Ethics statement

All experiments that involved mice were done in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals [18] and approved by the Brown University Institutional Animal Care and Use Committee.

Animals

Gpt2-null animals were obtained from Knockout Mouse Project at University of California, Davis (RRID: MMRRC_047980-UCD), genotyped and maintained as described previously [11]. The background of all mice was C57BL6/J (Jackson Laboratory Strain No. 664, RRID: IMSR_JAX:000664). The mice were maintained under a 12-h light/dark cycle (lights on at 7 am and off at 7 pm). The mice had ad libitum access to feed and water. All experiments were conducted on mice at postnatal day 18 (P18).

Synaptosome preparation

The protocol was adapted from [19] and [20]. Immediately after cervical dislocation, the brain was removed at an angle with double-ended surgical spatula so that the midbrain, pons, medulla, and cerebellum were excluded (Fig. 1A). Olfactory bulb was also excluded. The tissue was rinsed in ice-cold 1X brain homogenization buffer (0.32 M sucrose, 1 mM EDTA, 5 mM Tris, pH 7.4). Fresh 1X brain homogenization buffer (10 mL/g) was added, and the brain was cut into small pieces with surgical scissors. A Kimble pestle was used to homogenize the tissue with 12 strokes (each 4 s). The tube was spun at 1300 × g for 3 min at 4 °C. The first supernatant was stored on ice and the pellet was re-suspended in half starting volume with 1X brain homogenization buffer. The pellet was homogenized with a Kimble pestle with 8 strokes (each 4 s). The tubes were spun at 1300 × g for 3 min at 4 °C. The first and second supernatants were combined, and the sample was spun at 10,000 × g for 10 min at 4 °C with a Sorvall RC6Plus centrifuge using the SS34 rotor. The pellet was re-suspended in 10 mL per gram of original tissue with 15% Percoll (Percoll, Cytiva 17089102, prepared in 1X brain homogenization buffer). The suspension was carefully layered onto a Percoll gradient (23–40%) and the tube was spun at 22,000 × g for 10 min at 4 °C. The top layer contains myelin and associated membranes, the middle layer contains synaptosomes and associated membranes and the bottom layer contains the mitochondrial fraction. After carefully removing the upper layer, the synaptosomes were aspirated and re-spun with 4 volumes of 1X brain homogenization buffer, at 16,000 × g for 10 min at 4 °C. The supernatant was carefully removed, and 4 volumes of Krebs-like buffer (118 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.2 mM CaCl2, 0.1 mM Na2HPO4.H2O, 20 mM HEPES, 10 mM glucose, pH 7.4) was added. The tube was spun at 18,000 × g for 10 min at 4 °C. The final pellet was flash frozen in liquid nitrogen for experiments in which baseline metabolite levels were determined. For all other experiments, the final pellet was re-suspended in the Krebs-like buffer with 0.5 mM glutamine (Invitrogen, 25030081).

Fig. 1

GPT2 is the primary glutamate pyruvate transaminase in isolated synaptic terminals. A Fractionation of the mouse brain into fractions containing myelin and other membranes, synaptosomes (isolated synaptic terminals) and mitochondria. Note that the synaptosomes are extracted from brain tissue containing only the forebrain, excluding olfactory bulb, midbrain and the brain stem. B Metabolic pathways relevant to GPT2 and glutamate metabolism in synaptosomes including pre-synaptic neuron, post-synaptic neuron and astrocytic components. Ala: alanine, a-keto: alpha-ketoglutarate, Glu: glutamate, Gln: glutamine, Pyr: pyruvate, GDH: glutamate dehydrogenase, GLS: glutaminase, GLUL: glutamine synthetase. C GPT2 is enriched in synaptosomes and mitochondria. Western blotting of the homogenate, cytosolic, myelin and membranes, synaptosomes and mitochondria fractions obtained from a wild-type mouse at postnatal day 18 (P18) for GPT2, GPT1, MTCO1 (mitochondrial marker, subunit 1 of complex IV cytochrome c oxidase), SYN1 (synapsin I), LSD1 (nuclear marker, lysine-specific demethylase 1), GDH (glutamate dehydrogenase), GFAP (glial fibrillary acidic protein) and GAPDH (cytosolic marker, glyceraldehyde 3-phosphate dehydrogenase). D GPT2 is enriched in mitochondria of synaptosomes. Western blotting of synaptosomes that were fractionated further into synaptic membrane, synaptic cytosol and synaptic vesicles obtained from a wild-type mouse at P18 for GPT2, SV2 (synaptic vesicle protein 2), VGLUT1 (vesicular glutamate transporter 1), VGAT (vesicular GABA transporter), GLT1 (glutamate transporter 1, astrocytic), GRIA1 (AMPA receptor subunit 1), ACTIN, GAPDH. E Overall GPT enzyme activity in Gpt2-null is greatly reduced in Gpt2-null synaptosomes at P18. Each dot represents the total enzymatic activities of GPT2 and GPT1 in synaptosome samples obtained from different wild-type (black) or Gpt2-null (red) mice. The enzymatic activities are expressed as mU/mg per protein. P < 0.0001. F Glutamine cannot be used as a precursor for alanine in Gpt2-null synaptosomes at P18. Isotope tracing in wild-type (black) and Gpt2-null (red) synaptosomes at P18 using glutamine labeled with heavy nitrogen at the amine position. Each dot represents fractional enrichment of labeled pool of alanine in synaptosome from a different animal, expressed as percentage of total alanine pool. For pellet, **P = 0.006 and for medium, ***P = 0.00016. G Alanine and alpha-ketoglutarate can be used as precursors for glutamate in wild-type but not in Gpt2-null synaptosomes at P18. Isotope tracing in wild-type (black) and Gpt2-null (red) synaptosomes at P18 using heavy nitrogen labeled alanine. Each dot represents fractional enrichment of labeled pool of glutamate in synaptosome from a different animal, expressed as percentage of total glutamate pool. For pellet and medium, ***P < 0.0001

For isolation of synaptic vesicles, after the synaptosomes were washed with 4 volumes of Krebs-like buffer, the pellet was re-suspended in 450 µl ice-cold ddH2O. The pellet was homogenized by pipetting. 4.95 µl 1 M HEPES (pH7.4, with NaOH) was immediately added and the mixture was incubated on ice for 30 min. The mixture was spun at 43,100 × g with SS-34 rotor at 4 °C for 15 min. The supernatant was distributed equally to TLA 120.1 (Beckman Coulter) rotor thick-walled polycarbonate tubes (Beckman Coulter, 343776). The tubes were spun at 165,000 × g at 4 °C for 1 h. The pellet was re-suspended in 30 µl RIPA buffer (50 mM Tris, 0.15 M NaCl, 1 mM EDTA, 1% TritonX-100, 0.5% (w/v) sodium deoxycholate, 0.1% SDS) with 1X PhosStop and 1X PIC for Western blotting.

Western blotting

Tissues were homogenized in RIPA buffer (50 mM Tris, 0.15 M NaCl, 1 mM EDTA, 1% TritonX-100, 0.5% (w/v) sodium deoxycholate, 0.1% SDS) with 1X PhosStop and 1X PIC. Protein samples (20 µg) were incubated in 1X NuPage Sample buffer (Invitrogen, NP0007) and 1X NuPage reducing agent (Invitrogen, NP0004) at 70 °C for 10 min and immediately placed on ice. The samples were run in NuPage 4–12% Bis–Tris gel (Invitrogen, NP0321) at 170 V until the dye front was at the end of the gel. The gel was transferred to a nitrocellulose membrane (Invitrogen, LC2000) in 1X FisherScientific Pierce Western Blot Transfer Buffer (35040) with 20% methanol at 30 V for 1 h. The blot was blocked with Li-Cor Blocking Buffer (Li-Cor Biosciences, 927-50000) for 30 min and permeabilized for 5 min in TBST (FisherScientific, BP2411) with 0.05% Tween20 (Sigma-Aldrich, P7949). The blot was incubated with primary antibodies in Li-Cor Intercept Antibody Diluent (Li-cor Biosciences, 927-65001) at 4 °C overnight. The blot was washed 3 × 5 min with TBST, incubated with secondary antibodies diluted in TBST for 1 h, washed with TBST and finally placed in TBS. The blot was imaged using the Li-Cor Odyssey CLx Imaging System (resolution: 84 µm, background subtraction: median) and analyzed using Image Studio Lite software (RRID:SCR_013715). Protein amount in each sample was determined using the bicinchoninic acid (BCA) assay (ThermoScientific Pierce, PI23227). Primary antibodies used in the study: mouse anti-ACTIN (Sigma, A3853, RRID:AB_262137, 42 kDa, 1:2000), mouse anti-GAPDH (Sigma, G8795, RRID:AB_1078991, 37 kDa, 1:40,000), rabbit anti-GDH (Invitrogen, PA5-29492, RRID:AB_2546968, 55 kDa, 1:2000), rabbit anti-GLS1 (Proteintech, 20170-1-AP, RRID:AB_10665373, 66 kDa, 1:500), mouse anti-GLT1 (Millipore, MAB2262, RRID:AB_10615610, 62 kDa, 1:1000), mouse anti-GLUL (BD Biosciences, 610517, RRID:AB_397879, 42 kDa, 1:2000), mouse anti-GOT1 (Invitrogen, MA5-36093, RRID:AB_2890416, 45 kDa, 1:500), rabbit anti-GOT2 (Sigma, HPA018139, RRID:AB_1849903, 45 kDa, 1:1000), mouse anti-GPT1 (Sigma, SAB1412234, RRID:AB_2885183, 55 kDa, 1:1000), rabbit anti-GPT2 (Proteintech, 167571-AP, RRID:AB_2112098, 58 kDa, 1:600), rabbit anti-GRIA1 (abcam, ab31232, RRID:AB_2113447, 100 kDa, 1:1000), rabbit anti-LSD1 (CellSignaling, 2184, RRID:AB_2070132, 110 kDa, 1:1000), mouse anti-MTCO1 (abcam, ab14705, RRID:AB_2084810, 40 kDa, 1:2000), mouse anti-SV2 (DSHB, SV2, RRID:AB_2315387, 90 kDa, 1:1000), rabbit anti-SYNAPSINI (Millipore, AB1543P, RRID:AB_90757, 77&80 kDa, 1:1000), guinea pig anti-VGAT (SynapticSystems, 131,004, RRID:AB_887873, 50 kDa, 1:1000), guinea pig anti-VGLUT1 (SynapticSystems, 135304, RRID:AB_887878, 50 kDa, 1:1000). Secondary antibodies used in the study: goat anti-Rabbit IRDye 680RD (Li-Cor Biosciences, 926-68071, RRID:AB_10956166, 1:20,000), goat anti-Rabbit IRDye 800CW (Li-Cor Biosciences, 925-32211, RRID:AB_2651127, 1:20,000), goat anti-Mouse IRDye 680RD (Li-Cor Biosciences, 925-68070, RRID:AB_2651128, 1:20,000), goat anti-Mouse IRDye 800CW (Li-Cor Biosciences, 926-32210, RRID:AB_621842, 1:20,000), donkey anti-Guinea pig IRDye 680RD (Li-Cor Biosciences, 925-68077, RRID:AB_2814914, 1:20,000), donkey anti-Guinea pig IRDye 800CW (Li-Cor Biosciences, 925-32411, RRID:AB_2814905, 1:20,000).

GPT enzyme activity assay

GPT activity assay was performed using the Alanine Aminotransferase (ALT) Assay Kit (Sigma-Aldrich, MAK052) according to the manufacturer’s manual. Synaptosomes were resuspended in 1X ALT Assay buffer. 5 µg protein sample was used. Protein amount in each sample was determined using the BCA assay (ThermoScientific Pierce, PI23227). BioTek Cytation5 plate reader was used to detect fluorescence; Gen5 software was used to analyze the data (RRID:SCR_017317).

Metabolite detection by liquid chromatography—mass spectrometry (LC–MS)

For all experiments in which metabolites were detected in synaptosomes without any additional supplements, synaptosomes were collected as above and the final pellet was flash frozen in liquid nitrogen. The pellet was homogenized in 400 µL of 80% (vol/vol) methanol cooled to -80 °C with Kimble pestle and motor, then vortexed for 1 min, incubated at − 80 °C for 4 h. The samples were spun at 14,000 × g for 10 min at 4 °C and the pellet was resuspended in 400 µL of 80% methanol while the supernatant was stored at − 80 °C. The suspension was incubated at − 80 °C for 30 min and then spun at 14,000 × g for 10 min at 4 °C. The supernatants were combined and re-spun at 14,000 × g for 10 min at 4 °C. The supernatants were concentrated using a Savant SpeedVac vacuum concentrator (Thermo Scientific) at ambient temperature in 1.5-mL Eppendorf tubes. Liquid chromatography with tandem mass spectrometry was done at Beth Israel Deaconess Medical Center Mass Spectrometry Facility. The samples were re-suspended in liquid chromatography (LC) / mass spectrometry (MS)-grade water and run-in tandem LC–MS/MS. Samples were injected into hydrophilic interaction liquid chromatography (HILIC) at high pH using HPLC coupled to a 5500 QTRAP mass spectrometer (AB/SCIEX). Selected Reaction Monitoring (SRM) mode for 300 transitions with positive/negative polarity switching fragmented precursor ions and selected for product ions. Peak areas for each detected metabolite were integrated using MultiQuant software (AB/SCIEX). All data were normalized by median normalization of at least 100 metabolite peaks. Unpaired Student’s t-test assuming equal variance was used to generate a p-value.

For all experiments that involve additional supplement with 1.6 mM alanine (Sigma, A7627) and 1.6 mM alpha-ketoglutarate (Sigma, 75890), synaptosomes in all conditions were incubated in Krebs-like buffer with 0.5 mM glutamine for 15 min at 37 °C. For evoked glutamate release, a final concentration of 50 mM KCl was added, and the suspension was incubated for 1 min at 37 °C. The pellet was spun immediately after at 16,000 × g at 4 °C for 1 min. LC − MS analysis for soluble metabolites was achieved on the Q Exactive PLUS hybrid quadrupole-orbitrap mass spectrometer (Thermo Scientific) coupled with hydrophilic interaction chromatography (HILIC). To perform the LC separation, an XBridge BEH Amide column (150 mm × 2.1 mm, 2.5 μm particle size, Waters, Milford, MA) was used with a gradient of solvent A (95%:5% H2O: acetonitrile with 20 mM ammonium acetate, 20 mM ammonium hydroxide, pH 9.4), and solvent B (100% acetonitrile). The gradient was 0 min, 85% B; 2 min, 85% B; 3 min, 80% B; 5 min, 80% B; 6 min, 75% B; 7 min, 75% B; 8 min, 70% B; 9 min, 70% B; 10 min, 50% B; 12 min, 50% B; 13 min, 25% B; 16 min, 25% B; 18 min, 0% B; 23 min, 0% B; 24 min, 85% B; 30 min, 85% B. The flow rate was 150 μL/min; the injection volume was 10 μL; the column temperature was 25 °C. MS full scans were in negative ion mode with a resolution of 140,000 at m/z 200 and scan range of 75–1000 m/z. The automatic gain control (AGC) target was 1 × 106.

Isotope Labeling in Synaptosomes. Synaptosome metabolite pool was labeled with 0.5 mM [α-15N]-glutamine (Sigma-Aldrich, 486809), 0.5 mM [U-13C]-glutamine (Sigma-Aldrich, 605166) or 1.6 mM [α-15N]-alanine (Cambridge Isotope Laboratories, NLM-454) for 15 min at 37 °C. All conditions had labeled or unlabeled 0.5 mM glutamine in Krebs-like buffer.

Brain slice preparation and electrophysiological recordings

Coronal slices (250 µm) were prepared from anesthetized mice at postnatal day 18, as described previously [21]. Briefly, the brain was immediately extracted and placed in a vibratome (Leica, VT1000). Slices were cut in ice-cold oxygenated artificial cerebrospinal fluid (ACSF, in mM): 119 NaCl, 2.5 KCl, 2.5 CaCl2.2H2O, 1.0 NaH2PO4.H2O, 1.3 MgSO4.7H2O, 26.0 NaHCO3, 11 glucose. Slices were recovered in the same oxygenated ACSF solution at room temperature for 1 h and then transferred to a recording chamber where they were continuously submerged in ACSF at 28 °C with a flow rate of 1–2 ml/min.

Whole cell recordings of CA1 pyramidal neurons were performed using a MultiClamp 700B Amplifier (Molecular Devices). Signals were low-pass filtered at 3 kHz and digitized at 10 kHz via a Digidata 1550 digitizer (Molecular Devices). Patch electrodes were fabricated from borosilicate glass capillaries (Sutter Instruments) using a P-97 micropipette puller (Sutter Instruments). Patch pipettes had a resistance of 3–8 MOhm. For determining intrinsic cell properties, the patch pipette internal solution consisted of (in mM): 125.0 KCl, 2.8 NaCl, 10.0 HEPES, 2.0 MgCl2, 2.37 ATP-Mg, 0.32 GTP-Na, 0.6 EGTA, (pH 7.23–7.28, 270–278 mOsm). Capacitance and membrane resistance were calculated from voltage response during injection of -100 pA.

For measuring miniature synaptic currents, the patch pipette internal solution consisted of (in mM): 125.0 CsCl, 2.8 NaCl, 10.0 HEPES, 2.0 MgCl2, 2.37 ATP-Mg, 0.32 GTP-Na, 0.6 EGTA, (pH 7.23–7.28, 270–278 mOsm). At least 80 events were detected in ASCF with 1 µM TTX (Tocris, 1078) and 33 µM Bicuculline (Tocris, 0130) for miniature excitatory synaptic currents or with 1 µM TTX and 10 µM DNQX (Sigma, D0540) for miniature inhibitory synaptic currents, while the cell was voltage-clamped at − 80 mV. The series resistance was monitored throughout the recording and the cell was excluded from analysis if it changed by 20%. Miniature events with multiple consecutive peaks or “shoulders” were counted for frequency but excluded from amplitude calculations. For miniature excitatory currents only: the traces were low-pass (Gaussian) filtered with a -3 dB cutoff frequency of 1000 Hz for better visualization of miniature events with smaller amplitudes. The miniature events were detected using a template compiled from each recording individually in Clampfit software (SCR_011323).

For paired pulse ratio analysis, the patch pipette internal solution consisted of (in mM): 125.0 KCl, 2.8 NaCl, 10.0 HEPES, 2.0 MgCl2, 2.37 ATP-Mg, 0.32 GTP-Na, 0.6 EGTA, 5 mM QX-314 (Tocris, 2313) (pH 7.23–7.28, 270–278 mOsm). The external solution consisted of ASCF with 50 µM d-APV (Tocris, 0106) and 100 µM picrotoxin (Sigma, P1675). CA3 was surgically removed from the slice to minimize epileptic activity. To stimulate Schaffer collaterals, bipolar stainless steel microelectrodes with 2–5 MOhm impedance (FHC, UESMEGSEKNNM) were gently placed approximately 200 µm away from the recorded pyramidal neuron. Paired pulses, 50 ms apart, were used to achieve a stable current response of approximately 150–300 pA. Paired pulse ratio was calculated by dividing the peak amplitude of the second response to that of the first. All electrophysiological recordings were analyzed using Clampfit software.

Glutamate measurement by enzymatic method

Free glutamate levels in the cytosolic fraction, that is the supernatant obtained along with the pellet that was subsequently re-suspended in Percoll (see Synaptosome Preparation), were determined by enzymatic method using glutamate dehydrogenase. Protein concentration was quantified by the bicinchoninic acid (BCA) assay (ThermoScientific Pierce, PI23227). 100 µg of each sample was added to their respective wells in a black clear-bottom 96-well plate (Greiner Bio-One, 655,090) and the final volume was completed to 200 µl with Krebs-like buffer. NADP+ (Sigma, N5755) at a final concentration of 1 mM was added to each well. After 6 min of equilibration, five units of glutamate dehydrogenase (Sigma, G2626) was added to each well. The excitation and emission filters were 340 and 460 nm, respectively. All readings were taken at 37 °C. Glutamate standard (Sigma, G1251) was used for calibration.

To validate the viability of synaptosomes, 50 mM KCl was added to synaptosomes immediately after the fluorescence reached a plateau following glutamate dehydrogenase addition.

Electron microscopy

The animals were perfused with 1X PBS and then 2.5% glutaraldehyde, 2% paraformaldehyde, 2 mM calcium chloride in 0.15 M sodium cacodylate buffer. The 100 µm thick tissue sections were obtained using a vibratome and fixed in glutaraldehyde/paraformaldehyde/calcium chloride/sodium cacodylate buffer for 3 h at 4 °C. The sections were washed in cold 0.15 M sodium cacodylate with 2 mM calcium chloride for 3 min, 5 times and then in filtered ddH2O for 3 min, 5 times. The sections were placed in filtered thiocarbohydrazide (TCH) solution (10 mg/ml in ddH2O) for 20 min. The sections were rinsed with ddH2O for 3 min, 5 times. The sections were treated with 2% osmium oxide for 30 min and then washed with ddH2O for 3 min, 5 times. The sections were placed in 1% (w/v) uranyl acetate at 4 °C overnight. The tissue was washed with ddH2O for 3 min, 5 times and then placed in lead aspartate solution (6.6 mg/ml, pH 5.5) at 60 °C for 30 min. The tissues were washed in ddH2O for 3 min, 5 times and then dehydrated in increasing concentration of ethanol (70%, 90%, 95%, 20 min once each and then 100% for 10 min, 3 times.). Epon embedding was performed with Embed-812 (Electron Microscopy Sciences, 14,120), Araldite 502 (Electron Microscopy Sciences, 13,900), DDSA (Electron Microscopy Sciences, 13,700), DMP-30 (Electron Microscopy Sciences, 13,600) in Chien molds and the tissues were left to polymerase overnight at 60 °C overnight. The electron micrographs were obtained using Philips EM 410 transmission electron microscope. Fiji was used for electron micrograph image analysis. For synaptic vesicles, only clearly visible vesicles were counted and their areas were measured using freehand selection tool. Post-synaptic density length was measured with freehand line tool. Asymmetric spines and mitochondria were counted per 21000X magnification image; only mitochondria that were visible as a whole were counted, i.e. those that appeared cut along the edges/borders of the image were excluded.

Glutamate dehydrogenase activity assay

300 µg protein of synaptosomes prepared in Krebs-like buffer were used. Final concentrations of glutamate (Sigma, G1251) and NADP (Sigma, N5755) were 1 mM each. Total final volume per well in the black clear-bottom 96-well plate (Greiner Bio-One, 655090) was 200 µl. Immediately after addition of NADP, the readings were collected at 37 °C. Cytation 5 plate reader detected fluorescence every 45 s for 1 h, with excitation and emission wavelengths of 340 and 460 nm, respectively. The data were analyzed using Gen5 software. The slope (rate of product formation) was determined from the linear phase of the fluorescence curve. For the standard curve, various concentrations of glutamate dehydrogenase (Sigma, G2626) were used.

Glutaminase activity assay

5 µg protein of synaptosomes prepared in Krebs-like buffer were used. The assay was carried out according to the manufacturer's manual (BioVision, K455).

Aspartate aminotransferase activity assay

5 µg protein of synaptosomes prepared in Krebs-like buffer were used. The assay was carried out according to the manufacturer's manual (Sigma, MAK455).

ATP assay in synaptosomes

In addition to LC–MS/MS, ATP levels in wild-type and Gpt2-null synaptosomes at P18 were quantified with ATP Assay Kit (abcam, 83,355, lot: GR3299138) according to the manufacturer’s instructions. The samples were deproteinized by PCA/KOH method as recommended by the manual. Protein amount in each sample was determined using the BCA assay (ThermoScientific Pierce, PI23227).

Experimental design and statistical analyses

All data in the Figures and the Results section are presented as average ± standard error of the mean, unless otherwise noted. The statistical test for comparisons of two groups was unpaired two-tailed Student t-test unless otherwise noted. All statistical analyses were compiled using GraphPad Prism software (RRID:SCR_002798).