Animals

Male C57BL/6 J mice (n = 88, Janvier) arrived at 8 weeks of age in our animal facility (temperature = 22 ± 2 °C, relative humidity = 50 ± 5%). Upon arrival, mice were single-housed with food/water ad libitum in a light–dark cycle with lights on between 07:00 and 19:00. Animals were habituated to their new housing conditions for 1 week prior to experiments and belonged to one of five cohorts (Table 1). Male CD-1 mice (Janvier) at least 12 weeks of age upon arrival were retired breeders and used as aggressors in the CSD paradigm. Before experiments commenced, the CD-1 mice were tested for sufficient aggressive behavior.

Chronic social defeat

The social defeat paradigm is well established [14, 30]. All animal studies were conducted in accordance with national guidelines and approved by the appropriate animal protection committees (Landesuntersuchungsamt RLP, 23 177 07/G 20–17-058). Briefly, C57 mice were either exposed to CSD or kept as controls. CSD consisted of a social defeat in the home cage of the aggressor mouse (CD-1) lasting 10 s of aggressive encounter. These episodes were repeated thrice, with different aggressors, interspersed by intervals in which the intruder and aggressor were separated through a perforated metal grid, which removed the somatosensory component of the social interaction. Following the daily triple social defeat, C57 mice were overnight housed with their opponent separated by the metal grid. The CSD paradigm lasted for 10 consecutive days. Control mice were daily placed in a novel cage for 90 s for the same 10-day period as CSD-treated mice. Similarly, during this period, control mice were housed in pairs in which the individuals were separated by a metal grid, thus closely mimicking the conditions experienced by the defeated mice. After CSD (or the control conditions), mice were again single-housed in a novel cage. No animals required exclusion for excessive wounding. The week following CSD, animals were sacrificed by decapitation after which the brain was rapidly removed.

Peripheral blood glucose measurement

Mice were fasted for 1 h to exclude an immediate postprandial effect rise of blood glucose. Then, peripheral blood was obtained from the tail vein by a small nick using a scalpel under non-restrained conditions as before [31]. The first blood drop was discarded, and morning blood glucose levels were taken thrice (AccuChek®, Roche, Switzerland) and averaged to obtain reliable values.

ELISA for hepatic glycogen measurement

Liver tissue was dissected and stored at − 80 °C until processing. Tissue was homogenized in 200 µL ddH2O and boiled for 10 min to inactivate the enzymes. The supernatant was processed according to the manufacturer’s instructions. Glycogen levels were measured in liver tissue using the Glycogen Assay Kit II (colorimetric, ab 169,558, Abcam) at 450 nm with a microplate reader (MultiScan FC, Thermo Scientific).

Brain fractionation

Whole brains were extracted and cut in halves. One-half of the brains were used for MAM isolation, and the hippocampus, cerebellum, cortex, and striatum were dissected from the remaining brain halves. For MAM isolation, half brains were homogenized in isolation buffer (IBB; 225 mM D-mannitol, 75 mM sucrose, 30 mM Tris/HCl pH 7.4, 1 mM EGTA, 10 mM HEPES) and Percoll medium (225 mM D-mannitol, 25 mM HEPES/KOH pH 7.4, 1 mM EGTA, 30% (v/v) Percoll) were prepared. Reagents were obtained from Sigma-Aldrich. The fractionation of murine brain tissues was performed using an adjusted and optimized protocol, based on previously described methods [32, 33]. Before starting the experiment, buffers and the homogenizer with the teflon pestle were pre-cooled. Whole brains were transferred into 50-mL screw cap centrifugation tubes, washed with PBS 1 ×  + EDTA 10 mM and then cut into small pieces and trypsinized for 30 min at 4 °C. After a centrifugation step at 700 g (4 °C, 10 min), the brain pieces were resuspended into ice-cold IBB and put into a glass Teflon homogenizer. After homogenization (10 strokes at 1500 rpm on ice), samples were centrifuged at 800 g for 5 min at 4 °C and the supernatant fraction was collected. Lysate fractions were obtained after one additional centrifugation at 800 g for 5 min at 4 °C. The pellet was instead resuspended in IBB PNS fraction. A further centrifugation step led to separation of the mitochondrial pellet from the cytosolic fraction (from which microsomes and ER fractions were obtained upon centrifugation at 20,000 g for 30 min at 4 °C). Mitochondria were then gently resuspended in IBB, centrifuged twice at 8000 g for 10 min at 4 °C (crude mitochondria). The pure mitochondria fraction was then obtained upon ultracentrifugation through a Percoll gradient (swing-out rotor at 95,000 g for 30 min at 4 °C, “slow brake” mode, i.e., 10 min of deceleration). Following this step, the MAM fraction was visible as a dense white band in the middle of the gradient, and the pure mitochondria fraction being visible as a yellow band close to the bottom of the tube.

The hippocampus, cerebellum, cortex, and striatum were suspended in RIPA buffer and homogenized with a glass pestle. The samples were placed in a shaker and agitated for 2 h at 4 °C. After that, the samples were centrifuged at 20,000 g for 20 min at 4 °C. The supernatants were recovered, and the protein concentration was determined with the BCA assay.

Isolation of crude synaptosomes from brain tissues

Crude synaptosomes were isolated using the Syn-PER Synaptic Protein Extraction Reagent (Thermo Scientific, 8779) according to the manufacturer’s instructions. Briefly, striatal tissues (left and right for each animal) were homogenized in 400 µL Syn-PER reagent containing protease (Protease Inhibitor Cocktail Complete™, Roche, Germany) and phosphatase inhibitors (Phosphatase Inhibitor Cocktails PhosSTOP, Roche, Germany) using a small pellet pestle (Motor Cordless Kimble). The homogenate was centrifuged at 4 °C and 1200 g for 10 min. The supernatant S1 was collected and centrifuged at 15,000 g for 20 min at 4 °C. Supernatant S2, containing the cytosolic fraction, was collected and pellet P2 (enriched in crude synaptosomes) resuspended in 70 µL of Syn-PER reagent containing protease and phosphatase inhibitors. Protein concentrations were measured using the BCA protein assay kit (Thermo Scientific, 23,225).

Immunoblotting

The protein amount of the samples was quantified via the BC assay (Interchim). Denatured lysates (95 °C, 5 min) were then separated under reducing conditions on 4–15% Mini-PROTEAN® TGX Stain-Free™ gels (Bio-Rad) and subsequently transferred onto nitrocellulose membranes using the Trans-Blot® Turbo™ transfer system (Bio-Rad). Blocking of the membranes was conducted with 3% (w/v) milk powder in TBS-T (1 × TBS, 0.05% (v/v) Tween 20) for 1 h at room temperature (RT). The Chameleon Duo pre-stained protein ladder (LI-COR) served as protein size standard. The membranes were probed with the following primary antibodies: VDAC1 (voltage-dependent anion channel 1, Abcam ab14734), MFN2 (mitofusin 2, Abnova H00009927-M03), calnexin (Abcam ab22595), GRP75/mortalin (Neuromab 75–127), COXI (cytochrome c oxidase subunit 1, Abcam 14,715), COXIV (cytochrome c oxidase subunit IV, Cell Signaling 4844), hexokinase 3 (antibodies online, ABIN392753), hexokinase 1 (Cell Signaling C35C4), GAPDH (Abcam, 1:5000), synaptophysin 1 (synaptic systems, 1:1000, O/N), PSD95 (Cell Signaling Technology, 1:1000, O/N), and beta-III tubulin (R&D Systems, MAB1195 or Sigma-Aldrich, 1:500 for membrane fractions and 1:5000 for cytosolic fractions). Anti-mouse or anti-rabbit IgG (H + L) (DyLight 680 and 800 Conjugate, New England Biolabs) served as secondary antibodies and were diluted 1:15,000 in 3% (w/v) milk powder in TBS-T. Membranes were detected with the Odyssey infrared imaging system (LI-COR).

Protein digest and LC–MS analysis

Samples were prepared for LC–MS using the previously described SP3 protocol [34]. For LC–MS analyses, a NanoAQUITY UPLC system (Waters Corporation, Milford, MA) was connected online to a Synapt G2-S high-definition mass spectrometer (Waters Corporation) through a NanoLockSpray dual electrospray ion source (Waters Corporation). Samples of 200 ng, which were tryptically digested, were loaded onto a HSS-T3 C18 1.8 μm, 75 μm × 250 mm reversed-phase analytical column (Waters Corporation) heated at 55 °C. Mobile phase A consisted of 0.1% (v/v) FA and 3% (v/v) DMSO in water and mobile phase B of 0.1% (v/v) FA and 3% (v/v) DMSO in ACN. To separate peptides, a 90-min running gradient from 5 to 40% (v/v) mobile phase B was used at a flow rate of 300 nL/min. Eluting peptides were analyzed by MS using ion mobility separation (IMS) as described before [35]. In short, precursor ion information was gathered in low-energy MS mode at constant collision energy of 4 eV, while fragment ion information was collected in the elevated energy scan applying drift-time specific collision energies. With 0.6 s for the spectral acquisition time in each mode and 0.05-s interscan delay, the overall cycle time was 1.3 s for the acquisition of one cycle of low and elevated energy data. As lock mass [Glu1]-fibrinopeptide was used at 250 fmol/µL, with a flow rate of 1.5 µL/min and analyzed every 30 s. The lock mass entered the MS via the reference sprayer of the NanoLockSpray source.

Data processing and label-free quantification

The symphony (Waters ver. 1.0.0.191) pipeline was used for raw data processing and database search of LC–MS data. The database was custom made from murine proteome from UniProt (UniProtKB release September 2018, 16,991 entries), common contaminants, and reverse database entries. Search parameters included at least 2 fragment ions that had to be detected for peptide identification. For a protein to be reported, a minimum of 5 fragments summed over all assigned peptides had to be monitored. Furthermore, missed cleavages were set to 2, protease to trypsin, fixed modifiers to carbamidomethylation at cysteine, and variable modifiers to oxidation at methionine. The target decoy strategy was applied to determine the false discovery rate (FDR) for peptide and protein identification and was set to 0.01. The data of the experimental replicates of each condition were post-processed using the software ISOQuant ver. 1.8 including retention time alignments, exact mass retention time (EMRT) and IMS clustering, normalization, and protein homology filtering. Details have been described [35]. Additionally, the following settings were used: minimum peptide length of 6 amino acids, minimum 2 peptides per protein, minimum max score per cluster 6, FDR 0.01, no missed cleavage, and no variable modifications, e.g., methionine oxidation. Of those proteins with significant abundance difference, the proteins needed to have at least 3 consecutive amino acids in the peptide fragments identified, max score > 1000, and be reported in at least 2 out of the 4 measurements per condition. Absolute sample amounts were calculated for each protein using TOP3.

Statistics

All samples represent biological replicates. The experiments and analyses were performed in a blinded fashion. Sample sizes are indicated in the figure legends. Data are expressed as mean ± SEM. A p < 0.05 was considered statistically significant. Data were checked for normal distribution using the Kolmogorov–Smirnov test. Unpaired two-tailed Student’s t test (normally distributed) or Mann–Whitney U test (not normally distributed) were used to compare sets of data obtained from two independent groups of animals. Pearson’s or Spearman’s (non-parametric) correlation coefficient was used to measure linear correlation between two sets of data. Except for proteomics data (previous section), all data were analyzed using Prism (GraphPad Software).