Abstract The branched-chain aminotransferase isozymes BCAT1 and BCAT2, segregated into distinct subcellular compartments and tissues, initiate the catabolism of branched-chain amino acids (BCAAs). However, whether and how BCAT isozymes cooperate with downstream enzymes to control BCAA homeostasis in an intact organism remain largely unknown. Here, we analyze system-wide metabolomic changes in BCAT1- or BCAT2-deficient mouse models. Loss of BCAT2 but not BCAT1 leads to accumulation of BCAAs and BCKAs, causing morbidity and mortality that can be ameliorated by dietary BCAA restriction. Through proximity labeling, enzymatic and isotope tracing assays, we provide evidence for the formation of a mitochondrial BCAA metabolon involving BCAT2 and BCKDH. Disabling the metabolon contributes to BCAT2-deficiency-induced phenotypes, which can be reversed by BCAT1-mediated BCKA reamination. These findings establish a role for metabolon formation in BCAA metabolism in vivo, and suggest a new strategy to modulate this pathway in diseases involving dysfunctional BCAA metabolism. Introduction Branched-chain amino acids (BCAAs) leucine (Leu), isoleucine (Ile) and valine (Val) are essential amino acids required for normal growth and development. Dysregulation of BCAA metabolism is associated with various disorders including maple syrup urine disease (MSUD), cardiovascular diseases, obesity, diabetes, and cancer through various cellular mechanisms, highlighting the diverse functions of BCAAs by acting as major nutrient signaling molecules, building blocks for protein synthesis or molecules for energy production ^[52]1–[53]4. Intracellular BCAA levels are controlled at the first two steps in the BCAA catabolic pathway catalyzed by the branched-chain aminotransferase isozymes (cytosolic BCAT1 and mitochondrial BCAT2) and the branched-chain α-keto acid dehydrogenase (BCKDH) complex. BCAT isozymes catalyze the reversible transamination to transfer the BCAA α-amino group to α-ketoglutarate (α-KG) for generating glutamate (Glu) and corresponding branched-chain α-keto acids (BCKAs). After transamination, BCKAs can be reaminated by BCAT or oxidized to tricarboxylic acid (TCA) cycle intermediates for energy production and/or macromolecule synthesis ([54]Fig. 1a) ^[55]1. Figure 1. Generation of BCAT1 and BCAT2 Knockout Mice. Figure 1. [56]Open in a new tab (a) Diagram of the BCAA metabolic pathway. BCAT isozymes catalyze the transamination of BCAAs to generate BCKAs. BCKAs can be reaminated by BCAT or irreversibly oxidized by BCKDH, forming R-CoA compounds that are further oxidized to enter the TCA cycle. (b) Expression correlation of genes encoding enzymes involved in the BCAA metabolic pathway in mouse tissues across developmental times. The color and size of each circle indicate the Pearson’s correlation coefficient scores for positive (red) and negative (blue) correlations. P values by a two-sided t-distribution with n-2 degrees of freedom. (c) mRNA and protein expression for BCAT1 and BCAT2 across mouse tissues. Heatmap shows the normalized expression values for mRNA (red) and protein (orange) in the GTEx database. (d) Schematic of Bcat1^−/− and Bcat2^−/− constitutive knockout (KO) mouse models. (e) Validation of Bcat1^−/− and Bcat2^−/− KO by representative Western blot analysis of major tissue types. HEK293T cells with BCAT1 and BCAT2 double knockout (DKO) and DKO cells stably expressing BCAT1 or BCAT2 (BCAT1^OE and BCAT2^OE) were analyzed as controls. Note that BCAT1 protein expression was low or undetectable in mouse heart, kidney, liver and spleen, and BCAT2 expression was low or undetectable in mouse liver. (f) Body weight growth curves of the indicated genotypes up to 12 weeks. Results are mean ± SD (N = 5 WT, 9 Bcat1^−/−, and 11 Bcat2^−/− mice) and analyzed by two-way ANOVA mixed-effect analysis with multiple comparisons. (g) Kaplan-Meier survival curves of WT, Bcat1^−/− and Bcat2^−/− mice (N = 16 WT, 11 Bcat1^−/−, and 7 Bcat2^−/− mice). P values by a log-rank Mantel-Cox test. The metabolism of BCAAs presents a number of unique features. First, the catabolism of all three BCAAs is controlled by a common flux-generating step catalyzed by BCKDH in mitochondria ^[57]5. As such, various nutrient and hormone signaling pathways can modulate BCAA metabolism through phosphorylation or dephosphorylation of BCKDH. Second, the key enzymes associated with BCAA metabolism are segregated into distinct subcellular compartments. BCAT1 is localized in cytosol whereas BCAT2 and all other downstream enzymes are confined to mitochondria. Third, BCATs and other enzymes are also segregated into distinct cell or tissue types ^[58]1,[59]6,[60]7. In contrast to other essential amino acids, BCAAs bypass ‘first pass’ metabolism in the liver due to low expression of BCAT isozymes in hepatocytes. Consequently, BCAAs reach the circulation, allowing them to act as signaling molecules for nutrient availability. Then, the predominant expression of BCAT2 in most other tissues allows extra-hepatic BCAA metabolism. Lastly, BCKDH activity can be modulated through phosphorylation in different tissues by BCKDK and PPM1K, respectively ^[61]8,[62]9. This results in significant inter-organ shuttling of BCAAs and BCKAs for complete degradation ^[63]1,[64]2. Due to these unique features, prior studies using in vitro assays or cultured cells cannot fully recapitulate mechanisms that control BCAA metabolism and homeostasis in an intact organism. Hence, the biological function of BCAT isozymes and how they interact with downstream enzymes to control compartmentalized BCAA metabolism in living organisms have remained largely unknown. The concept of a structural-functional complex formed between sequential metabolic enzymes, or a ‘metabolon’, was first introduced for the enzymes of the Krebs cycle ^[65]10,[66]11. Like many sequential metabolic enzymes, BCAT2 and BCKDH were also found to physically associate in organized supramolecular complexes in vitro by affinity purification assays ^[67]12,[68]13. The interaction between BCAT2 and the E1 subunit of BCKDH may enable substrate channeling of BCKAs (formed via BCAT2 activity) directly to BCKDH for subsequent oxidation. More importantly, BCAT2 but not BCAT1 stimulates BCKDH activity in enzyme reconstitution assays ^[69]12. Although these in vitro biochemical studies using purified enzymes or homogenized tissues provide initial evidence to support a ‘BCAA metabolon’ model, whether and how these enzymes act in concert within a multi-enzyme complex under physiological conditions are not known. In addition, how BCAT isozymes contribute to system-wide metabolism of BCAAs and other nutrients, and how these processes are deregulated in the absence of either isozyme or in diseases have not been addressed. Accordingly, the in vivo function of metabolon formation and its relevance to the metabolic health of an intact organism remain largely unknown. Here we generated mouse models deficient in BCAT1 or BCAT2 isozyme and analyzed systemic metabolic changes by quantitative metabolomic profiling of major mouse organs. Loss of BCAT2 but not BCAT1 causes BCAA and BCKA accumulations and acute mortality that can be rescued by dietary restriction of BCAAs. We provided evidence for the in vivo formation of the BCAA metabolon between BCAT2 and BCKDH multi-enzyme complexes within mitochondria. Disabling the metabolon contributes to BCAT2 deficiency-induced metabolic phenotypes, which can be reversed by BCAT1-mediated BCKA reamination. Our results not only establish in vivo function for metabolon formation in controlling BCAA homeostasis in a living organism, but also provide new insights into leveraging BCAT isozyme activity for dysfunctional BCAA metabolism in genetic or metabolic disorders. Results Expression and segregation of BCAT1 and BCAT2 isozymes BCAAs (Leu, Ile and Val) are multi-functional essential amino acids that share the first two steps of catabolism. The first step is initiated by the cytosolic BCAT1 or mitochondrial BCAT2 isozyme through reversible transamination to form Glu and BCKAs. The second enzymatic complex BCKDH, consisting of E1 (α-ketoacid dehydrogenases encoded by the BCKDHA and BCKDHB genes), E2 (dihydrolipoyl transacylase encoded by DBT) and E3 subunits (dihydrolipoamide dehydrogenase encoded by DLD), catalyzes the irreversible oxidative decarboxylation of BCKAs to form branched-chain acyl-CoAs for energy production and/or macromolecule synthesis ([70]Fig. 1a). Previous studies of individual components revealed distinct expression patterns for genes encoding the major BCAA metabolic enzymes. Specifically, the cytosolic BCAT1 is expressed in only a few tissues in mice and humans, while the mitochondrial BCAT2 is expressed almost ubiquitously, along with all other downstream enzymes involved in BCAA oxidation ^[71]1,[72]6,[73]7,[74]14. To comprehensively examine gene expression correlation, we analyzed the mRNA expression of all 13 genes associated with BCAA metabolism in seven human and mouse organs (brain, cerebellum, heart, kidney, liver, ovary and testis) across developmental time points from early organogenesis to adulthood ^[75]15. BCAT1 is highly expressed during organogenesis but progressively downregulated in most adult human tissues, whereas BCAT2 expression remains largely unchanged or increased in adulthood ([76]Extended Data Fig. 1a,[77]b). More importantly, we observed strong and positive correlation between the expression of BCAT2 and other BCAA metabolic genes across tissues and developmental time in mice and humans, whereas the expression of BCAT1 negatively correlates with BCAT2 or other genes ([78]Fig. 1b, [79]Extended Data Fig. 1c). Consistent with these findings, BCAT1 mRNA is only detected in several tissues including ovary, uterus, brain, and spinal cord, whereas BCAT2 is ubiquitously expressed in nearly all mouse and human tissues ([80]Fig. 1c, [81]Extended Data Fig. 1d). BCAT1 and BCAT2 mRNA expression patterns are further corroborated by protein expression across mouse and human tissues by quantitative proteomics ([82]Fig. 1c, [83]Extended Data Fig. 1d) ^[84]16–[85]18. Of note, the expression of BCKDK shows weaker correlation with other genes ([86]Fig. 1b), consistent with its inhibitory role of BCAA catabolism, although gene expression is not always correlated with enzyme activity ^[87]19. It is also important to note that, while both BCAT1 and BCAT2 are detected in several tissues, their expression may be segregated in distinct cell types within the same tissue. For instance, BCAT1 is only expressed in neuronal cells whereas BCAT2 is highly expressed in astrocytes and glia cells in mouse, rat and human brains ^[88]7,[89]20. In addition, while BCAT2 protein is undetectable in mouse hepatocytes, it is highly expressed in nonparenchymal liver cells such as hepatic stellate cells, Kupffer cells and liver sinusoidal endothelial cells, along with other downstream BCAA catabolic enzymes ^[90]21 ([91]Extended Data Fig. 1e,[92]f). Therefore, the expression of BCAT isozymes is segregated into distinct subcellular compartments and across organ systems, raising major questions about how they contribute to BCAA metabolism in an intact organism. Generation of mouse models to study systemic BCAA metabolism To study BCAT1- and BCAT2-mediated BCAA metabolism in intact organisms, we generated Bcat1 and Bcat2 constitutive knockout (KO) mouse models by CRISPR targeting. We designed paired sgRNAs flanking exon 6 of Bcat1 and exons 4 to 6 of Bcat2 genes, respectively ([93]Fig. 1d). After microinjecting sgRNAs with Cas9 ribonucleoproteins into C57BL/6 mouse zygotes, the founder animals were screened for the KO allele by genotyping and Sanger sequencing ([94]Extended Data Fig. 2a,[95]b). Homozygous Bcat1^−/− and Bcat2^−/− mice were generated by breeding positive founders and confirmed for the loss of BCAT1 or BCAT2 protein in major organs including heart, kidney, liver, spleen and brain, respectively ([96]Fig. 1e, [97]Extended Data Fig. 2c). Heterozygous Bcat1^+/− and Bcat2^+/− mice displayed no abnormalities compared to wild-type (WT) littermates. Homozygous Bcat1^−/− mice also developed normally without gross defect in body weight and size ([98]Fig. 1c). By contrast, Bcat2^−/− mice failed to thrive and developed weaknesses with decreased spontaneous movement after weaning. Bcat2^−/− mice diverged from WT and Bcat1^−/− mice starting at 6 weeks of age, resulting in significantly lower body weight, smaller body size, and acute mortality between 8 and 12 weeks old ([99]Fig. 1f,[100]g, [101]Extended Data Fig. 2d). Postmortem examination revealed lower epididymal fat pad weight, smaller fat cell size, decreased hindlimb skeletal muscle weight, and enlarged heart and kidney ([102]Extended Data Fig. 2e–[103]g). We also noted slightly increased plasma aspartate transaminase (AST) activity and significantly increased alanine transaminase (ALT) activity, indicating impaired liver function, but no apparent histological abnormalities of heart, kidney, and liver tissue sections that may explain the acute lethality ([104]Extended Data Fig. 2h–[105]k). Therefore, despite the shared enzymatic function, BCAT1 and BCAT2 are differentially required for the growth and metabolic health of living organisms. Metabolomic profiles in BCAT1- and BCAT2-deficient tissues To gain insights into the differential phenotypes, we employed metabolomics analysis to determine the effects of BCAT loss on whole-body metabolism across major tissues. We collected 11 organs from 6 to 8 weeks old male and female WT, Bcat1^−/−, and Bcat2^−/− mice, extracted soluble metabolites, and performed targeted metabolite profiling using a quadrupole time-of-flight (QTOF) mass spectrometer ([106]Fig. 2a, [107]Supplementary Table 1). Among 295 annotated metabolites, 1.0% to 7.1% were differentially enriched between Bcat1^−/− and WT tissues (|fold change| ≥ 1.5, P value ≤ 0.05 by ANOVA). By contrast, significantly more metabolites (12.2% to 20.0%) were differentially enriched in Bcat2^−/− relative to WT tissues ([108]Fig. 2b). The metabolomic profiles of Bcat2^−/− samples also formed a distinct cluster away from that of WT and Bcat1^−/− by principal component analysis (PCA) ([109]Extended Data Fig. 3a). Of note, among all tissues analyzed, BCAT1 KO led to the highest number of altered metabolites in the heart, consistent with a critical role of BCAT1 in reaminating BCKAs for protein synthesis in heart physiology ^[110]22. Loss of BCAT1 affected metabolic pathways including ascorbate and aldarate metabolism (BM, heart) and biosynthesis of unsaturated fatty acids (liver, muscle, pancreas, thymus) ([111]Extended Data Fig. 3b). In total, we detected 58 and 107 significantly altered metabolites in Bcat1^−/− and Bcat2^−/− tissues relative to WT controls, respectively. The most significantly altered metabolites in Bcat2^−/− but not Bcat1^−/− include BCAAs (Ile, Leu and Val), BCKAs (KIC and KMV), Ala, Asp, and Lys ([112]Fig. 2c). Figure 2. Distinct Metabolomic Profiles of BCAT1 and BCAT2 Knockout Mice. Figure 2. [113]Open in a new tab (a) Workflow for targeted metabolomic profiling of 11 tissues from 6 to 8 weeks old WT, Bcat1^−/− or Bcat2^−/− mice (N = 4 mice per genotype). (b) Fraction (%) of differentially enriched metabolites in Bcat1^−/− or Bcat2^−/− tissues relative to WT controls (|fold change| ≥ 1.5, P ≤ 0.05; N = 4 mice per genotype). (c) Significantly changed metabolites in Bcat1^−/− or Bcat2^−/− relative to WT tissues. The x- and y-axis denote the negative log10 transformed P values calculated by two-sided linear model covariate analysis of MetaboAnalyst 5.0 ^[114]23,[115]24 for differentially enriched metabolites among tissues and genotypes, respectively (N = 4 mice per genotype). (d) Metabolic pathways associated with enriched or depleted metabolites in Bcat2^−/− mice. Heatmap shows the negative log10 transformed P values from the pathway enrichment analysis of increased (red) or decreased (blue) metabolites (N = 4 mice per genotype) which were identified by two-sided unpaired t-test in the indicated tissues. (e) Relative abundance of metabolites associated with the top enriched pathways in Bcat1^−/− and Bcat2^−/− tissues. Heatmap shows the log2 fold changes of the indicated metabolites in each tissue of Bcat1^−/− or Bcat2^−/− mice relative to WT controls (N = 4 mice per genotype) which were identified by two-sided unpaired t-test in the indicated tissues. Given the more extensive metabolomic changes in Bcat2^−/− tissues, we identified the associated metabolic pathways using MetaboAnalyst 5.0 ^[116]23,[117]24. BCAA biosynthesis and degradation, aminoacyl-tRNA synthesis and lysine degradation were significantly associated with the top enriched metabolites across all tissues, whereas alanine, aspartate and glutamine metabolism, arginine biosynthesis and histidine metabolism were associated with the top depleted metabolites in most tissues ([118]Fig. 2d, [119]Extended Data Fig. 3b,[120]c). Consistent with these findings, BCAA (Leu, Ile and Val) and BCKA (KIC, KMV and KIV) levels were elevated in most Bcat2^−/− tissues relative to WT or Bcat1^−/− tissues ([121]Fig. 2e, [122]Extended Data Fig. 4). Tissue-specific alterations in the levels of metabolites associated with lysine degradation, alanine, aspartate and glutamine metabolism, such as Lys, 5-hydroxylysine, Glu, Gln, Ala, and Asp, were also noted in some Bcat2^−/− tissues including heart, liver, muscle and bone marrow ([123]Fig. 2e, [124]Extended Data Fig. 4). Compared to BCAT2 KO, loss of BCAT1 resulted in fewer altered metabolites and metabolic pathways, consistent with low or absent BCAT1 expression in most tissues and the lack of overt defects in Bcat1^−/− mice ([125]Fig. 2b–[126]e, [127]Extended Data Fig. 3 and [128]4). BCAT2 deficiency causes BCAA and BCKA accumulation Bcat2^−/− tissues displayed BCAA accumulation consistent with the predominant role of BCAT2-mediated BCAA transamination and the lack of BCAT1 expression in these tissues ([129]Fig. 2e). However, Bcat2^−/− mice also had significant accumulation of BCKAs, the metabolic products of BCAT isozymes, in most tissues including heart, kidney and liver ([130]Fig. 2c,[131]e). The accumulation of both BCAAs and BCKAs in Bcat2^−/− mice resembles human maple syrup urine disease (MSUD) caused by mutations in BCKDH, the enzyme complex immediately downstream of BCAT2 ([132]Fig. 3a). Impaired BCKDH in MSUD leads to BCAA and BCKA accumulation which can be detected in urine, blood and cerebrospinal fluid of affected patients ^[133]25. Elevation of BCAAs and BCKAs, specifically the most toxic of the group, Leu and KIC, causes developmental delay, neurological complications, lethargy, seizures, weight loss and death if untreated ^[134]25,[135]26. Early treatment through dietary restriction of BCAAs alleviates BCAA and BCKA accumulation in MSUD patients and improves clinical outcomes ^[136]27. Figure 3. BCAT2 Deficiency Induces Metabolic Phenotypes Resembling MSUD. Figure 3. [137]Open in a new tab (a) BCAT2 deficiency induces phenotypes similar to MSUD caused by mutations in the BCKDH enzyme complex. (b) Total food intake normalized to mouse body weight, measured weekly for 12 weeks. Results are mean ± SD (N = 15 WT and 14 Bcat2^−/− mice) and analyzed by two-way ANOVA mixed-effect analysis with multiple comparisons. (c) Fraction (%) of BCAA-low diet consumed by the indicated genotypes. Results are mean ± SD (N = 6 WT and 8 Bcat2^−/− mice) and analyzed by two-way ANOVA mixed-effect analysis with multiple comparisons. (d) Body weight growth curves of indicated genotypes up to 12 weeks. Results are mean ± SD (N = 5 WT and 5 Bcat2^−/− mice on normal diet and 13 Bcat2^−/− mice on BCAA-choice diet) and analyzed by two-way ANOVA mixed-effect analysis with multiple comparisons. (e) Kaplan-Meier survival curves of WT or Bcat2^−/− mice with BCAA normal or choice diet (N = 11 WT and 6 Bcat2^−/− mice on normal diet and 10 Bcat2^−/− mice on BCAA-choice diet). P values by a log-rank Mantel-Cox test. (f) Relative abundance of Leu and Ile and their respective BCKAs, KIC and KMV, across different tissues in Bcat2^−/− mice with normal or choice diet. Results are mean ± SEM (N = 3 thymus samples and 4 samples for other tissues with normal diet, N = 3 samples for Leu and Ile measurements with choice diet, and N = 2 lung samples and 3 samples for other tissues for KIC and KMV measurements with choice diet) and analyzed by two-way ANOVA. Of note, a range of mutations can occur within the BCKDH complex to variably reduce enzyme activity. Bcat2^−/− mice most resemble intermediate MSUD (iMSUD), where 3–30% of BCKDH activity is retained and disease onset ranges from juvenile to adulthood ^[138]25. In a mouse model of iMSUD, substantial BCAA and BCKA accumulations altered the biochemical, behavioral and neuropathological profiles, resulting in death in early adulthood ^[139]28,[140]29. We noted similar behavioral changes including hind limb dystonia, uncoordinated gait, and seizures in Bcat2^−/− mice. We also observed biochemical changes including decreases in brain L-DOPA, a dopamine precursor, threonine, tryptophan, and tyrosine, and increases in T2 (transverse relaxation time) signals in the brain of 6-week old Bcat2^−/− mice compared to WT mice ([141]Extended Data Fig. 5a,[142]b). Moreover, we noted alterations to brain structure including increased vacuolation in the striatum and brain stem of Bcat2^−/− mice ([143]Extended Data Fig. 5c,[144]d). Together, these phenotypes are similar to those in iMSUD mouse models ^[145]25,[146]28,[147]29. Given the similarities between Bcat2^−/− mice and human MSUD patients, we provided WT and Bcat2^−/− mice with the BCAA choice diet upon weaning at 4 weeks old, where mice had access to both chow containing normal levels of BCAAs (17.7% of total L-amino acids; see [148]Methods) and chow containing low levels of BCAAs (0.5% of L-amino acids). It is important to note that previous analysis of BCAT2-deficient mice has relied exclusively on BCAA-restricted diets and the measurements of plasma metabolites, thus the extent to which the BCAT2 deficiency-induced phenotypes are affected by dietary BCAAs across tissue types remains unknown ^[149]30,[150]31. Bcat2^−/− mice had increased total food consumption but a progressive preference for BCAA-low chow compared to WT mice ([151]Fig. 3b,[152]c), likely owing to the rodent’s ability to discriminate between diets with different amino acid composition ^[153]32. Bcat2^−/− mice on BCAA choice diet showed significantly improved body weight and increased survival compared to mice on normal diet ([154]Fig. 3d,[155]e), similar to human MSUD patients and mouse models of MSUD ^[156]27,[157]29,[158]30,[159]33. To determine the effects of dietary BCAA restriction on systemic metabolism, we measured metabolite levels by targeted metabolomics of 11 tissues from Bcat2^−/− mice after 8 weeks on normal or BCAA choice diet. Compared to Bcat2^−/− mice on normal diet, mice on BCAA choice diet had completely or partially normalized BCAA and BCKA levels in most tissues ([160]Fig. 3f, [161]Supplementary Table 2). These findings demonstrate that loss of BCAT2 but not BCAT1 leads to significant BCAA and BCKA accumulations resembling human MSUD, causing acute morbidity and mortality that can be rescued by dietary restriction of BCAAs. Formation of BCAA metabolon in living cells The systemic elevation of BCKAs in Bcat2^−/− mice raises a fundamental question about how loss of BCAT2 impacts downstream BCKA oxidation. Given that BCKAs are irreversibly oxidized by BCKDH, we reasoned that BCKA accumulation may be caused by impaired BCKDH-catalyzed oxidative decarboxylation ([162]Fig. 3a). We further hypothesized that BCKDH activity is impaired due to altered expression, phosphorylation, and/or formation of organized supramolecular complexes in the absence of BCAT2. BCKDH is located on the inner surface of mitochondrial inner membrane and consists of three subunits: a decarboxylase heterotetramer (α[2]β[2]) encoded by BCKDHA and BCKDHB (E1), a 24-meric dihydrolipoyl transacylase encoded by DBT (E2), and a dihydrolipoyl dehydrogenase encoded by DLD (E3) ^[163]2. BCKDH activity can be regulated by phosphorylation/dephosphorylation of the catalytic E1α subunit (BCKDHA) in a tissue- and signal-dependent manner ^[164]34. Specifically, the BCKDH kinase (BCKDK) phosphorylates BCKDHA to suppress BCKDH activity, whereas the PPM1K phosphatase dephosphorylates BCKDHA to reactivate BCKDH ^[165]8,[166]9. We first examined the expression of BCKDHA, BCKDK and PPM1K in Bcat2^−/− mouse tissues but noted no or minimal changes in protein levels compared to WT controls ([167]Extended Data Fig. 6a). The levels of phosphorylated E1α (P-BCKDHA), the inactive form of BCKDH, also remained unchanged in all Bcat2^−/− tissues except liver, which had lower P-BCKDHA relative to WT controls. We further quantified BCKDHA and P-BCKDHA levels in livers from a cohort of twelve WT and eleven Bcat2^−/− mice, respectively, and noted a slight increase of BCKDHA level but a significant decrease of P-BCKDHA ([168]Extended Data Fig. 6b). Thus, the impaired BCKA oxidation observed in Bcat2^−/− mice cannot be explained by the altered expression or phosphorylation of BCKDH. Through in vitro affinity purification assays, it was noted that BCAT2 and BCKDH physically associate as organized supramolecular complexes ^[169]12,[170]13. The interaction between BCAT2 and BCKDH may enable substrate channeling for efficient BCKA oxidation. Consistent with this notion, BCAT2 but not BCAT1 was able to stimulate BCKDH activity by 12-fold in reconstitution assays ^[171]12. Despite these in vitro biochemical studies supporting a ‘BCAA metabolon’ model, whether and how these enzymes act in concert within a multi-enzyme complex under physiological conditions have remained unknown. To this end, we designed a proximity labeling assay to assess BCAA metabolon formation in mitochondria and to identify the interacting proteins of BCAT2-BCKDH-associated complexes in living cells ([172]Fig. 4a). We first generated BCAT1 and BCAT2 double knockout (DKO) HEK293T cells by CRISPR using paired sgRNAs flanking exons 3 to 6 of BCAT1 and exons 4 to 6 of BCAT2 genes, respectively, and confirmed the loss of BCAT1 and BCAT2 proteins in DKO cells ([173]Fig. 4b). We next fused the miniTurboID (mTD) promiscuous biotin ligase ^[174]35 to the C-terminus of human BCAT2 cDNA together with a V5 tag under the doxycycline (Dox)-inducible promoter. We selected mTD over TurboID or other biotin ligases ^[175]35 due to its smaller size for more efficient mitochondrial localization ([176]Extended Data Fig. 6c) and high enzyme activity for labeling mitochondrial proteins upon supplementation of exogenous biotin ([177]Extended Data Fig. 6d). In addition, the fast labeling kinetics and short labeling radius made mTD suitable for the dissection of dynamic protein complexes in live cells based on proximity biotinylation ^[178]35. As controls, we generated BCAT1-mTD fusion cDNA or mTD alone with the 27-amino-acid mitochondrial targeting signal peptide from BCAT2. These transgenes were then stably expressed in BCAT DKO cells to generate control, BCAT1 or BCAT2 reconstituted cells, respectively ([179]Fig. 4a). We confirmed the comparable expression levels of individual transgenes, the exclusive mitochondrial localization, and the efficient biotinylation of mitochondrial proteins in the presence of Dox and exogenous biotin ([180]Extended Data Fig. 6c,[181]d). Figure 4. Formation of BCAT2-BCKDH Complexes in Living Cells. Figure 4. [182]Open in a new tab (a) Schematic of the proximity labeling assays to determine BCAT2-interacting proteins in the mitochondria of living cells. (b) Schematic for the generation of BCAT1 and BCAT2 DKO HEK293T cells and the Western blot validation of single cell-derived clones. (c) Representative Western blot analysis of in vivo biotinylated proteins purified by streptavidin IP in DKO cells reconstituted with mTD, BCAT1-mTD or BCAT2-mTD. The BCKDH subunits E1 (BCKDHA and BCKDHB), E2 (DBT), E3 (DLD) and GLUD1 were specifically labeled by BCAT2-mTD and confirmed by mass spectrometry-based proteomics. (d) Validation of BCAT2-interacting proteins by Western blot analysis of streptavidin-purified mitochondrial proteins using antibodies specific to the indicated protein. Using this assay, we induced proximity labeling, isolated mitochondria, and purified biotin-labeled protein complexes using streptavidin magnetic beads in control (mTD alone) and BCAT1 or BCAT2 reconstituted cells, respectively ([183]Fig. 4a). We first noted the significant enrichment of a number of distinct signals by Western blot in BCAT2-mTD but not BCAT1-mTD or mTD-expressing cells ([184]Fig. 4c). More importantly, the proteins labeled by BCAT2-mTD were determined by mass spectrometry-based proteomics of the corresponding gel slices ([185]Supplementary Table 3). We identified all three subunits of BCKDH including E1 (BCKDHA and BCKDHB), E2 (DBT) and E3 (DLD). We also identified glutamate dehydrogenase 1 (GLUD1), which binds 5’-phosphate (PMP)-BCAT2 to regenerate pyridoxal 5’-phosphate (PLP)-BCAT2 through oxidative deamination of glutamate ^[186]13. We confirmed the significant enrichment of individual proteins using antibodies specific to each protein in streptavidin-purified mitochondrial complexes in BCAT2-mTD but not BCAT1-mTD or mTD-expressing cells ([187]Fig. 4d). As additional support of these findings, we performed co-immunoprecipitation (co-IP) experiments in WT and Bcat2^−/− kidney samples. We validated the interaction between BCAT2 and BCKDHA in WT kidney mitochondrial extracts in the presence of Val, pyridoxal phosphate (PLP), α-KG, thiamine diphosphate, and Coenzyme A to stimulate BCAT2 and BCKDH activity and metabolon formation ([188]Extended Data Fig. 6e). As an important control, the exclusion of these additives resulted in a loss of BCAT2-BCKDH interaction by co-IP ^[189]36. Together, our results provide evidence that BCAT2 associates with BCKDH within mitochondria. The BCAT2-BCKDH interactions enable the formation of multi-enzyme supramolecular complexes to facilitate BCAA catabolic reactions, consistent with the presence of a BCAA metabolon in living cells. Loss of BCAT2 impairs BCKDH-catalyzed BCKA oxidation The proximity labeling studies provided evidence for the presence of BCAT2-BCKDH complexes ([190]Fig. 4), but it remained unknown whether the formation of the multi-enzyme complexes is required for efficient BCAA oxidation in living cells. Metabolons function to facilitate metabolic flux, enzyme efficiency, and/or sequestration of toxic intermediates from competing pathways or damaging cells ^[191]37. Since BCKA accumulation causes oxidative stress, mitochondrial energy deficit, and cell apoptosis ^[192]38–[193]40, it is plausible that the formation of BCAT2-BCKDH complexes is required for efficient BCKA oxidation through substrate channeling. To determine whether the formation of BCAT2-BCKDH complexes is required for BCAA metabolism in vivo, we performed stable isotope tracing with uniformly ^13C-labeled α-ketoisovalerate ([U-^13C]-KIV) ([194]Fig. 5a). We quantified the production of 3-hydroxyisobutyrate (3-HIB_M+4), the only downstream intermediate without coenzyme A (CoA) conjugation in the BCAA catabolic pathways ^[195]19, as a readout of BCKDH activity. Additionally, we quantified Val_M+5 generated by [U-^13C]-KIV reamination and the fractional enrichment of succinate_M+3, M+2 and M+1 generated from the first, second, and third turns of the TCA cycle, respectively ([196]Fig. 5a). Compared to WT tissues, Bcat2^−/− had significantly reduced levels of Val_M+5 in most tissues, consistent with BCAT2 loss and no or minimal BCAT1 expression in these tissues ([197]Fig. 5b). More importantly, the levels of 3-HIB_M+4 and succinate_M+3,M+2,M+1, the downstream metabolites of KIV oxidation, were significantly lower in Bcat2^−/− relative to WT tissues ([198]Fig. 5b), suggesting that the BCKDH-catalyzed BCKA oxidation is impaired in the absence of BCAT2. Figure 5. BCAT2 Deficiency Impairs BCKDH-Catalyzed BCKA Oxidation. Figure 5. [199]Open in a new tab (a) Schematic of stable isotope tracing with [U-^13C]-KIV in WT and Bcat2^−/− mice. [U-^13C]-KIV can be reaminated to Val_M+5 by BCAT1 or BCAT2 in WT mice, or by BCAT1 in Bcat2^−/− mice. [U-^13C]-KIV is oxidized by BCKDH to generate HIB_M+4. Succinate_M+3,2,1 are generated by the first, second and third turns of the TCA cycle, respectively. (b) Quantification of Val_M+5 and HIB_M+4 and fraction labeling of succinate_M+3,2,1 are shown. Values were normalized to infusion rate and tissue weight relative to WT samples for the indicated genotypes and tissues with the number (N) of independent samples shown. Results are mean ± SEM and analyzed by two-sided unpaired t-test. (c) Schematic of the radiochemical measurement of total BCKDH activity. BCKDH oxidizes [^14C]-KIV, releasing ^14CO[2] molecules which are trapped in the NaOH soaked wick and quantified by a scintillation counter. (d) Total BCKDH activity in heart, kidney and liver from WT or Bcat2^−/− mice with the number (N) of independent samples shown. Results are mean ± SD and analyzed by two-sided unpaired t-test. (e) Loss of BCAT2 impairs BCKDH activity to cause BCKA accumulation in Bcat2^−/− mice. Since BCKA accumulation was not due to changes in expression or phosphorylation of BCKDH ([200]Extended Data Fig. 6a,[201]b), we next measured total BCKDH activity using a radiochemical assay with ^14C-labeled KIV ^[202]41,[203]42. We focused on liver, heart and kidney due to the abundance of cells for rapid processing and assay sensitivity. Upon BCKDH-catalyzed oxidative decarboxylation, the ^14C-labelled CO[2] molecules were released and collected by a CO[2] trap, followed by quantification using a scintillation counter ([204]Fig. 5c). BCKDH activity in Bcat2^−/− samples accounted for only 42%, 23% and 41% of activity in WT liver, heart and kidney tissues, respectively ([205]Fig. 5d). Therefore, despite the lower levels of phosphorylated and inactive BCKDH in Bcat2^−/− tissues ([206]Extended Data Fig. 6b), the enzymatic activity of BCKDH in catalyzing BCKA oxidation was significantly impaired upon loss of BCAT2. Collectively, these studies demonstrate that the integrity of the BCAT2-BCKDH multi-enzyme complexes is required for efficient BCKA oxidation to maintain BCAA metabolic homeostasis in a living organism ([207]Fig. 5e). Ectopic BCAT1 ameliorates BCAT2-deficiency-induced phenotype The identification of BCAT2-BCKDH complex formation raised the question about the role of BCAA metabolon in regulating system-wide BCAA homeostasis. Given the deleterious effects of elevated BCKAs on mammalian cells, animal models and MSUD patients, we reasoned that one of the primary functions of BCAA metabolon is to prevent BCKA accumulation. Therefore, lowering BCKAs may ameliorate defective BCAA metabolons caused by BCAT2 deficiency in Bcat2^−/− mice or by BCKDH mutations in MSUD patients. Of note, BCAT1 is physically segregated from BCAT2 due to subcellular location and tissue-restricted expression ^[208]1,[209]6,[210]7,[211]14 ([212]Fig. 1b,[213]c). A preference for BCAT1-catalyzed BCKA reamination was observed under various physiological and pathological conditions ^[214]22,[215]43,[216]44. Based on these findings, we reasoned that ectopic BCAT1 expression may alleviate the phenotypes caused by BCKA accumulation in Bcat2^−/− mice. To this end, we generated a Bcat1 overexpression mouse model by CRISPR-mediated site-specific knockin (KI) of mouse Bcat1 cDNA with an HA tag under the loxP-STOP-loxP (LSL) cassette and a CAG promoter into the Rosa26 locus. By crossing the Bcat1^LSL KI founder mice with the CMV-Cre transgenic mice ^[217]45 to delete LSL in all tissues, we obtained BCAT1 constitutive overexpression (Bcat1^OE) mice ([218]Fig. 6a) and confirmed the cytosolic BCAT1 overexpression by Western blot analysis ([219]Extended Data Fig. 7a,[220]b). Figure 6. Ectopic BCAT1 Expression Restores BCAT2-Deficiency-Induced Metabolic Defects. Figure 6. [221]Open in a new tab (a) Schematic of the Bcat1 knockin (Bcat1^LSL KI) and constitutive overexpression (Bcat1^OE) mouse models. (b) Body weight growth curves of the indicated genotypes. Results are mean ± SD (N = 4 WT, 6 Bcat1^OE, 6 Bcat2^−/−, and 5 Bcat1^OE;Bcat2^−/− mice) and analyzed by two-way ANOVA mixed-effects analysis with multiple comparisons. (c) Kaplan-Meier survival curves of WT, Bcat1^OE, Bcat2^−/−, and Bcat1^OE;Bcat2^−/− mice (N = 8 WT, 8 Bcat1^OE, 5 Bcat2^−/−, and 8 Bcat1^OE;Bcat2^−/− mice). P values by a log-rank Mantel-Cox test. (d) BCAA and BCKA levels in Bcat1^OE, Bcat2^−/−, and Bcat1^OE;Bcat2^−/− mouse tissues. Heatmap shows the log2 fold changes of the indicated metabolites in each mutant tissue relative to WT controls (N = 3 WT, 3 Bcat1^OE, 3 Bcat2^−/−, and 4 Bcat1^OE;Bcat2^−/− mice) which were identified by two-sided unpaired t-test in the indicated tissues. (e) Quantification of Val_M+5 and HIB_M+4 and fractional labeling of succinate_M+3,2,1 are shown. Values were normalized to infusion rate and tissue weight relative to WT in the indicated genotypes and tissues with the number (N) of independent samples shown. Results are ± SEM and analyzed by one-way ANOVA. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. n.s. not significant. The exact P values are provided in [222]Supplementary Table 5. (f) Total BCKDH activity in liver, heart and kidney with the number (N) of independent samples shown. Results are mean ± SD and analyzed by two-way ANOVA. (g) Quantification of puromycin incorporation in mice of the indicated genotypes relative to WT samples with the number (N) of independent samples shown. Results are mean ± SD and analyzed by two-sided unpaired t-test. (h) [^13C] Val_M+5 incorporation into protein in the indicated tissues and genotypes relative to WT samples with the number (N) of independent samples shown. Values were normalized to infusion rate and tissue weight. Results are mean ± SD and analyzed by two-way ANOVA. (i) Quantification of urinary 3-methyl-histidine and 3-methyl-histidine/creatinine ratio in the indicated genotypes with the number (N) of independent samples shown. Results are mean ± SD and analyzed by ordinary one-way ANOVA. (j) Schematic of the working model. Strikingly, while BCAT1 overexpression alone had no significant effect on body weight and survival, BCAT1 overexpression together with BCAT2 KO (Bcat1^OE;Bcat2^−/−) rescued BCAT2-deficiency-induced phenotypes and early mortality ([223]Fig. 6b,[224]c). To determine the effect on systemic BCAA metabolism, we examined metabolomic profiles by comparing WT, Bcat1^OE, Bcat2^−/−, and Bcat1^OE;Bcat2^−/− mice ([225]Fig. 6d,[226]e). Of note, BCAT1 overexpression normalized BCAA and BCKA accumulations to WT levels in major tissues of Bcat1^OE;Bcat2^−/− mice, whereas BCAT1 overexpression alone had minimal effects on BCAA and BCKA levels in most tissues ([227]Fig. 6d). By in vivo isotope tracing with [U-^13C]-KIV, we observed that the levels of labeled 3-HIB_M+4 and succinate_M+3,M+2,M+1 were comparable between Bcat2^−/− and Bcat1^OE;Bcat2^−/− tissues, which were significantly lower than those in WT or Bcat1^OE tissues ([228]Fig. 6e). It is important to note that the blood enrichment of labeled KIV were comparable among genotypes and the plasma levels of Val and KIV were not significantly altered by KIV infusion ([229]Extended Data Fig. 7c,[230]d). The similar labeling patterns in Bcat2^−/− and Bcat1^OE;Bcat2^−/− tissues indicate that the impaired BCKDH-catalyzed BCKA oxidation caused by BCAT2 deficiency was not restored by BCAT1 overexpression. Consistent with these findings, total BCKDH activity remained lower in Bcat2^−/− and Bcat1^OE;Bcat2^−/− relative to WT tissues ([231]Fig. 6f). As systemic BCAT activity may be influenced by nitrogen acceptor or donor availability, we measured the glutamate/α-KG ratio in major tissues. Bcat2^−/− mice had significant decreases in the glutamate/α-KG ratios in heart, kidney, liver and spleen, consistent with the loss of BCAT2 activity ([232]Extended Data Fig. 7e). The glutamate/α-KG ratios were increased in Bcat1^OE;Bcat2^−/− tissues compared to Bcat2^−/− tissues, consistent with BCAT1-catalyzed BCKA reamination. In addition, BCKDH activity can be regulated by other mechanisms such as NAD^+/NADH ratios and insulin levels. However, we found no alterations to the NAD^+/NADH ratios in heart, kidney or pancreas, suggesting that NAD^+ is likely not a limiting factor for BCKDH activity in these tissues. We also found a trend of increased plasma insulin levels that cannot explain the decreased BCKDH activity in Bcat1^OE;Bcat2^−/− mice ([233]Extended Data Fig. 7f,[234]g). Taken together, our studies using genetic mouse models provide evidence that ectopic BCAT1 expression is sufficient to alleviate BCAA and BCKA accumulations caused by BCAT2 deficiency. BCKA reamination modulates BCAA transport and turnover It is important to note that, despite BCAT1 expression in Bcat1^OE;Bcat2^−/− mice, the levels of labeled Val_M+5 remained lower in heart, kidney, liver and pancreas relative to WT tissues, whereas there was a significant increase in labeled Val_M+5 in the plasma ([235]Fig. 6e). BCAT1 expression or activity associates with increased BCAA export and/or protein synthesis ^[236]22,[237]44 that may contribute to increased BCAA consumption, resulting in the observed tracing profiles in Bcat1^OE and Bcat1^OE;Bcat2^−/− mice. Consistent with this notion, the levels of the plasma membrane amino acid transporter LAT1 (encoded by Slc7a5) were significantly increased by BCAT1 overexpression, whereas the levels of BCKDHA, P-BCKDHA, BCKDH and PPM1K remained unchanged in major tissues ([238]Extended Data Fig. 7b and [239]8a,[240]b). To further explore the underlying basis for the observed metabolic alterations, we examined the activation of mTORC1 and its downstream targets including the phosphorylated 4EBP1 (P-4EBP1) and p70 S6 kinase (P-p70S6K) in different genotypes and conditions. We first noted that BCAT2 KO led to enhanced mTORC1 activation as indicated by significantly increased P-4EBP1 and P-p70S6K levels relative to WT tissues ([241]Extended Data Fig. 8c,[242]d). The enhanced mTORC1 activation correlated with the depletion of some but all amino acids in Bcat2^−/− tissues ([243]Extended Data Fig. 8e). Moreover, Bcat2^−/− mice fed with BCAA choice diet for 4 weeks led to decreased P-4EBP1 relative to Bcat2^−/− mice fed with normal diet ([244]Extended Fig. 8c). Bcat1^OE alone led to increased mTORC1 activity and Bcat1^OE;Bcat2^−/− tissues displayed comparable mTORC1 activity as Bcat2^−/− tissues ([245]Extended Fig. 8d). We next quantified protein synthesis using the surface sensing of translation (SUnSET) assay ^[246]46, which measures the incorporation of puromycin into newly synthesized proteins. We noted significantly increased protein synthesis in Bcat1^OE and Bcat1^OE;Bcat2^−/− relative to WT tissues, with the highest increases observed in Bcat1^OE;Bcat2^−/− tissues ([247]Fig. 6g; [248]Extended Data Fig. 8f). Consistent with these findings, the incorporation of labeled Val_M+5 into proteins was significantly increased in most Bcat1^OE and Bcat1^OE;Bcat2^−/− tissues ([249]Fig. 6h), further contributing to the decrease in Val_M+5 labeling within tissues. Lastly, we measured protein degradation by urinary excretion of 3-methylhistidine and creatinine, commonly used as an index of skeletal muscle protein breakdown ^[250]47. We found that BCAT2 KO significantly increased protein degradation ([251]Fig. 6i), consistent with a BCAT2 deficiency-induced futile protein turnover cycle ^[252]30. More importantly, the levels of excreted 3-methylhistidine and creatinine were comparable between WT and Bcat1^OE;Bcat2^−/− mice, indicating that BCAT1 overexpression ameliorated BCAT2 KO-induced protein degradation. Taken together, these findings demonstrate that BCAT2 deficiency in mice led to enhanced mTORC1 activation, protein synthesis, and protein degradation, whereas ectopic BCAT1 expression reverses BCAT2-deficiency-induced metabolic phenotypes through enhanced BCKA reamination, BCAA export, and rescued protein turnover ([253]Fig. 6j). Therefore, through stringent genetic studies and systemic analysis of in vivo metabolism, we provide evidence that the metabolic defects caused by impaired BCAA metabolon (e.g. BCAT2 deficiency or BCKDH mutations) can be alleviated through modulating BCAT1 activity to control BCAA homeostasis in vivo. These findings not only establish in vivo function for BCAA metabolon in a living organism, but also provide insights into rational targeting of dysfunctional BCAA metabolism in genetic or metabolic disorders. Discussion Functional and molecular features of BCAT isozymes Enzyme heterogeneity is common across species, although it remains difficult to determine whether and how homologous enzymes display redundant or distinct biological functions. Isozymes, also known as isoenzymes, were first described as different enzyme variants catalyzing the same chemical reaction in the same individual ^[254]48. Many enzyme complexes of intracellular metabolism contain isozymes, such as lactate dehydrogenase (LDH), creatine phosphokinase (CPK), cytochrome P450, and alkaline phosphatase ^[255]49. However, the functional and molecular properties that discriminate these isozymes, in particular their roles in controlling metabolic homeostasis in an intact organism, remain poorly understood. BCAT1 and BCAT2 isozymes have several unique properties. First, they are largely segregated into distinct subcellular compartments (cytosol versus mitochondria) and organs (tissue-restricted versus ubiquitously expressed) ^[256]1,[257]6,[258]7,[259]14. Second, BCAT-catalyzed reaction is rapid and likely in equilibrium in most cases due to a low free energy change ^[260]19. Third, BCAT-catalyzed BCAA transamination requires other metabolites including the nitrogen acceptor and donor (α-KG and Glu). Thus, the concentrations of these metabolites in different subcellular compartments and tissues likely influence the ‘reversibility’ and kinetics of BCAT-catalyzed reaction. An important example is the formation of ‘BCAA shuttle’ between astrocytes and neurons for glutamate metabolism in the brain ^[261]5. Lastly, due to extensive inter-organ shuttling of BCAAs and BCKAs ^[262]1,[263]2, the use of purified proteins, homogenized tissues or cultured cells is not ideal for studying BCAA metabolic pathways. Furthermore, the use of WT animals or healthy humans does not provide sufficient information about the consequences of defective metabolic enzymes or their interactions. As such, whether and how the unique properties of BCAT isozymes contribute to system-wide metabolism in an intact organism has remained largely unknown. To address these challenges, we generated mouse models deficient in BCAT1 or BCAT2 and examined the metabolomic alterations caused by BCAT loss across major tissue types. We further characterized BCAT-dependent metabolic processes using orthogonal approaches including dietary perturbation, isotope tracing, enzymatic function, and genetic reconstitution experiments. Our studies uncover a number of molecular features and mechanistic insights that help explain the differential metabolic roles of BCAT isozymes in a living organism. Specifically, loss of BCAT2 results in systemic accumulation of BCAAs and BCKAs, causing morbidity and mortality resembling human MSUD that can be rescued by restricting BCAA intake. By contrast, BCAT1 is dispensable for mouse development, whole-body metabolism, and overall health. In addition, BCAT2 physically interacts with the downstream BCKDH to form multi-enzyme complexes within mitochondria, consistent with the presence of the BCAA metabolon for efficient BCAA catabolism in living cells. Loss of BCAT2 affects BCKA oxidation due to impaired BCKDH activity, resulting in the accumulation of both the substrates (BCAAs) and the products (BCKAs) of BCAT2-catalyzed reaction. These findings further support the relatively minor roles for BCAT1 in regulating systemic BCAA oxidation under physiological conditions, as the tissue-restricted BCAT1 activity is not sufficient to prevent BCAT2-deficiency-induced metabolic defects. Our findings support a model for the presence of BCAT2-BCKDH-associated metabolon formation in living cells, whereas disabling the BCAA metabolon by BCAT2 deficiency or BCKDH mutations leads to MSUD-associated metabolic phenotypes such as BCKA accumulation. To test this, we generated a BCAT1 knockin overexpression mouse model and found that ectopic BCAT1 expression was sufficient to mitigate BCAT2-deficiency-induced growth defects and mortality. Importantly, the amelioration of BCAT2-deficiency-induced phenotypes was not due to restored BCKDH activity, but instead caused by the BCAT1-catalyzed BCKA reamination and the modulation of BCAA transport and protein turnover. Hence, although BCAT1 has little contribution to systemic BCAA metabolism under physiological conditions likely due to its tissue-restricted expression, ectopic BCAT1 expression is capable of modulating whole-body BCAA homeostasis in the context of defective BCAA metabolons. Therefore, by systematic analysis of metabolic consequences caused by loss of BCAT isozymes in living organisms, our studies provide new insights into the fundamental mechanisms that regulate BCAA homeostasis in physiology and pathologies. Our findings reveal that the coordination of compartmentalized and tissue-specific BCAT isozymes together with BCAT2-BCKDH-mediated metabolon formation is required for fine-tuning BCAA metabolism to meet the needs of a given tissue or developmental stage. Given that dysregulated BCAA metabolism is associated with various disorders such as MSUD, heart diseases, diabetes, and cancer, our studies provide the proof-of-principle evidence that leveraging critical nodes of the BCAA metabolic pathways may alleviate morbidities caused by enzyme deficiency or mutations in genetic or metabolic disorders. Function of BCAA metabolon in health and disease Dynamic protein-protein interactions that bring sequential metabolic enzymes together have been theorized since the 1970s ^[264]37. The term ‘metabolon’ was later introduced following the discoveries of enzyme compartmentalization within the TCA cycle ^[265]10,[266]11,[267]50. Subsequent advances in cell-based techniques led to the identification of the ‘purinosome’, a putative multi-enzyme complex for de novo purine biosynthesis within living cells ^[268]51–[269]53. While these discoveries laid the foundation for understanding the biochemical basis of metabolon formation, the use of in vitro assays or cell culture-based approaches does not provide strong evidence for establishing the functional relevance of metabolon formation in living organisms. As such, the proposed roles for various metabolons have not been rigorously tested by genetic studies in vivo, and the effect of defective metabolon formation on whole-body metabolic homeostasis is not known. By combining proximity labeling proteomics with metabolic analysis of genetic mouse models, our studies establish a role for the formation of BCAT2-BCKDH multi-enzyme complexes in living cells. We show that impaired BCAA metabolon in BCAT2-deficent mice leads to similar BCKA accumulations as those caused by BCKDH mutations in human MSUD, indicating that the integrity of the BCAT2-BCKDH complexes is required for maintaining BCAA homeostasis in vivo. The identification of in vivo BCAA metabolon in living cells raises new questions about their function in regulating cellular metabolism. At the conceptual level, there are a number of benefits to metabolon formation: 1) the increase local concentrations of enzymes and metabolites may improve channeling of intermediates into specific pathways; 2) the channeling of metabolites between sequential enzymes provides increased control over pathway fluxes, thus increasing the efficiency of chemical reactions; 3) the closed compartments can limit diffusion to sequester toxic or reactive intermediates; and 4) the formation of metabolon can also counteract futile cycles of metabolites due to reversible reactions. This is likely relevant to BCAT2-catalyzed BCAA transamination given its fast kinetics and reversibility, thus the formation of BCAA metabolon may provide a means of coordinating the reversible transamination step with the downstream flux-generating, irreversible BCAA oxidation. In addition, various nutrient or hormone signals may affect metabolon formation to allow integration of specific metabolic needs by partitioning BCAAs into distinct cellular compartments and/or catabolic pathways. It is therefore important to investigate in future studies whether any or all these features are regulated by the formation of BCAA metabolon, and how dysregulation of these processes contributes to cellular defects associated with dysfunctional BCAA metabolism. Leveraging metabolons for disease management and therapy Establishing evidence for the in vivo function of metabolon formation not only advances our understanding of the fundamentals of BCAA metabolism in mammalian physiology, but also provides new insights into potential therapeutic opportunities. Dysregulated BCAA metabolism is associated with a number of disorders although various cellular or metabolic mechanisms have been proposed. Our findings help explain some of these findings by providing evidence for the formation of BCAA metabolon in regulating BCAA homeostasis in vivo. In particular, the expression and/or activity of BCAT2 is required for the optimal activity of BCKDH-catalyzed BCAA oxidation. Aberrant BCAT1 activity may also contribute to BCAA homeostasis under physiological and pathological conditions, as observed in our genetic complementation studies. Given the relevance of BCAA metabolism to human diseases, the BCAT isozymes and BCAA metabolon may represent pharmacological opportunities for BCAA-targeted therapeutic interventions. BCAT2 KO mice displayed clinical and metabolic features resembling human MSUD and, like MSUD, responded to BCAA-restricted diet. Although rare and less investigated, BCAT2 deficiency caused by mutations in the BCAT2 gene are also described in humans ^[270]54,[271]55. The phenotypic manifestation of human BCAT2 deficiency ranges from largely asymptomatic to global developmental delay and microcephaly, and associates with increased plasma BCAAs but not BCKAs in most cases ^[272]55. Unlike BCAT2 KO mice, the human variants resulted in reduced but not absence of BCAT2 activity ^[273]54,[274]55. Therefore, the observed phenotypic and metabolic defects in our studies could be due to the variable residual enzyme activities among the patients, compared to the complete loss of BCAT2 in our knockout mice, resulting in the most severe disease. Given that only plasma levels of BCAAs and BCKAs were measured in human cases, it remains unclear whether the human BCAT2 variants may also cause BCKA accumulation in tissues. Lastly, it remains unknown whether the human BCAT2 variants affect BCAT2-BCKDH interaction and BCKDH activity that contribute to BCKA accumulation. Our findings uncover that ectopic BCAT1 expression is sufficient to normalize BCAA and BCKA accumulation caused by BCAT2 deficiency, suggesting that BCAT1-mediated BCKA reamination may be an effective approach to mitigate pathological BCKA accumulation underlying MSUD and other disorders associated with dysregulated BCAA catabolism. It remains unknown whether BCAT1 activity in one or multiple tissues was required for restoring the metabolic defects. BCKDH activity in liver is known to be critical for the removal of excessive BCKAs, and liver transplantation is an effective treatment for MSUD ^[275]56,[276]57. Based on these findings, the development of gene therapy approaches is underway to repair mutant BCKDH by adeno-associated virus or mRNA-mediated gene replacement ^[277]58,[278]59. As more than 150 distinct mutations affecting different BCKDH subunits can lead to MSUD with different severity and etiology ^[279]25,[280]26, these strategies require tissue-directed delivery of specific BCKDH gene products for the corresponding mutations found in each affected individual. Our findings that ectopic BCAT1 expression is sufficient to ameliorate aspects of MSUD-associated pathologies suggest a strategy for mitigation of dysfunctional BCAA metabolism in a broad spectrum of MSUD conditions. Hence, this work provides the rationale for further investigation of BCAT isozymes and their interactions with BCAA metabolon for the development of more effective and generally applicable strategies to target dysregulated BCAA metabolism in genetic and metabolic disorders. Methods Mice. BCAT1 and BCAT2 constitutive knockout (KO) mice were generated by the Easi-CRISPR-based targeting, respectively ^[281]60. Briefly, sgRNAs flanking the exon 6 of Bcat1 gene or exons 4 to 6 of Bcat2 gene were selected by CHOPCHOP ([282]http://chopchop.cbu.uib.no/) and gRNA checker ([283]https://www.idtdna.com/site/order/designtool/index/CRISPR_SEQUENC E). BCAT1 constitutive knockin (KI) overexpression (OE) mouse model was generated by site-specific integration of N-terminal HA-tagged mouse Bcat1 cDNA into the Rosa26 locus following sgRNA-mediated cleavage. CRISPR/Cas9 crRNAs and targeting sgRNAs were synthesized by Integrated DNA Technologies. N-terminal HA-tagged mouse Bcat1 cDNA (TransOMIC, Cat. # [284]BC053706) was cloned into a modified targeting vector pR26 CAG AsiSI/MluI (Addgene #74286) by removing the puromycin cassette. After microinjection into mouse zygotes, the founder animals were screened for the KO or KI allele by genotyping PCR and confirmed by Sanger sequencing. The sequences for sgRNAs and primers are listed in [285]Supplementary Table 4. Both male and female mice were used unless otherwise specified. All mice were housed under a 12-hour light-dark cycle, 75°F and 35% humidity in the Animal Resource Center at the University of Texas Southwestern Medical Center (UTSW). All mouse experiments were performed under protocols approved by the Institutional Animal Care and Use Committee of UTSW. Cells and cell culture. To generate BCAT1 and BCAT2 double knockout (DKO) and reconstituted cell lines, HEK293T cells (ATCC, Cat# CRL-3216) were transduced with pSpCas9(BB)-2A-GFP (pX458; Addgene #48138) vector co-expressing Cas9 and sgRNAs flanking exons 3 to 6 of BCAT1 or exons 4 to 6 of BCAT2, respectively. Single cell-derived clonal lines were screened and confirmed for BCAT1 and BCAT2 DKO by genotyping PCR and Western blot analyses. The cells were then transduced with lentiviral BCAT2-mTD, BCAT1-mTD, or mTD alone under the doxycycline (Dox)-inducible promoter. Lentiviruses were packaged in HEK293T cells as previously described ^[286]61. Specifically, 2 μg of psPAX2, 1 μg of pMD2.G and 5 μg lentiviral vectors were transfected into HEK293T cells seeded in a 10 cm dish. Lentiviruses were harvested by collecting the supernatant 48 to 72 hours post-transfection, followed by precipitation using PEG-it (System Biosciences) following the manufacturer’s protocol. Plasmids. To generate Dox-inducible BCAT2 lentiviral expressing vectors, human BCAT2 cDNA was fused to the 19 amino-acid linker containing glycine and serine (GS) and miniTurboID (mTD) sequences ^[287]35, together with a C-terminal V5 epitope tag (BCAT2-mTD), was cloned to the BamHI and EcoRI sites of the pLVX-TRE3G-IRES-Puro vector (Clontech #631362) using the In-Fusion HD cloning kit (Takara). As controls, we also generated Dox-inducible BCAT1 lentiviral expressing vector by fusing BCAT1 cDNA with the mitochondria targeting signal sequences from BCAT2 at N-terminus, 19 amino-acid GS linker, mTD and a C-terminal V5 tag (BCAT1-mTD), or mTD alone with N-terminal mitochondria targeting signal sequences and C-terminal V5 (mTD). To generate sgRNA vectors for knockout of BCAT1 and/or BCAT2, sgRNA sequences were cloned into the pX458 vector (Addgene #48138). sgRNAs were designed to minimize off-targets based on filtering tools ([288]https://zlab.bio/guide-design-resources). Oligonucleotides were synthesized, annealed, and cloned into vectors as previously described ^[289]61. The sequences for sgRNAs and primers are listed in [290]Supplementary Table 4. BCAT1 and BCAT2 gene expression analysis. The mRNA expression values of human and mouse BCAT1 and BCAT2 across different tissue types were obtained from The Genotype-Tissue Expression (GTEx) Portal on 06/24/2020. The GTeX project was supported by the Common Fund of the Office of the Director of the National Institutes of Health, and by NCI, NHGRI, NHLBI, NIDA, NIMH, and NINDS. The mRNA expression of all genes associated with BCAA metabolism in seven human and mouse organs (brain, cerebellum, heart, kidney, liver, ovary and testis) across developmental time points from early organogenesis to adulthood was performed using the database for Evo-devo mammalian organs ([291]https://apps.kaessmannlab.org/evodevoapp/) ^[292]15. The expression of Bcat2, Bckdha, Bckdhb, Bckdk and Ppm1k mRNA and protein fractionated liver cell types including hepatocytes (HC), hepatic stellate cells (HSC), Kupffer cells (KC), and liver sinusoidal endothelial cells (LSEC) were obtained from previous studies ^[293]21 by RNA-seq and proteomics, respectively. Metabolomics. Tissue samples were collected and flash frozen in liquid nitrogen. For extraction of metabolites, 50 mg of tissue was added to 1 ml of ice-cold methanol/water 80% (vol/vol) and homogenized. After three freeze/thaw cycles, samples were vortexed and 200 μl of the sample was transferred to a new tube containing 800 μl 80% methanol. Samples were centrifuged at 20,000xg for 30 min at 4°C. Then 800 μl of the metabolite containing supernatant was transferred to a new tube and placed in a SpeedVac to dry the supernatant. Data acquisition was performed by reverse-phase chromatography on a 1290 UHPLC liquid chromatography (LC) system interfaced to a high-resolution mass spectrometry (HRMS) 6550 iFunnel Q-TOF mass spectrometer (MS) (Agilent Technologies). The MS was operated in both positive and negative (ESI+ and ESI−) modes. Analytes were separated on an Acquity UPLC HSS T3 column (1.8 μm, 2.1 × 150 mm, Waters). The column was kept at room temperature. Mobile phase A composition was 0.1% formic acid in water and mobile phase B composition was 0.1% formic acid in 100% acetonitrile. The LC gradient was 0 min: 1% B; 5 min: 5% B; 15 min: 99%; 23 min: 99%; 24 min: 1%; 25 min: 1%. The flow rate was 250 μL/min−1. The sample injection volume was 5 μL. ESI source conditions were set as follows: dry gas temperature 225°C and flow 18 L/min−1, fragmentor voltage 175 V, sheath gas temperature 350°C and flow 12 L/min−1, nozzle voltage 500 V, and capillary voltage +3500 V in positive mode and −3500 V in negative. The instrument was set to acquire over the full m/z range of 40–1700 in both modes, with the MS acquisition rate of 1 spectrum per second in profile format. Raw data files were processed using Profinder B.08.00 SP3 software (Agilent Technologies) with an in-house database containing retention time and accurate mass information on 600 standards from Mass Spectrometry Metabolite Library (IROA Technologies) analyzed under the same conditions. The in-house database matching parameters were: mass tolerance 10 ppm; retention time tolerance 0.5 min. Peak integration results were manually curated in Profinder for improved consistency. For some compounds, when standards were analyzed by the same experimental setup, more than one chromatographic peak with similar qualification but different intensity were detected, most likely due to interaction with the column and high pressure; in these cases, each chromatographic peak is denoted by a suffix (−1 or −2, etc.) in the [294]supplemental tables. The peak area for each detected metabolite was normalized against the total ion count of each sample. The normalized areas were used as variables for the multivariate and statistical analyses. Quantification of 3-Methyl-Histidine and Creatinine. Urine samples were collected and immediately added to 1 ml of ice-cold methanol/water 80% (vol/vol) for extraction of metabolites. After three freeze/thaw cycles, samples were vortexed and 200 μl of the sample was transferred to a new tube containing 800 μl 80% methanol. Samples were centrifuged at 20,000xg for 30 min at 4°C. Then 800 μl of the metabolite containing supernatant was transferred to a new tube and placed in a SpeedVac to dry the supernatant. Mass spectrometry was performed on an AB SCIEX 5500 QTRAP liquid chromatography/mass spectrometer (Applied Biosystems SCIEX), equipped with a vacuum degasser, a quaternary pump, an autosampler, a thermostatted column compartment, and a triple quadrupole/ion-trap mass spectrometer with electrospray ionization interface, controlled by AB SCIEX Analyst 1.6.1 Software. SeQuant^® ZIC^®-pHILIC (150mm×2.1mm, 5μm) PEEK coated column was used. Solvents for the mobile phase were 10mM NH[4]Ac in H[2]O (pH = 9.8, adjusted with conc. NH[4]OH) (A) and acetonitrile (B). The gradient elution was: 0–15 min, linear gradient 90–30% B; 15–18 min, 30% B, 18–19 min, linear gradient 30–90% B; and finally reconditioning it for 9 min using 90% B. The flow-rate was 0.25 ml/min, and injection volume was 10 μL. Columns were operated at 35°C. Declustering potential (DP), collision energy (CE) and Collision Cell Exit Potential (CXP) were optimized for each metabolite by direct infusion of reference standards using a syringe pump prior to sample analysis. Samples were quantified against standard solutions of 3-methyl-histidine (Methyl-D3) and creatinine (N-Methyl-D3) with concentrations escalated from 1ng to 1000ng. The MRMs used are listed as follows: 3-methyl-L-histidine Q1/Q3, 170/126 (m+0), 173/129 (m+3, for standard curve), CE: 16; creatinine Q1/Q3, 114/86 (m+0), 117/89 (m+3, for standard curve), CE: 15. The following chemicals were used: Tau-Methyl-L-Histidine (Methyl-D3) (Cambridge Isotope Laboratories, Cat# DLM-2949-PK) and Creatinine (N-Methyl-D3) (Cambridge Isotope Laboratories, Cat# DLM-3653-PK). Isotope tracing. In vivo infusion experiments were performed using 6- to 8-week old mice maintained on the BCAA choice diet. Mice were anaesthetized using ketamine and xylazine (120 mg/kg and 16 mg/kg, intraperitoneally) and maintained under anesthesia with subsequent doses of ketamine/xylazine as needed. A 25-guage catheter was placed in the tail vein. A total dose of 0.05 g/kg [U-^13C]KIV (Cambridge Isotope Laboratories, #CLM-4418) was dissolved in 750 μl saline, administered as a bolus of 125 μl/min for 1 min, followed by an infusion rate of 2.5 μl/min for 2 hours. Retro-orbital blood draws were taken every 30 min to determine isotope enrichment in plasma. After 2 hours, mice were sacrificed, and tissues were collected and immediately flash frozen. Mass spectrometry was performed on iFunnel QTOF (described above) or on an AB SCIEX 5500 QTRAP liquid chromatography/mass spectrometer (Applied Biosystems SCIEX), equipped with a vacuum degasser, a quaternary pump, an autosampler, a thermostatted column compartment, and a triple quadrupole/ion-trap mass spectrometer with electrospray ionization interface, controlled by AB SCIEX Analyst 1.6.1 Software. A Cogent Bidentate C18 HPLC column (150 × 2.1 mm) was used at 35°C. Mobile phase A composition was 0.03% formic acid in water and mobile phase B composition was 0.03% formic acid in acetonitrile. The gradient elution was a linear gradient in 0–4 min, 0–20% B, then the column was washed with 100% B for 2min before reconditioning it for 4 min using 0% B. The flow rate was 0.8 ml/min, and the injection volume was 20 μl. The mass spectrometer used with an electrospray ionization (ESI) source in multiple reaction monitoring (MRM) mode. Declustering potential (DP), collision energy (CE) and Collision Cell Exit Potential (CXP) were optimized for each metabolite by direct infusion of reference standards using a syringe pump prior to sample analysis. For quantification of KIV, HIB and Val, samples were homogenized in 80% methanol containing internal standards (L-Valine-D8, Cambridge Isotope Laboratories, #DLM-311-PK, final concentration 0.125 μmol/L; 3-HB-C2, final concentration 1 μmol/L, Santa Cruz Biotechnology, #sc-236902; and C2-KIV, final concentration 1 μmol/l, Cambridge Isotope Laboratories, #CLM-6821) and quantified against standard solutions with concentrations escalated from 0.0763 μmol/L to 0.625 mmol/L. The MRMs used are listed as follows: KIV Q1/Q3 (115/71 (m+0), 120/75 (m+5), 117/73 (m+2, IS), CE:−10); HIB Q1/Q3 (103/73 (m+0), 107/76 (m+4), CE:−13); 3-HB Q1/Q3 (103/59 (m+0), 105/60 (m+2, IS), CE:−13); Val Q1/Q3 (118/55 (m+0); 123/59 (m+5), 126/62 (m+8, IS), CE: 27), Succinate Q1/Q3, 117/99 (m+0), 118/100 (m+1), 119/101 (m+2), and 120/102 (m+3), CE: −18. The following chemicals were used: [U-^13C]-KIV (Cambridge Isotope Laboratories, Cat# CLM-4418), L-Valine-D8 (Cambridge Isotope Laboratories, Cat# DLM-311-PK), 3-HB-C2 (Santa Cruz Biotechnology, Cat# sc-236902), and C2-KIV (Cambridge Isotope Laboratories, Cat# CLM-6821). BCAA diet restriction. At weaning, animals were placed on a BCAA choice diet, with access to chow containing normal levels of BCAAs (12 g/kg leucine, 8 g/kg valine, and 8 g/kg isoleucine) (Research Diets, #A06030501) and chow containing low levels of BCAAs (0.278 g/kg leucine, 0.205 g/kg valine, and 0.205 g/kg isoleucine) (Research Diets, #A06030502). Mouse weight and food intake was measured weekly for 12 weeks, and mouse survival was monitored for 30 weeks. BCKDH activity assay. BCKDH activity was measured as previously described with modifications ^[295]42. Briefly, flash frozen tissue samples were homogenized in 1 ml of ice-cold buffer 1 (30 mM KPi, pH 7.5, 3 mM EDTA, 5 mM DTT, 3% fetal bovine serum, 5% Triton X-100, protease inhibitor tablets (Thermo-Fisher, #A32963). Samples were spun at 25,000xg for 10 min at 4°C and supernatants were transferred to new tubes. Heart, kidney and liver samples were diluted 1:4 in buffer 2 (50 mM HEPES, pH 7.5, 0.5 mM DTT, 0.1% Triton X-100, 3% fetal bovine serum and protease inhibitor tablets) and Lambda Protein Phosphatase (LPP) (New England BioLabs, #P0753) and incubated at 30°C for 30 min. Fifty μl of sample was then added to 295 μl assay buffer (30 mM KPi, pH 7.5, 0.4 mM CoA, 3 mM NAD^+, 5% fetal bovine serum, 2 mM thiamine diphosphate, 2 mM MgCl[2] and 42 μg/ml human E3). To initiate the reaction, 25 μl [1-^14C] KIV (7.4 mM cold) substrate added to each well along with a 2M NaOH soaked filter wick. The plate was then sealed with a clear mylar adhesive film and incubated at 37°C for 40 min. Plates were cooled for 10 min before 50 μl of a 20% TCA solution was added to stop the reaction. The samples were incubated for another 45 min at 37°C. ^14CO[2] molecules trapped on the NaOH soaked filter wicks were then counted in a scintillation counter. Proximity labeling and proteomics. To identify BCAT1- and BCAT2-interacting protein complexes in living cells, the BCAT1 and BCAT2 DKO or reconstituted HEK293T cells were induced with 10 ng/ml Dox for BCAT1-mTD and BCAT2-mTD and 300 ng/ml Dox for mTD for three days. On the last day, cells were treated with 200 mM biotin (Sigma-Aldrich, #B4639) for 30 min or 50 mM for one hour, washed with ice-cold 1x PBS, and fractionated by the mammalian mitochondria isolate kit (BioVision, #K288–50). The mitochondria fraction was lysed by RIPA buffer (Boston BioProducts, #BP-115) and the protein concentration was measured by BCA protein assay kit (Pierce, #23225). The same amount of mitochondrial lysates (250 μg of protein) from each sample were incubated with 120 μl Dynabeads^™ MyOne^™ Streptavidin T1 magnet beads (Invitrogen, #65602) one hour at room temperature (RT) followed by overnight incubation at 4°C. The beads were washed twice with RIPA lysis buffer (1 ml, 2 min at RT), once with 1 M KCl (1 ml, 2 min at RT), once with 0.1 M Na[2]CO[3] (1 ml, 10 sec), once with 2 M urea in 10 mM Tris-HCl (pH 8.0) (1 ml, 10 sec), and twice with RIPA lysis buffer (1 ml per wash, 2 min at RT). After the final wash, the beads were resuspended in 1 ml RIPA lysis buffer and transferred to new tubes, followed by elution with boiling each sample in 30 μl of 3x protein loading buffer supplemented with 2 mM biotin and 20 mM DTT at 95°C for 10 min. The same amount of total lysates before immunoprecipitation (input) and elutes were loaded for Western blot analyses. For the identification of differentially enriched proteins by proteomics, the protein complexes in the elutes were separated by SDS-PAGE and stained using the silver stain kit (Pierce #24612). Gel slices corresponding to the identified proteins were excised and processed for mass spectrometry by the Proteomics Core Facility at UTSW. Western blot analysis. Western blot was performed as described ^[296]44 using antibodies against DBT (Thermo-Fisher, #PA5–29727), DLD (Proteintech, #16431–1-AP), GLUD1 (Invitrogen, #PA5–28301a), BCAT1 (OriGene Technologies, #TA504360), BCAT2 (ABconal, #A7426), BCKDHA (Cell Signaling Technology, #90198), P-BCKDHA (Cell Signaling Technology, #40368), BCKDHB (Santa Cruz Biotechnology, #sc-374630), MT-CO2 (Proteintech, #55070–1-AP), P70-S6K (Cell Signaling Technology, #2708S), PPM1K (Abcam, #ab135286), LAT1 (Santa Cruz Biotechnology, #sc-374232), Puromycin (EMD Millipore, #MABE343), β-actin (Cell Signaling Technology, #5125), Vinculin (Cell Signaling Technology, #4650), and GAPDH (Cell Signaling Technology, #3683) with 1:1000 dilutions. For analysis of mTOR activation, tissue lysates were prepared in Cell Lysis Buffer (Cell Signaling Technologies, 9083) containing protease inhibitor tablets (Roche) and phosphatase inhibitor cocktails 2 and 3 (Sigma) and 10 mM PMSF with antibodies against p70 S6 Kinase (Cell Signaling Technology, #9202), phospho-p70 S6 Kinase (Cell Signaling Technology, #9204), 4EBP1 (Cell Signaling Technology #9452), and phospho-4EBP1 (Cell Signaling Technology, #9459). Briefly, whole cell lysates were prepared, separated on a SDS-PAGE gel, and transferred to Amersham^™ Hybond^™ P 0.45 PVDF blots (GE Healthcare, #10600023). The blots were incubated with primary antibodies with 5% non-fat milk in TBS/T (20mM Tris-HCl, pH7.5, 150mM NaCl, 0.1% Tween-20) at 4°C overnight with shaking. After washing 3 times with TBS/T, the blots were incubated with secondary antibodies with 5% non-fat milk in TBS/T for one hour at room temperature. The blots were then washed 3 times with TBS/T and developed using Plus-ECL (PerkinElmer, #NEL104001EA). Co-Immunoprecipitation. Mouse kidney mitochondria were isolated using the mammalian mitochondria isolation kit (BioVision, #K288–50). Mitochondrial proteins were extracted as previously described through gentle homogenization of the mitochondrial pellets in solubilizing buffer (SB) containing 50 mM potassium phosphate, 0.2 mM EDTA, 0.75% CHAPS, 25 mM KCl, 1 mM DTT, pH 7.4, and protease inhibitor mixture set III (Calbiochem) at a mitochondrial protein concentration pf 25 mg protein/ml ^[297]13. The solubilized mitochondria were centrifuged at 15,000xg for 30 min. Then the supernatant was removed and the protein concentration was measured by BCA protein assay kit (Pierce, #23225). Next, 50 μl magnetic Dynabeads Protein A (Invitrogen, #10001) were cross-linked to 5 μg BCKDHA antibody (Cell Signaling Technology, #90198) using bis(sulfosuccinimidyl)suberate (BS^3) (Thermo Scientific, #21580) according to the manufacturer’s cross-linking immunoprecipitation protocol. Antibody conjugated beads were incubated with the same amount of mitochondrial protein in SB with or without the addition of Valine (5 mM), α-KG (20 mM), thiamine diphosphate (1 mM), and Coenzyme A (0.4 mM) ^[298]13 with rotation for 10 min at room temperature. The beads were washed 3 times with SB with or without Valine, α-KG, thiamine diphosphate and Coenzyme A. After the final wash, the beads were resuspended in SB and transferred to a clean tube followed by elution with elution buffer (50 mM glycine, pH 2.8), NuPAGE LDS Sample Buffer (Invitrogen, #NP0007) and NuPAGE Sample Reducing Agent (Invitrogen, #NP0004) and heated for 10 min at 70°C. Samples were separated from the magnetic beads and loaded for Western blot analysis. Surface sensing of translation (SUnSET) assay. SUnSET assays were carried out as previously described ^[299]46,[300]62. Briefly, ad libitum fed mice were administered a puromycin dose of 40 nmol/g of body weight intraperitoneally. After 30 min, mice were sacrificed and tissue samples were flash frozen and processed for Western blot analysis as described previously. Valine protein incorporation. Extraction and hydrolysis of protein from mouse tissues was performed as previously described ^[301]63. Briefly, approximately 50 mg mouse tissues from KIV_M+5 isotope tracing experiments were homogenized in lysis buffer (25 mM HEPES pH 7.2, 10 mM EDTA, 2% SDS, and cOmplete^™ protease inhibitor cocktail (Roche)). Lysates were then sonicated and centrifuged at 16,000xg for 5 min. Then 400 μl methanol, 200 μl chloroform, and 300 μl water were added sequentially to 100 μl supernatant and vortexed between each addition. The mixture was centrifuged at 16,000xg for 2 min and the upper layer was discarded. Then 300 μl of methanol was added to the mixture, vortexed and centrifuged at 16,000xg for 2 min. The supernatant was discarded and the pellet was air-dried. Then 200 μl 6 M HCl was added to the pellets. Samples were heated at 110°C overnight and air-dried in a SpeedVac, followed by derivatization as described above. NAD^+/NADH measurement. NAD^+/NADH measurement was performed using the NAD^+/NADH-Glo assay kit (Promega, #G9071) with modifications as previously described ^[302]64. Briefly, approximately 10 mg of tissues were homogenized in 100 μl cold lysis buffer (1% Dodecyltrimethylammonium bromide (DTAB) in 0.2N NaOH diluted 1:1 with PBS) and placed on ice. To isolate NADH, 20 μl of the sample was incubated at 75°C for 30 min, followed by equilibration to room temperature (10 min) and the addition of 20 μl 0.2 M Tris in 0.2 N HCl to quench the reaction. To isolate NAD^+, 20 μl lysis buffer and 20 μl 0.4 N HCl was added to 20 μl sample and incubated at 60°C for 15 min. After 10 min equilibration to room temperature, the reaction was quenched with 20 μl 0.5 M Tris, pH 10.7. The NAD^+/NADH-Glo assay kit was used to quantitate NAD^+ and NADH levels according to the manufacturer’s protocol. Histology. Tissues were collected and fixed in formalin, dehydrated and embedded in paraffin. Sectioned slides were rehydrated and stained with hematoxylin and eosin as described previously ^[303]44. Quantification and statistical analysis. Statistical details including N, mean, and statistical significance values are indicated in the text, figure legends, or methods. Error bars in the experiments represent SEM or SD from either independent experiments or independent samples. All statistical analyses were performed using GraphPad Prism software unless specified otherwise, and the detailed information about statistical methods is specified in figure legends or methods. The numbers of independent experiments or biological replicate samples and P values (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, n.s. not significant) are provided in individual figures. P < 0.05 was considered statistically significant. Panels in [304]Fig. 1e, [305]4b–[306]d, [307]Extended Data Fig. 2b–[308]e, [309]2i–[310]k, [311]5c,[312]d, [313]6a,[314]c–[315]e, [316]7a,[317]b, and [318]8a,[319]f show a representative image of three independent experiments or biological replicate samples with similar results. Extended Data Extended Data Figure 1. Distinct Expression Patterns of BCAT Isozymes in Human and Mouse Tissues. Extended Data Figure 1. [320]Open in a new tab (a) Expression of BCAT1 mRNA in seven human organs across developmental time. (b) Expression of BCAT2 mRNA in seven human organs across developmental time. (c) Expression correlation of genes encoding enzymes involved in the BCAA metabolic pathway in human tissues across developmental times. The color and size of each circle indicate the Pearson’s correlation coefficient scores for positive (red) and negative (blue) correlations. P values by a two-sided t-distribution with n-2 degrees of freedom. (d) mRNA and protein expression for BCAT1 and BCAT2 across human tissues. Heatmap shows the normalized expression values for mRNA (pink) and protein (brown) in the GTEx database. (e) Expression of Bcat2, Bckdha, Bckdhb, Bckdk and Ppm1k mRNA in fractionated liver cell types including hepatocytes (HC), hepatic stellate cells (HSC), Kupffer cells (KC), and liver sinusoidal endothelial cells (LSEC). mRNA expression values by RNA-seq (FKPM) are shown. (f) Expression of BCAT2, BCKDHA, BCKDHB, BCKDK and PPM1K protein in fractionated liver cell types. Protein expression values by quantitative proteomics (fraction of total) are shown. Extended Data Figure 2. Phenotypic Analysis of BCAT1 and BCAT2 Knockout Mice. Extended Data Figure 2. [321]Open in a new tab (a) Schematic of the genotyping PCR for Bcat1^−/− and Bcat2^−/− constitutive KO mice. The locations and sizes of the genotyping primers and PCR products are shown. (b) Representative genotyping PCR and Sanger sequencing results for Bcat1^−/− and Bcat2^−/− mice. (c) Validation of Bcat1^−/− KO by Western blot analysis of mouse brain and heart. HEK293T cells with BCAT1 and BCAT2 double knockout (DKO) and DKO cells with BCAT1 or BCAT2 overexpression (BCAT1^OE and BCAT2^OE) were analyzed as controls. (d) Representative images for male and female mice of the indicated genotypes. (e) Representative H&E staining of epididymal adipose tissue in WT and Bcat2^−/− mice. Scale bar, 50 μm. (f) The organ weight relative to body weight is shown for heart, kidney and spleen of WT, Bcat1^−/− and Bcat2^−/− mice with the number (N) of samples shown. Results are mean ± SD and analyzed by one-way ANOVA. *P < 0.05, n.s. not significant. (g) Hindlimb skeletal muscle and epididymal adipose weight of the indicated genotypes. Results are mean ± SD (N = 4 WT and 3 Bcat2^−/− mice for muscle, and 4 WT and 4 Bcat2^−/− mice for fat) and analyzed by two-sided unpaired t-test. (h) Plasma AST and ALT levels in mice of the indicated genotypes. Results are mean ± SD (N = 5 WT and 4 Bcat2^−/− mice) and analyzed by two-sided unpaired t-test. (i-k) Representative H&E staining of heart (i), kidney (j), and liver (k) from WT or Bcat2^−/− mice. Scale bar, 50 μm. Extended Data Figure 3. Metabolomic Alterations Caused by BCAT1 or BCAT2 Deficiency. Extended Data Figure 3. [322]Open in a new tab (a) Principle component analysis of metabolic profiles of liver, pancreas, plasma and muscle in WT, Bcat1^−/− and Bcat2^−/− mice. (b) Enriched pathways associated with increased or decreased metabolites across major organs in Bcat1^−/− mice. Color scale indicates negative log10 transformed P values of the pathways associated with increased (red) or decreased (blue) metabolites determined by MetaboAnalyst 5.0 pathway analysis using hypergeometric test. (c) Enriched pathways associated with increased or decreased metabolites across major organs in Bcat2^−/− mice. Color scale indicates negative log10 transformed P values of the pathways associated with increased (red) or decreased (blue) metabolites determined by MetaboAnalyst 5.0 pathway analysis using hypergeometric test. Extended Data Figure 4. Differentially Enriched Metabolites Caused by BCAT1 or BCAT2 Deficiency. Extended Data Figure 4. [323]Open in a new tab Hierarchical clustering heatmaps are shown for the top 50 differentially enriched metabolites in each indicated tissue between WT, Bcat1^−/− and Bcat2^−/− mice from one-way ANOVA analysis. The red and blue color indicate higher and lower metabolite abundance by MetaboAnalyst 5.0, respectively. Extended Data Figure 5. Neuropathology of Bcat2^−/− mice. Extended Data Figure 5. [324]Open in a new tab (a) Transverse relaxation time (T[2]) maps of WT (left) and Bcat2^−/− (right) mouse brains. Color bar indicates T[2] values. (b) Relative abundance of L-DOPA, threonine, tryptophan, and tyrosine in brain tissue of Bcat2^−/− mice. Results are mean ± SD (N = 4 WT and 4 Bcat2^−/−) and analyzed by two-sided unpaired t-test. (c,d) Representative H&E staining of the striatum (c) and brainstem (d) in WT and Bcat2^−/− mice. Scale bar, 50 μm. Extended Data Figure 6. Generation of BCAT1 or BCAT2 Reconstituted Cell Models for Proximity Labeling. Extended Data Figure 6. [325]Open in a new tab (a) Western blot analysis of enzymes in the BCAA metabolic pathway in major tissues of WT and Bcat2^−/− mice. (b) Western blot analysis and quantification of BCKDHA and P-BCKDHA in WT and Bcat2^−/− liver samples. Results shown as mean ± SD (N = 12 WT and 11 Bcat2^−/− mice) and analyzed by two-sided unpaired t-test. (c) Generation of BCAT1 or BCAT2 reconstituted HEK293T cells. Western blot analysis is shown for the mitochondrial location of mTD, BCAT1-mTD and BCAT2-mTD, and efficient biotinylation of mitochondrial proteins in the presence of doxycycline (Dox) and exogenous biotin. MT-CO2 and p70 S6 kinase (P70-S6K) were analyzed as controls for the fractionated mitochondria and cytoplasm, respectively. (d) Validation of efficient biotinylation of mitochondrial proteins by proximity labeling using mTD, BCAT1-mTD or BCAT2-mTD in the presence of Dox and exogenous biotin. (e) Co-immunoprecipitation of BCAT1 and BCKDH in mouse kidney mitochondrial lysates using α-BCKDHA antibody cross-linked Dynabeads Protein A. Input or immunoprecipitated samples were blotted with α-BCKDHA and α-BCAT2 antibodies, respectively. Co-immunoprecipitation was performed in the presence or absence of additives (Val, PLP, α-KG, thiamine diphosphate and Coenzyme A), whereas co-immunoprecipitation using Dynabeads Protein A without cross-linking with α-BCKDHA antibody was performed as a negative control. Extended Data Figure 7. Ectopic BCAT1 Expression Ameliorates BCAT2 Deficiency-Induces Metabolic Defects. Extended Data Figure 7. [326]Open in a new tab (a) Western blot analysis to validate the cytosolic expression of BCAT1 in WT and Bcat1^OE kidney lysates. (b) Western blot results are shown for the validation of Bcat1^OE and Bcat1^OE;Bcat2^−/− mice and the expression of enzymes in the BCAA metabolic pathway in major tissues. (c) Fractional enrichment of KIV_M+5 in the plasma of the indicated genotypes during the 2-hour stable isotope infusion experiments. Results are mean ± SD (N = 4 WT, 4 Bcat2^−/−, 3 Bcat1^OE and 3 Bcat1^OE;Bcat2^−/− mice). (d) Relative abundance of Val_M+0,5 and KIV_M+0,5 in plasma samples before (0 hour) and after (2 hour) KIV_M+5 stable isotope infusion with the number (N) of independent samples shown. Results are mean ± SD and analyzed by two-way ANOVA. (e) Glutamate/α-KG ratios in major tissues of the indicated genotypes with the number (N) of independent samples shown. Results are mean ± SD and analyzed by one-way ANOVA. (f) NAD^+/NADH ratio in major organs of the indicated genotypes with the number (N) of samples shown. Results are mean ± SD and analyzed by two-way ANOVA. (g) Insulin levels in the serum of the indicated genotypes. Results are mean ± SD (N = 5 WT, 4 Bcat2^−/−, 4 Bcat1^OE and 4 Bcat1^OE;Bcat2^−/− mice) and analyzed by one-way ANOVA. Extended Data Figure 8. BCAT1-Catalyzed BCKA Reamination Increases BCAA Transport and Protein Synthesis. Extended Data Figure 8. [327]Open in a new tab (a) Western blot analysis of LAT1 (encoded by Slc7a5) in major tissues of WT, Bcat1^OE and Bcat1^OE;Bcat2^−/− mice. (b) Quantification of LAT1 expression in mice of the indicated genotypes relative to WT samples with the number (N) of samples shown. Results are mean ± SD and analyzed by two-sided unpaired t-test. (c) Representative Western blot analysis of 4EBP1, P-4EBP1, p70S6K, and P-p70S6K in WT or Bcat2^−/− liver lysates. Quantification is shown on the bottom. Results are mean ± SD (N = 3 WT and 3 Bcat2^−/− mice on BCAA-normal diet, and 3 Bcat2^−/− mice on BCAA-choice diet) and analyzed by one-way ANOVA. (d) Representative Western blot analysis of 4EBP1, P-4EBP1, p70S6K, and P-p70S6K in liver lysates from mice of the indicated genotypes. Quantification is shown on the bottom. Results are mean ± SD (N = 3 WT, 3 Bcat2^−/−, 3 Bcat1^OE and 3 Bcat1^OE;Bcat2^−/− mice) and analyzed by one-way ANOVA. (e) Amino acid levels in Bcat2^−/−, Bcat1^OE, and Bcat1^OE;Bcat2^−/− mouse tissues. Heatmap shows the log2 fold changes of the indicated metabolites in each mutant sample relative to WT controls (N = 3 Bcat2^−/−, 3 Bcat1^OE, and 4 Bcat1^OE;Bcat2^−/− mice). (f) Representative Western blot analysis of puromycin incorporation in major tissues of WT, Bcat1^OE and Bcat1^OE;Bcat2^−/− mice. Supplementary Material Supplementary Information [328]NIHMS2063434-supplement-Supplementary_Information.docx^ (32.8KB, docx) Supplementary Table 1 Supplementary Table 1. Metabolomic Profiling of BCAT1 and BCAT2 KO Mice The name, mass, HMDB, associated KEGG pathway, fold change relative to WT, raw P value and FDR-corrected P values are shown for each metabolite in the indicated mouse genotype and tissue type. [329]NIHMS2063434-supplement-Supplementary_Table_1.xlsx^ (602.6KB, xlsx) Supplementary Table 2 Supplementary Table 2. Metabolomic Profiling of BCAT2 KO with Normal or BCAA-Choice Diet The name, mass, HMDB, associated KEGG pathway, fold change relative to WT, raw P value and FDR-corrected P values are shown for each metabolite in the indicated mouse with BCAA normal or choice diet. [330]NIHMS2063434-supplement-Supplementary_Table_2.xlsx^ (404.8KB, xlsx) Supplementary Table 3 Supplementary Table 3. Proteins and Peptides Identified by Proximity Labeling Proteomics The peptide sequence, modifications, number of PSMs are shown for each identified protein by BCAT2-mTD-mediated proximity labeling. [331]NIHMS2063434-supplement-Supplementary_Table_3.xlsx^ (13.7KB, xlsx) Supplementary Table 4 Supplementary Table 4. List of Primers and sgRNAs [332]NIHMS2063434-supplement-Supplementary_Table_4.xlsx^ (12.4KB, xlsx) Supplementary Table 5 Supplementary Table 5. Statistical Analysis of [333]Figure 6e [334]NIHMS2063434-supplement-Supplementary_Table_5.xlsx^ (21.8KB, xlsx) Acknowledgments