0.05% Tween 20 in TBS, and treated with the appropriate IRDye secondary antibody (LI-COR Biosciences) for 1 h at room temperature. The blots were washed again 3 times, 5 min each, in TBS and were imaged using the ODYSSEY^® CLx (LI-COR Biosciences). Western blotting was repeated at least 3 times for each protein and tissues were taken from at least 5 different mice. The signal intensity of the blots was quantified using ImageJ. Statistical differences were analyzed using GraphPad Prism 6 software. SLN expression was compared using unpaired t-tests, while for all other proteins, 2-way ANOVA for multiple comparisons using Šidák’s multiple comparisons test was performed.

Previous studies have reported that SLN expression is regulated in a developmental and muscle fiber type-specific manner in C57BL/6 strain (Babu et al., 2007). During neonatal stages of muscle development, SLN expression is very abundant in all the muscle groups, but by day 21 (weaning age), SLN expression is down-regulated in glycolytic muscle types and continues to be expressed at high levels in oxidative muscles. First, we compared the SLN level in the quadriceps and found that neither of the strains had a detectable amount of SLN expression using standard western blotting methods. Interestingly, our data show that the diaphragm and soleus from C57BL/6 mice express a 2-fold higher level of SLN than those from FVB/N mice (Fig. 1A). The soleus muscle is involved in posture maintenance and the diaphragm is closely associated with respiratory function and therefore undergoes continuous contraction. Both muscle groups in mice contain mixed muscle fiber types, with ∼40–50% of fibers being slow-twitch and the rest being fast oxidative fibers in C57BL/6 mice. Some strains such as CBA/J have much fewer oxidative fibers in the soleus (Stelzer and Widrick, 2003); although data for the FVB/N strain are currently not available, it might have a fiber-type distribution similar to that of the CBA/J strain. We also analyzed the two major SERCA isoforms expressed in the skeletal muscles: SERCA1a and SERCA2a (Fig. 1B). A moderately higher expression of both isoforms was observed in the diaphragm from C57BL/6 mice versus FVB/N mice. The glycolytic quadriceps predominantly expressed the SERCA1a isoform, and the expression level of the SERCA2a isoform was very low. SERCA1a expression was higher in the quadriceps of FVB/N mice than in C57BL/6. The oxidative soleus expressed high levels of the SERCA2a isoform, while SERCA1a expression was less than ∼10%. The mixed diaphragm also expressed the SERCA2a isoform predominantly in both strains, while SERCA1a was moderately higher in C57BL/6. The expression of calsequestrin (CASQ)2 was also analyzed in the three muscles; it was undetectable in the fast glycolytic quadriceps but showed relatively high expression in the soleus and diaphragm. Interestingly, CASQ2 did not reveal any strain-specific differences in expression. From these results, the SERCA/SLN ratio is higher in FVB/N mice, suggesting that there would be a lesser degree of uncoupling of SERCA-mediated Ca²⁺ uptake in these mice compared with C57BL/6.

It is unclear whether FVB/N mice, which show lower SLN expression, are more efficient in terms of energy conservation and consume less ATP for the same work. It can also be argued that as SLN expression is low in FVB/N mice, ectopic expression might give rise to a better phenotype. But, the physiological effect of SLN-mediated SERCA uncoupling is reliant on the expression and spatial arrangement of additional subcellular components. One of these might be mitochondrial health and the abundance of its ETC proteins. Therefore, next we analyzed the abundance of ETC proteins in the three selected muscles (Fig. 2). In soleus muscle, expression of complexes 3 and 4 was significantly higher, while complex 1 was moderately more abundant in C57BL/6 mice than in FVB/N mice. Interestingly, expression of complex 5 or ATP synthase was moderately lower in FVB/N mice than in C57BL/6 mice. This may indicate that soleus muscle, which is primarily oxidative, could have lower oxidative capacity in FVB/N mice. In FVB/N mice, the quadriceps, a glycolytic muscle (very low SLN), showed drastically lower abundance of complexes 1, 3 and 4; and ATP synthase was also significantly lower. In the diaphragm, expression of complexes 1 and 4 was drastically lower whereas expression of complexes 3 and 5 was mildly lower in FVB/N. None of the muscle showed any difference in complex 2 expression, which suggests that FADH-mediated electron transfer might be working at similar levels while NADH-mediated transfer might be limited. These observations agree with an earlier study that showed soleus muscle of FBV/N mice has lower state 4 (substrate-induced) respiration, ATP production capacity and expression of several energy-sensing mitochondrial biogenesis markers than that of C57BL/6 mice (Boudina et al., 2012). Interestingly, TFAM, a mitochondrial transcription factor essential for maintaining baseline mitochondrial number, and adenine nucleotide translocase (ANT), which is responsible for ATP delivery to cytosol, did not show much difference between the strains (Fig. 2). Our data suggest that the two mouse strains have intrinsic differences in protein expression profile, which may lead to a differential response to increased fat intake.

During conditions of higher energy demand in the skeletal muscle, ATP production is boosted by Ca²⁺ uptake into the mitochondria. The uptake of Ca²⁺ into muscle mitochondria is mediated by two major membrane-bound proteins: voltage-dependent anion channel (VDAC)1, located in the outer membrane, and mitochondrial calcium uniporter (MCU), located in the inner membrane. The degree of elevation of mitochondrial metabolism by Ca²⁺ influx is reliant on efflux of Ca²⁺ from the mitochondria that is facilitated primarily by mitochondrial Na⁺/Ca²⁺ exchanger (mNCX) (Haumann et al., 2019; Tarasov et al., 2012). Among mitochondrial Ca²⁺ transporters, VDAC1 showed significant differences in expression between the two strains; VDAC1 was found to be ∼3-fold higher in the soleus, ∼2-fold higher in the quadriceps and slightly higher in the diaphragm of C57BL/6 versus FVB/N mice (Fig. 3). FVB/N mice had lower levels of MCU in the soleus (40%) and diaphragm (32%) as compared with C57BL/6 mice. Interestingly, mNCX was found to be ∼60% lower in the soleus muscle but not in other muscle types in FVB/N mice. Further, we found lower expression of mitofusin (MFN)1, which is important in mediating SR-mitochondrial Ca²⁺ crosstalk, in FBV/N mice muscles although expression of MFN2 was not different between strains. These data suggest that FVB/N mice express significantly lower levels of proteins involved in Ca²⁺ influx into and efflux from the mitochondria. This indicates that under normal chow diet feeding, several muscle types of FVB/N mice rely less on Ca²⁺ signaling to boost mitochondrial ATP production and other mechanisms enhancing ATP synthesis may play a more primary role. It is also conceivable that this is a reflection of mitochondrial density and/or fiber type composition that varies between the different mouse strains.

Based on these results, we suggest that there are mouse strain-specific differences in the relative expression of SLN, ETC proteins and mitochondrial Ca²⁺-transport proteins. It is also likely, based on previous studies (Bal et al., 2017; Maurya et al., 2015, 2018), that a lower level of SLN affects mitochondrial content. It is not clear whether this is due to differences in the composition of fiber types in each muscle group. A few other studies also cautioned against the use of the FVB/N mice model for metabolism-related questions because of their resistance to diet-induced obesity and inherent differences in substrate utilization properties (Chadt et al., 2008; Haramizu et al., 2009; Nascimento-Sales et al., 2017; Seeger and Murphy, 2016; Sinasac et al., 2016; Vaillant et al., 2014). Intriguingly, white adipose tissue of the FVB/N mice exhibits a lack of transcriptional pathways to undergo angiogenesis and expansion (Kim et al., 2013). Differences in mediators of metabolism in skeletal muscle and white adipose tissue between the two strains might indicate that they show intrinsic divergence in energy homeostasis. In addition, there have been technical difficulties in quantifying SLN expression by some groups because it is relatively small and highly hydrophobic, and there are limited commercial antibodies that are proven to work in western blotting. These inherent technical difficulties may very well contribute some of the variance in the published data. Interestingly, in a recent study, Wang et al. (2020) developed a new SLN^−/− mice model in a C57BL/6 genetic background and confirmed the importance of SLN in Ca²⁺ homeostasis and its association with thermogenic and metabolic defects, supporting our earlier interpretations. More recent studies have shown a crucial role of SLN in adaptive thermogenesis and Ca²⁺ handling in muscles (Kaspari et al., 2020; Liu et al., 2020; Morales-Alamo et al., 2020; Nicolaisen et al., 2020). We therefore suggest that future studies should consider strain-specific differences when interpreting data. SLN expression is significantly higher in the skeletal muscles of large mammals than in those of rodents, underscoring the importance of SLN in muscle energy metabolism. Thus, the role of SLN in larger mammals including humans will provide more pertinent answers to target muscle energy homeostasis to counter metabolic disorders.

The current study does not include physiological/functional experiments or analysis of other metabolically active organs, including adipose tissue depots. An earlier study (Totsuka et al., 2003) reported differences in fiber type and cross-sectional area of muscle fibers in the soleus between C57BL/6J and BALB/cA (another inbred albino mouse strain) mice. The C57BL/6J group was also shown to perform mildly better than BALB/cA when an acute exercise challenge was imposed (Totsuka et al., 2003). Another study showed significant differences in the response of muscle to ischemia between C57BL/6J and BALB/c mice (McClung et al., 2016). A few other studies have also reported differences in the degree of response to physiological challenges between inbred mouse strains including C57BL/6J and FVB (Shababi et al., 2019). Further analysis of differences between these two strains should use RNA sequencing, physiological challenges such as exercise, cold, pCa–force measurements in isolated muscles and measurement of mitochondrial ETC protein activity (Spinazzi et al., 2012). The differences at the protein level reported here may arise as a result of strain-specific variation in transcriptional factors, which was not studied here.

We thank Valerie Bergdall and Xin-An Pu of the Ohio State University, Columbus, OH, USA, for providing WT FVB/N mice.