If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Cell Biology Program, Sloan Kettering Institute, Memorial Sloan Kettering Cancer Center, New York, New York, USALouis V. Gerstner Jr. Graduate School of Biomedical Sciences, Memorial Sloan Kettering Cancer Center, New York, New York, USA
The tricarboxylic acid (TCA) cycle, otherwise known as the Krebs cycle, is a central metabolic pathway that performs the essential function of oxidizing nutrients to support cellular bioenergetics. More recently, it has become evident that TCA cycle behavior is dynamic, and products of the TCA cycle can be co-opted in cancer and other pathologic states. In this review, we revisit the TCA cycle, including its potential origins and the history of its discovery. We provide a detailed accounting of the requirements for sustained TCA cycle function and the critical regulatory nodes that can stimulate or constrain TCA cycle activity. We also discuss recent advances in our understanding of the flexibility of TCA cycle wiring and the increasingly appreciated heterogeneity in TCA cycle activity exhibited by mammalian cells. Deeper insight into how the TCA cycle can be differentially regulated and, consequently, configured in different contexts will shed light on how this pathway is primed to meet the requirements of distinct mammalian cell states.
Cellular metabolism comprises a complex network of biochemical reactions that convert nutrients into metabolic building blocks that fuel the growth and survival of living organisms. Metabolic processes require energy exchange, which is achieved through oxidation–reduction reactions that transfer high-energy electrons from one molecule onto another. Outputs of metabolic pathways can broadly be separated into three categories: energy, reducing equivalents, and macromolecular precursors. These outputs are required in all living cells, but the relative degree to which cells rely on each output is dependent upon the specific requirements of distinct cell types and cell states. Accordingly, myriad signaling pathways and regulatory networks control the balance between catabolic pathways, which break down molecules to harness chemical energy, and anabolic pathways, that orchestrate macromolecular synthesis.
At the nexus of both catabolic and anabolic metabolism lies the tricarboxylic acid (TCA) cycle, a broadly conserved metabolic pathway consisting of a cyclic series of chemical reactions that harness high-energy electrons from fuel sources (
). The chemical reaction that initiates each “turn” of the TCA cycle is the condensation of the four-carbon metabolite oxaloacetate (OAA) with the two-carbon molecule acetyl-CoA to generate citrate. Subsequent reactions oxidize citrate to produce two molecules of CO2 and one GTP or ATP molecule (Fig. 1) (
). During each turn of the cycle, three hydride ions (six electrons) are transferred to three NAD+ molecules and one pair of hydrogen atoms (two electrons) are transferred to one FAD molecule, producing four reducing equivalents total (three NADH and one FADH2) (Fig. 1). Importantly, each turn of the cycle ends with the regeneration of the starting molecule, OAA. Thus, OAA acts as a catalyst in the TCA cycle: only a small amount is required for the oxidation of large amounts of acetyl-CoA (
The TCA cycle itself neither consumes molecular oxygen nor produces meaningful amounts of ATP; rather, the TCA cycle removes electrons (reducing equivalents) from inputs (e.g., acetyl-CoA) and transfers them to electron carriers that deposit their electrons onto the electron transport chain (ETC). Electron funneling through the complexes of the ETC is coupled to generation of the mitochondrial membrane potential that is ultimately used to power the production of the cellular currency (ATP) in a process known as oxidative phosphorylation (OXPHOS), as oxygen serves as the terminal acceptor for electrons that transit through the ETC (Fig. 1) (
). The TCA cycle is thus linked to oxygen consumption, as the oxidized electron carriers required to continue to turn the cycle are regenerated by the ETC, and the concerted activity of these two pathways allows for the generation of significant amounts of ATP. In this configuration, the central function of the TCA cycle is to convert fuel sources into energy, and therefore, the TCA cycle is generally considered a catabolic process. However, several chemical intermediates of the TCA cycle also serve as critical precursors for biosynthetic reactions, which we will discuss in detail later (
Cells must carefully calibrate anabolic and catabolic pathways to ensure balance between nutrient supply and demand. Given the central role of the TCA cycle in both provisioning key anabolic substrates and in maintaining energy production, it is not surprising that TCA cycle activity is under tight physiologic regulation. Multiple metabolic signals control both TCA cycle inputs and directionality, and increasing work demonstrates the importance of such TCA cycle flexibility for optimizing cellular fitness both under physiological conditions and in the context of disease. In mammals, the consequences of disrupting TCA cycle function can be found in a group of disorders known as inborn errors of metabolism, which are caused by inherited mutations in genes encoding metabolic enzymes. Human patients who present with mutations in TCA cycle–associated genes display neonatal symptoms, developmental defects, and failure to thrive (
). These observations support the notion that the TCA cycle plays a significant role in sustaining mammalian tissue function. The goal of this review is to revisit the wiring of the TCA cycle in detail, leveraging both historical work and more recent studies to provide an updated view of the mammalian TCA cycle as a dynamic metabolic network at the heart of cell biology.
Evolutionary perspectives on the TCA cycle
Cellular life emerged approximately four billion years ago, at a time when earth’s atmosphere is predicted to have been largely devoid of oxygen (
). The existence of life in any capacity on earth requires the generation of basic cellular constituents that enable organization necessary to combat entropy. Ancient metabolic pathways thus must have acted to assemble organic molecules from inorganic precursors for the synthesis of biomolecules, like proteins, lipids, and nucleic acids. A major theory in the study of primordial metabolism is that these ancient anabolic pathways originated spontaneously based on geochemical conditions before the emergence of enzymes, genetic material, or cells (
). The TCA cycle plays a central role in theories for the chemical origins of life because it supplies acetyl-CoA, pyruvate, OAA, succinate, and alpha-ketoglutarate (αKG), the five universal metabolic precursors for biosynthetic molecules (e.g., lipids, glucose, nucleic acids, amino acids, and cofactors). Suggestively, many of the reactions of the oxidative TCA cycle can occur nonenzymatically in the presence of iron as a catalyst and oxidizing agents other than oxygen (
). However, given that prebiotic chemistry would necessarily have initially required the production, not the breakdown, of organic molecules, primordial metabolism likely centered on reductive CO2-fixing pathways.
Clues about primordial metabolism have been sought by studying the metabolism of CO2-fixing chemoautotrophs, which lie at the root of the tree of life, and thus may offer insights into the earliest biosynthetic pathways (
). Accordingly, while the oxidative TCA cycle oxidizes acetyl-CoA to CO2 to generate reducing equivalents and ATP, the rTCA cycle captures CO2 to generate acetyl-CoA in a process requiring electron donors and ATP. The rTCA cycle is an appealing candidate for early prebiotic metabolism because, similar to the oxidative TCA cycle, it produces the five universal precursors to biological metabolism (
). Intriguingly, certain reaction sequences of the cycle have been shown to occur nonenzymatically in the presence of minerals under UV radiation or using metal ions under acidic and aqueous conditions, not unlike those proposed on early Earth (
). However, demonstration of the C-C bond forming reactions of the rTCA cycle has yet to be achieved efficiently and nonenzymatically under experimental conditions, raising the possibility that this portion of the pathway required organic catalysts and, thus, emerged later (
Ultimately, the ability of the rTCA cycle and/or other anabolic pathways to produce organic molecules allowed for the emergence of enzymes that facilitated these biochemical reactions. Many of the enzymes of the oxidative and forward TCA cycle are reversible and, thus, function similarly to the enzymes that comprise the rTCA cycle. One of the biggest differences between the two cycles is that the rTCA cycle requires the conversion of succinyl-CoA to αKG. This reaction, mediated by an αKG synthase, is highly unfavorable and requires a strong reducing agent in the form of reduced ferredoxin (
). In the forward TCA cycle, citrate synthase (CS) is the key enzyme that condenses acetyl-CoA with OAA to form citrate. The rTCA cycle, on the other hand, requires that citrate be cleaved to liberate acetyl-CoA and regenerate OAA, a reaction that is catalyzed by the ATP-citrate lyase (ACL) enzyme. Human ACL contains a citryl-CoA synthetase module and a citryl-CoA lyase (CCL) domain, which coordinate to catalyze the multistep cleavage of citrate (
). However, in the rTCA cycle active in the deep branching bacterial phylum Aquificae, CCL and citryl-CoA synthetase function as two distinct enzymes that mediate citrate cleavage in a stepwise manner. Recently, cross-kingdom structural analysis revealed conserved structural features between CCL in Aquificae and CS in a species of Archaea, suggesting that the process of citrate condensation originated from citrate cleavage during evolution (
The oxygenation of Earth’s atmosphere over time increased the availability of molecular oxygen, whose function as an electron acceptor facilitates oxidation of organic molecules derived from carbon assimilation (
). This development likely enabled the emergence of the oxidative TCA cycle and its coupling with OXPHOS, allowing the transfer of electrons derived from the breakdown of carbon substrates onto molecular oxygen. The evolution of increasingly complex TCA cycle enzymes was a critical development allowing for both improved efficiency of this metabolic pathway and enhanced capacity for regulation of its activity (
). Being able to fine-tune the function of metabolic pathways like the TCA cycle—in particular, to toggle between catabolic and anabolic activity—set the stage for early unicellular life forms to evolve into multicellular organisms, which require coordination of metabolic activity beyond that dictated by local nutrient availability (
At the beginning of the 20th century, cellular respiration remained one of the major unsolved problems in biology. Scientists knew that cellular respiration involved oxygen consumption and also suspected that the process was somehow linked to the catabolism of nutrients like carbohydrates. It was also known that dicarboxylic acids played some sort of role in respiration, but it was assumed that these compounds acted as substrates and thus were consumed in the process. Moreover, the intermediate reaction sequences and fundamental biochemistry of respiration had yet to be worked out. Seeking to understand the oxidative breakdown of carbohydrates, the Hungarian scientist Albert Szent-Györgyi turned to pigeon breast muscle as an experimental system because of its high rates of respiration and ease of accessibility. In 1935, Szent-Györgyi worked out the sequence of reactions from succinate to fumarate to malate to OAA and demonstrated the rapid oxidation of these substances by pigeon muscle suspensions (
) and Baumann found that very small quantities of dicarboxylic acids were sufficient to cause an increase in oxygen consumption and that this increase was beyond what was necessary for the oxidation of the quantities of substrate added. Significantly, they also found that added dicarboxylic acids were not used up and could still be detected in the medium, indicating that these compounds were functioning in a catalytic capacity (
) showed that succinate could be synthesized by animal tissues in the presence of pyruvate, leading him to speculate that succinate may arise from citrate oxidation downstream of pyruvate. A key finding in that same year came from Martius and Knoop (
) in Enzymologia in 1937, in which they discovered that citrate was readily oxidized by minced pigeon breast muscle and that citrate addition increased oxygen uptake beyond the amount necessary for the complete oxidation of the added citrate, indicating that citrate itself was stimulating consumption of molecular oxygen. Moreover, they found that citrate did not disappear during this process, indicating that it was being continuously reformed (
). Their work also showed that large quantities of citrate could be synthesized anaerobically by muscle in the presence of OAA and that OAA, when added to muscle, condensed with two carbon atoms from an unknown precursor to form citrate (
) determined that the two carbon atoms that condensed with OAA to form citrate were derived from pyruvate and that oxygen consumption in muscle increases when pyruvate is added. Later work by Lipmann et al. (
) elucidated that the two-carbon molecule required for citrate synthesis was, in fact, acetyl coenzyme A or acetyl-CoA.
Wiring of the TCA cycle
In theory, OAA regeneration allows for infinite turning of the TCA cycle, assuming a continuous supply of acetyl-CoA. Following multiple turns of the cycle, the metabolic intermediates of the TCA cycle can be entirely derived from and replaced by acetyl-CoA carbons (
). The metabolite αKG can be converted to glutamate, which in turn is diverted from the cycle and used in the synthesis of amino acids and nucleotides. Succinyl-CoA may be siphoned from the cycle to serve as a precursor of porphyrins like heme (
). OAA itself provides the carbon backbone for the amino acid aspartate, a critical input into the urea cycle and protein and nucleotide biosynthesis, and may be converted to phosphoenolpyruvate, a substrate for gluconeogenesis (
). Thus, in living cells, robust and sustained TCA cycle function requires both a continuous source of acetyl-CoA and replenishment of TCA cycle intermediates (particularly OAA) via anaplerotic reactions (Fig. 3). Here, we outline the various sources of acetyl-CoA production and OAA regeneration.
Sources of acetyl-CoA
The TCA cycle is the final step for oxidation of all cellular nutrients, including glucose, fatty acids, and amino acids, which primarily enter the cycle at the level of acetyl-CoA. While the mechanisms dictating a cell’s preferred source of acetyl-CoA are not well understood, recent work has found that different tissues in the body have distinct fuel preferences (
). Once in the mitochondrial matrix, pyruvate can undergo oxidative decarboxylation to form acetyl-CoA. This irreversible reaction is catalyzed by the pyruvate dehydrogenase complex (PDHC), a supramolecular assembly of multiple catalytic subunits (
). By converting the glycolytic product pyruvate into a substrate for TCA cycle oxidation, the PDHC functions as a gatekeeping enzyme that links glycolysis to the TCA cycle and mitochondrial respiration.
Glycolysis-derived pyruvate has alternative fates outside the mitochondrial oxidation. Pyruvate can be used as a gluconeogenic substrate, but acetyl-CoA cannot; consequently, the PDHC reaction is a key step that is tightly regulated to control whether pyruvate is used in the TCA cycle or for gluconeogenesis (
). Beyond its critical role in gluconeogenesis, pyruvate utilization also has implications for redox homeostasis. Lactate dehydrogenase (LDH) can reduce pyruvate to lactate in the cytosol, and this reaction is coupled with oxidation of NADH to NAD+. In the absence of oxygen, this process of fermenting pyruvate to form lactate, also known as anaerobic glycolysis, occurs in mammals as a means of sustaining biosynthetic processes under hypoxia (
) found that cancer cells produce significant amounts of lactate even in the presence of oxygen. Subsequent investigation has revealed that proliferating cells significantly engage in this process, termed aerobic glycolysis or the Warburg effect. While the field continues to debate the underlying benefits of engaging in aerobic glycolysis, one clear advantage to the process is that it allows cells to robustly regenerate NAD+ in the cytosol (
). Therefore, glycolysis provides several biosynthetic advantages to proliferating cells beyond producing pyruvate. Proliferating cells thus must balance use of pyruvate toward oxidation in the TCA cycle with demand for cytosolic NAD+ regeneration to support continued glycolytic flux.
While glucose-derived pyruvate is a considerable source of acetyl-CoA carbons for TCA cycle oxidation in cultured cells in vitro, recent work has revealed that glucose may not be the preferred pyruvate source of cells in vivo (
). In rapidly growing cells, the LDH-mediated conversion of pyruvate to lactate is a major source of cytosolic NAD+ that supports sustained glycolytic flux. Proliferating cells in culture typically excrete this lactate as a waste product to support redox and pH homeostasis (
). However, recent studies in vivo have demonstrated that cells import lactate from the circulation via monocarboxylate transporter 1 and subsequently oxidize lactate to form pyruvate, resulting in a large fraction of pyruvate and downstream TCA cycle intermediates being derived from circulating lactate (
). In most tissues and some tumors, the contribution of circulating lactate to TCA cycle intermediates exceeds that of glucose, supporting the idea that lactate is a fundamental TCA cycle substrate in living organisms (
), this observation raises the possibility that a basal level of TCA cycle oxidation in cells can persist independent of exogenous growth factor stimulation. Future work should be aimed at understanding the compartmentalization of lactate oxidation, as the conversion of lactate to pyruvate is coupled to reduction of NAD+ to NADH. Absent compensatory measures, this change in the cytosolic NAD+/NADH ratio will hamper metabolic pathways—like glycolysis—that depend on oxidized NAD+. Suggestively, one study found that LDH can localize to the mitochondria, raising the possibility that lactate can be directly imported into the mitochondria for further oxidation (
Acetyl-CoA is both the precursor for fatty acid synthesis and the final product of fatty acid breakdown for oxidation in the TCA cycle. To separate these functions, cells synthesize fatty acids in the cytosol via ACL-mediated citrate cleavage and, conversely, import fatty acids into the mitochondrial matrix for degradation and subsequent oxidation. The process of breaking down fatty acids to produce acetyl-CoA, known as β-oxidation, supplies the majority of acetyl-CoA for TCA cycle oxidation in certain tissues such as the heart (
). Because the mitochondrial membrane is impermeable to acyl-CoAs, fatty acids must first be transported from the cytosol into the mitochondrial matrix via the carnitine shuttle prior to undergoing β-oxidation (
). Fatty acids are first converted to fatty acyl-carnitines on the outer mitochondrial membrane by carnitine acetyltransferase I (CPT1), then transported across the membrane via the carnitine-translocase protein, and reconverted back to acyl-CoA esters inside the mitochondrial matrix by carnitine acetyltransferase II (
). The four steps of the β-oxidation process are repeated until the fatty acyl-CoA is entirely oxidized, with each iteration shortening the molecule by two carboxy-terminal carbons, which are liberated as acetyl-CoA (
). To prevent the futile cycle of simultaneous fatty acid synthesis and oxidation, the rate-limiting enzyme of fatty acid oxidation—CPT1—is regulated in both a transcriptional and an allosteric manner. Sensing of free fatty acid availability within the cell primarily occurs through peroxisome proliferator–activated receptors, which function as fatty acid–activated transcription factors that drive the expression of CPT1 (
Acetyl-CoA can also be derived from the breakdown of amino acids and ketone bodies. Removal of the amino group from ketogenic amino acids, including lysine and the branched-chain amino acids, leucine and isoleucine, results in a carbon skeleton that can be catabolized to acetyl-CoA directly or the ketone body acetoacetate (
). Amino acids can also contribute to acetyl-CoA pools via a pyruvate intermediate. For example, catabolism of glucogenic amino acids, including serine and cysteine, and the transamination of alanine by alanine aminotransferase produces pyruvate, which can be converted to acetyl-CoA by PDHC. Beyond amino acids, the ketone body β-hydroxybutyrate can also be converted to acetoacetate, providing another potential source of acetyl-CoA.
Acetate can function as a source of acetyl-CoA through the activity of acetyl-CoA synthetase (ACSS), which catalyzes the ATP-dependent ligation of acetate and CoA to produce acetyl-CoA. In cancer cells, citrate becomes labeled following supplementation of cells with isotopically labeled acetate ([13C]acetate), indicating that acetate can undergo oxidation in the TCA cycle in cultured cell lines (
). The relevance of acetate as a potential fuel source has also been observed in vivo: in glioblastoma tumors resected from patients, up to half of the intramitochondrial acetyl-CoA pool is derived from circulating acetate (
). Distinct isoforms of ACSS localize specifically to the cytosol and mitochondria, but how conversion of acetate to acetyl-CoA is coordinated between these two compartments remains poorly understood (
). While glutamine can be synthesized directly by most mammalian cells, it has been recognized for several decades that glutamine supplementation is necessary for the growth and viability of many cultured cell lines, particularly cancer cell lines (
). Glutamine taken up by cells through transporters such as alanine–serine–cysteine transporter 2 and l-type amino acid transporter 1 is converted to glutamate by glutaminase, producing ammonia, or through nitrogen-donating reactions involved in purine and pyrimidine nucleotide synthesis (
). Glutamate can then be converted to αKG, which can enter the TCA cycle for further oxidation.
Production of αKG from glutamate occurs through two mechanisms: (1) by deamination via glutamate dehydrogenase (GDH), releasing ammonia and the reducing equivalent NAD(P)H or (2) by transaminases that transfer the amino group from glutamate to a keto-acid, generating αKG and an amino acid (
). Notably, transamination is freely reversible, meaning that aspartate aminotransferases (glutamic-oxaloacetic transaminase 1; glutamic-oxaloacetic transaminase 2 [GOT2]) can alternatively directly generate glutamate and OAA using αKG and aspartate. Cells can become particularly reliant on glutamine oxidation via GDH when their ability to oxidize glucose is impaired, for example, upon MPC inhibition (
), which may be abundant in mitochondria of cultured cells. The importance of this allosteric regulation is underscored by patients harboring germline mutations in the GTP-binding region of GDH: persistently high GDH activity drives hyperinsulinism–hyperammonemia syndrome, marked by aberrant glutamate catabolism and mitochondrial ATP production leading to excessive insulin release from pancreatic beta cells (
). GOT2-driven glutamate metabolism thus offers several benefits to proliferating cells: GOT2 circumvents physiological limitations on glutamate catabolism, preserves the amine nitrogen of glutamate for aspartate synthesis, and reduces production of toxic ammonia.
Across cultured cell lines, the majority of the carbons in aspartate and other TCA cycle intermediates are supplied by glutamine (
). As a result, most cultured cells are exquisitely dependent upon glutamine to support anaplerosis. However, the extent to which cells depend on glutaminolysis for anaplerosis has been shown to depend on multiple factors, including glucose versus glutamine availability, cellular capacity to deal with ammonia toxicity, and overall cellular demand for glutamine, which also serves as a nitrogen source in biosynthetic pathways (
). Anaplerotic substrate preference is likely also governed by cellular redox demands given that GDH-mediated production of αKG also produces NAD(P)H, which can be used for oxidative stress management and other biological processes (
). PC-derived OAA can either contribute to the TCA cycle or serve as a substrate for phosphoenolpyruvate carboxykinase, which catalyzes the decarboxylation of OAA to phosphoenolpyruvic acid in the gluconeogenic pathway (
). Consequently, while PC is expressed in most tissues, it is particularly active in gluconeogenic tissues like the kidney and liver. Even in nongluconeogenic tissues, PC responds to physiologic nutrient shifts to maintain anaplerosis. PC activity is allosterically activated by abundant acetyl-CoA, making it a potent sensor of OAA insufficiency in the mitochondria (
). When glutamine anaplerosis is disrupted, PC becomes a critical source of OAA. PC is essential for growth of succinate dehydrogenase (SDH)–mutant cells, which exhibit a truncated TCA cycle and thus require an alternative source of OAA to support TCA cycle function and sustain aspartate production for anabolic pathways (
). Likewise, cells that are addicted to glutamine become dependent on PC for growth when glutamine metabolism is suppressed; conversely, cells with high basal PC activity are generally more resistant to interruption of glutamine metabolism (
). Thus, while glutamine anaplerosis efficiently allows cells both to capture reducing equivalents from substrate oxidation and maintain OAA production, PC provides a critical backup when glutamine is not available or when nutrient status is so high that the ATP-dependent circumventing of the oxidative TCA cycle mediated by PC may actually help cells cope with nutrient overload.
Other sources of anaplerosis
Succinyl-CoA and fumarate are additional anaplerotic entry points within the TCA cycle. Mitochondrial β-oxidation of fatty acids with an odd number of carbon atoms yields acetyl-CoA and, ultimately, the 3-carbon propionyl-CoA, which cannot be further oxidized through the β-oxidation pathway. Rather, propionyl-CoA is converted to the TCA cycle intermediate succinyl-CoA through a pathway involving propionyl-CoA carboxylase and methylmalonyl-CoA mutase (
), but the degree to which cells can activate propionyl-CoA catabolism to sustain anaplerotic flux under physiologic conditions remains largely unknown. Suggestively, some cultured cells maintain large propionyl-CoA pools fueled by isoleucine catabolism, and it will be interesting for future work to investigate the significance of this flux for TCA cycle metabolism and cell proliferation (
). Following entry into the pathway and oxidation to OAA, these anaplerotic substrates can support continuous cycling of the TCA cycle. The signals that may control these fluxes—and whether usage of these substrates modulates other aspects of TCA cycle wiring—remains to be explored.
The reducing equivalent NADH is produced in the mitochondria through the reactions of the TCA cycle but is also a byproduct of glycolysis in the cytosol. Notably, while NAD+ itself can be imported into mitochondria through SLC25A51–MCART1 (
), NADH is not efficiently transported. The impermeability of the inner mitochondrial membrane to reduced electron carriers necessitates pathways that mediate the transport of reducing equivalents from the cytoplasm into the mitochondrial matrix for oxidation in the ETC. These so-called “electron shuttles” transport electrons by linking the oxidation of cytosolic NADH with reduction of a substrate that can be imported into the mitochondria and subsequently reoxidized. Beyond supporting net NADH “import” into the mitochondria, these shuttles also facilitate the oxidation of cytosolic NADH, which is required to support continued glycolysis and the de novo synthesis of serine and lipids (
). While this shuttle does not involve TCA cycle intermediates, it does represent an electron input into the ETC. This pathway relies on two glycerol 3-phosphate dehydrogenase (GPD) enzymes, GPD1 and GPD2, located in the cytosol and mitochondrial membrane, respectively, to coordinate cytosolic NAD+ regeneration with electron donation to the ETC (Fig. 4A). The glycerol 3-phosphate shuttle is not considered the major route of NADH shuttling in most mammalian cells but has been shown to be particularly active in brown adipose tissue and may be implicated in thermoregulation (
). Moreover, activity of this pathway may become upregulated to compensate for disruption of other electron shuttles, for example, in the case of the developmental disorder caused by deficiency in the malate–aspartate shuttle enzyme malate dehydrogenase 1 (MDH1) (
The malate–aspartate shuttle, or Borst cycle, was first proposed by Piet Borst in the early 1960s during his study of Ehrlich ascites tumor cells. He determined that it was improbable that any of the known electron shuttles of that time, including the glycerol 3-phosphate shuttle, were active in these cells and proposed the reactions of the malate–aspartate shuttle as a means of driving the oxidation of cytosolic NADH (
). The malate–aspartate shuttle uses the complementary reactions of glutamic-oxaloacetic transaminase 1 and GOT2 transaminases paired with MDH1 and MDH2 to oxidize NADH in the cytosol and reduce NAD+ in the mitochondria (Fig. 4B). While theoretical at first, the reactions of the malate–aspartate shuttle provided an explanation for the earlier finding that GOT2 transaminase was highly active in the mitochondria and present in high concentrations that were proportional to those of TCA cycle enzymes (
) indicated that the NAD+/NADH ratio is significantly higher in the cytosol than the mitochondria, making it unclear how electron shuttles could sustain the continuous oxidation of cytosolic NADH. Thus, the energetics underlying the shuttle were debated until the early 1970s when it was discovered that efflux of aspartate from the mitochondria was energy dependent and coupled to the import of a proton, allowing the shuttle to be powered by the proton motive force (
). Moreover, given that aspartate efflux is dependent upon mitochondrial membrane potential, variability in TCA cycle and ETC function will both influence and be influenced by malate–aspartate shuttle activity (
). However, how TCA cycle flux and malate–aspartate shuttle activity are coordinated to support cross-compartment metabolism is poorly understood. Future work should aim to elucidate the mechanisms underlying how intermediates are partitioned between the TCA cycle and the malate–aspartate shuttle, including factors controlling the fate of OAA—whether continuing the TCA cycle or exiting as aspartate.
Prior to the discovery that the malate–aspartate shuttle was powered by the proton motive force, Piet Borst proposed the reactions of the citrate–malate shuttle as a potential alternative electron shuttle that could drive continuous NADH oxidation despite the cytosol's high NAD+/NADH ratio. While both shuttles rely on MDH1 and MDH2, the citrate–malate shuttle is coupled with an energy-expending reaction to favor cytosolic NADH oxidation (
). Here, TCA cycle–derived citrate is transported by the citrate–malate antiporter SLC25A1 to the cytosol where energy (ATP) is required for cleavage by ACL, a reaction that liberates acetyl-CoA and OAA. OAA is then reduced to malate by MDH1, and malate is reimported into the mitochondria to complete the electron shuttle (Fig. 4B). Given that the mitochondrial membrane is impermeable to acetyl-CoA, it was appreciated early on that cleavage of cytosolic citrate by ACL provides a route for delivering acetyl-CoA into the cytosol and, thus, supporting fatty acid synthesis (
). However, experimental evidence that citrate cleavage also impacts NADH shuttling to the mitochondria came with the observation that hydroxycitrate, an inhibitor of ACL, significantly increases the lactate over pyruvate ratio, a proxy for the cytosolic NADH/NAD+ ratio, in rat livers perfused with ethanol (
). More recently, it was demonstrated that the citrate–malate shuttle performs functions beyond that of an electron shuttle and, in fact, supports continuous citrate regeneration and TCA cycle intermediate homeostasis (
). In cancer cells and embryonic stem cells (ESCs), the citrate–malate shuttle comprises a major alternative to the traditional TCA cycle and the degree to which cells engage the traditional TCA cycle as opposed to the citrate–malate shuttle, which may represent a “noncanonical TCA cycle,” is cell state dependent (
). Given that electron shuttles translocate the reducing equivalent (NADH) that both drives ETC activity and modulates flux through the TCA cycle, future work will likely continue to uncover functional links between these metabolic pathways and the TCA cycle.
Flexibility of TCA cycle wiring
For many decades after its initial discovery, the TCA cycle was generally thought to operate as a fixed pathway in most cellular contexts. However, it has become increasingly appreciated that TCA cycle function is not “one size fits all,” and that components of the pathway are differentially and flexibly engaged in a context-specific fashion. For example, under conditions of hypoxia or impaired mitochondrial respiration, certain reactions of the TCA cycle reverse to support glutamine-dependent reductive carboxylation (
). As low mitochondrial NAD+/NADH inhibits production of acetyl-CoA by PDH, reductive carboxylation provides a critical alternative mechanism to generate citrate required to sustain cytosolic acetyl-CoA supply for de novo lipid synthesis even during reducing conditions. ACL cleavage of citrate derived from reductive carboxylation may also provide an alternative anaplerotic source of OAA, allowing cells to bypass a truncated or otherwise defective TCA cycle (
Beyond reductive carboxylation, other portions of the TCA cycle have also been shown to exhibit reversibility. Krebs found that GDH operates at close to equilibrium in vitro, and some evidence indicates that GDH may be reversible in some tissues in vivo (
). For example, providing isotopically labeled ammonia to perfused livers or to rat portal veins results in the appearance of heavy labeled amino acids—reactions that would all depend on GDH as the initiating step for ammonia assimilation (
). Conditions of high ammonia availability may also favor reverse GDH activity as a means of ammonia detoxification and/or biomass accumulation. Indeed, breast cancer cells can engage in reductive amination, and GDH activity supports ammonia incorporation and tumor growth in mouse models of breast cancer (
Extremely reducing conditions can ultimately drive additional steps of the TCA cycle to operate in reverse. For example, ETC inhibition, which prevents cells from using molecular oxygen as a terminal electron acceptor, can force the SDH complex to operate in reverse, reducing fumarate to produce succinate (
). By siphoning electrons onto fumarate, reverse SDH activity allows the ETC to continue to accept electrons from other oxidation reactions, such as that catalyzed by dihydroorotate dehydrogenase, which is critical for nucleotide biosynthesis (
). Even CS may reverse: CS reversibility, which occurs in prokaryotes under high CO2 conditions (see aforementioned), has been observed in mammalian cells in the context of SDH deficiency in which reductive glutamine metabolism produces citrate and, strikingly, citrate-derived OAA that can be used to sustain biosynthesis (
). Whether mitochondrial citrate cleavage by CS provides any specific advantages over cytoplasmic citrate cleavage by ACL, or whether reverse CS flux is secondary to biochemical conditions that disfavor citrate efflux, remains to be determined. More broadly, these reversible networks show that metabolic perturbations can ripple throughout cellular compartments and trigger pathway rewiring to ensure resilience in the face of stress.
Nodes of TCA cycle control
Because the TCA cycle functions as a critical biosynthetic hub and is coupled with cellular energy production via OXPHOS, cells have evolved multiple nodes of TCA cycle regulation. In this section, we will outline three different mechanisms by which cells achieve tight control over TCA cycle activity.
Allosteric regulation of TCA cycle enzyme activity
Metabolic flux through the TCA cycle is tightly coordinated by both negative and positive allosteric regulation of TCA cycle–associated enzymes. Three enzymes in particular—CS, IDH, and oxoglutarate dehydrogenase (OGDH)—catalyze rate-controlling steps in the TCA cycle and are considered regulatory enzymes within the pathway (
). High NADH thus serves as a key signal of ETC overloading to shutdown TCA cycle flux and prevent delivery of excess electrons to the ETC that can generate potentially dangerous reactive oxygen species (ROS). OXPHOS activity also signals to the TCA cycle at the level of ATP: high ATP/ADP ratios allosterically inhibit IDH, resulting in a deceleration of the cycle upon excess energy supply (
). Altogether, allosteric regulation of core TCA cycle enzymes tunes the TCA cycle according to reactant and product availability, preventing excess nutrient catabolism in the absence of demand.
While not technically part of the TCA cycle, the PDHC generates the starting molecule of the cycle, acetyl-CoA, thereby functioning as a gatekeeper of TCA cycle activity. During PDHC catalysis, a carboxyl group is removed from pyruvate and released as CO2, and the remaining two-carbon molecule is oxidized, with NAD+ accepting the electrons to form NADH (Fig. 5). The PDHC is a multienzyme complex composed of multiple copies of three catalytic subunits: pyruvate dehydrogenase (E1), dihydrolipoamide acetyltransferase (E2), and dihydrolipoamide dehydrogenase (E3). These subunits carry out the sequential conversion of pyruvate into acetyl-CoA along with five different coenzymes, including coenzyme A (CoA-SH), the universal carrier for acyl groups. The complex is activated by its substrate, pyruvate, and inhibited by its product, acetyl-CoA. Like IDH, the PDHC is also allosterically inhibited by a high ATP/ADP ratio, allowing it to respond to cellular energy status. Importantly, like the three regulatory enzymes of the TCA cycle described previously, the PDHC is negatively regulated by high NADH levels. Inhibition of the complex by NADH links PDHC flux and, thus, acetyl-CoA production to ETC function and mitochondrial redox status (Fig. 5). Recent work has shown that enhancing pyruvate oxidation via the PDHC can result in a metabolic imbalance wherein NADH production outstrips NADH oxidation via the ETC, limiting the availability of NAD+ for key biosynthetic reactions such as aspartate biosynthesis (
). These results indicate that cell proliferation can be impaired when the demand for mitochondrial NAD+ exceeds that of ATP and may explain why highly proliferative cells often engage in aerobic glycolysis over pyruvate oxidation and OXPHOS, even in the presence of sufficient oxygen (
). By converting pyruvate to lactate in lieu of oxidizing it in the TCA cycle, cells can avoid excess production of NADH relative to ATP and the concomitant effect on proliferation that this produces.
Covalent modification of TCA cycle–related enzymes
Beyond allosteric regulation, the PDHC is also regulated by covalent modification. The activity of the mammalian complex is regulated by the phosphorylation status of three serine residues on the pyruvate dehydrogenase E1 alpha subunit (
). Phosphorylation of the complex is inhibitory but reversible via removal by pyruvate dehydrogenase phosphatases (PDPs; PDP1 and PDP2). While the PDHC is activated by NAD+, ADP, and pyruvate, the PDKs are conversely activated by NADH, ATP, and high acetyl-CoA levels (
) (Fig. 5). Collectively, the PDKs and PDPs provide cells with the ability to fine-tune PDHC activity through multiple signaling inputs. For example, low oxygen levels activate the hypoxia-inducible transcription factors (HIFs), which bind to hypoxia-response elements within a number of target genes including LDH-A (LDHA) and PDK1, leading to suppression of PDHC activity with a concomitant increase in aerobic glycolysis (
). To preferentially engage aerobic glycolysis, cancer cells frequently co-opt several pathways, including the HIF/hypoxia transcriptional program, to enhance PDK1 expression and achieve PDHC repression (
). Whether core TCA cycle enzymes are also controlled by phosphorylation remains to be shown, but there are some examples of TCA cycle enzymes like SDH undergoing alternative post-translational modification (
). Future work should aim to uncover other potential TCA cycle regulatory mechanisms exerted by covalent modification.
Ion-mediated TCA cycle regulation
Calcium (Ca2+) functions as an intracellular messenger in an array of biological processes, including mitochondrial metabolism. Mitochondrial Ca2+ uptake primarily occurs via the mitochondrial calcium uniporter, a highly selective ion channel in the inner mitochondrial membrane. Mitochondrial calcium uniporter–mediated Ca2+ influx is driven by the membrane potential established by ETC activity (
). Once inside the mitochondria, Ca2+ ions directly activate IDH and OGDH enzymes. Ca2+ binding to IDH leads to a decrease of the KM for isocitrate, and this effect is enhanced in the context of low ATP/ADP ratios (
). Thus, intramitochondrial Ca2+ regulates both carbon entry into and flux through the TCA cycle. Maintenance of Ca2+ homeostasis is critical for cell viability as mitochondrial Ca2+ overload favors opening of the permeability transition pore, leading to collapse of mitochondrial membrane potential and activation of cell death (
Iron also plays a significant role in modulating the TCA cycle by regulating both the activity and expression of ACO. Both the mitochondrial and cytosolic isoenzymes of ACO—ACO2 and ACO1, respectively—contain an iron–sulfur (4Fe–4S) cluster, and catalytic activity requires substrate coordination to a specific iron atom within this cluster. This enzymatic requirement for iron makes ACO activity sensitive to iron levels within cells (
). Iron-responsive elements (IREs) are motifs found within genes related to iron metabolism. Upon iron depletion, iron regulatory proteins bind to IREs in the 5′ end of mRNA transcripts, typically resulting in reduced mRNA translation and expression (
). Other metal ions play significant roles in TCA cycle enzyme activity: IDH, OGDH, and PDP all bind Mg2+, but as free Mg2+ concentrations are considerably higher than enzyme KMs, whether fluctuations in Mg2+ are sufficient to modulate TCA cycle flux remains to be determined (
TCA cycle metabolites dictate cell fate and function
As a central metabolic hub, the TCA cycle frequently undergoes significant remodeling during both normal and pathological changes in cell fate. This section will cover some examples of how both inputs and outputs of the TCA cycle are exploited to alter cell function and, in turn, cell fate.
Metabolic control of gene expression and cell fate
A handful of TCA cycle intermediates are increasingly implicated in cell fate control. While transcription factors are the ultimate controllers of cell fate, their ability to act can be shaped by chemical modifications on DNA and histones that can modulate DNA accessibility and cofactor recruitment (
). Notably, steady-state levels of TCA cycle metabolites can affect the activity of chromatin-modifying enzymes, including αKG-dependent dioxygenases and histone acetyltransferases (Fig. 6). By co-opting the production and localization of select TCA cycle metabolites, cells can modulate the activity of chromatin-modifying enzymes and exert control over gene expression and cell fate (
Succinate, fumarate, and αKG are all implicated in regulation of αKG-dependent dioxygenases, a family of enzymes that includes Jumonji C-domain–containing histone demethylases and ten–eleven translocation (TET) DNA methylcytosine dioxygenases, which remove repressive histone marks and 5-methylcytosine, respectively (Fig. 6). These enzymes all require αKG, ferrous iron, and oxygen as cosubstrates and are inhibited by high levels of succinate and fumarate (
), thereby integrating multiple metabolic pathways in the control of gene expression programs and, ultimately, cell fate decisions.
Modulation of αKG production
Alterations in intracellular αKG levels have been shown to modulate cell fate in cancer and stem cells in a variety of model systems. The role of αKG in cell fate is most clearly demonstrated by studies that manipulate nutrient availability to modulate intracellular αKG pools. Depriving APC-mutant intestinal organoids of glutamine depletes intracellular αKG levels and promotes stem-like features and adenocarcinoma formation in vivo (
). Endogenous metabolic networks may determine αKG availability: the branched-chain amino acid transaminase 1 (BCAT1) enzyme, which transaminates αKG to initiate catabolism of valine, leucine, and isoleucine, constrains αKG pools in acute myelogenous leukemia. Accordingly, high BCAT1 activity leads to a hypermethylated chromatin landscape that blunts normal myeloid differentiation; reciprocally, blocking BCAT1 triggers αKG accumulation and induces myeloid differentiation (
Intracellular αKG pools may also be controlled by oncogenic mutations. In cell lines derived from mouse models of pancreatic ductal adenocarcinoma driven by mutant Kras and reversible silencing of the tumor suppressor p53, restoring p53 function rewires TCA cycle metabolism and drives accumulation of αKG relative to succinate (
). An increased αKG/succinate ratio in this system, whether driven by p53 reactivation or suppression of OGDH, enhances activity of αKG-dependent dioxygenases like the TETs and drives reacquisition of premalignant gene expression patterns associated with tumor differentiation (
). Together, these studies provide examples of how αKG accumulation—whether enabled by changes in TCA cycle wiring or upstream pathways—can function in a tumor suppressive capacity by stimulating chromatin remodeling that drives a more differentiated and less aggressive cancer phenotype. However, the effect of αKG on cell fate is likely context specific as interventions that increase αKG levels in mouse ESCs favor self-renewal over differentiation (
). How αKG levels are set in cells—in particular, whether changes in mitochondrial TCA cycle wiring can result in alterations in nucleocytosolic αKG pools—and what determines cellular response to changes in αKG, remains to be determined.
Germline and somatic mutations in genes encoding TCA cycle enzymes are directly implicated in αKG-dependent dioxygenase activity. Mutations that impair fumarate hydratase (FH) and SDH activity are associated with pathologic accumulation of substrates fumarate and succinate, respectively. Oncogenic mutations in genes encoding IDH1/2 result in neomorphic enzyme activity that favors reduction of αKG to d-2-hydroxyglutarate (D-2HG) (
). Succinate, fumarate, and 2HG are all considered “oncometabolites” because they are linked to development of certain human cancers and because they affect cancer-relevant processes by virtue of their ability to act as competitive inhibitors of αKG-dependent dioxygenases (
). By blocking the enzymatic activity of histone and DNA demethylases, d-2HG can function to lock cells in a hypermethylated state that blocks differentiation and reinforces a malignant and stem cell–like phenotype (
), raising the possibility that 2HG accumulation may play a regulatory role under normal physiological conditions.
Like 2HG accumulation, increased levels of succinate and fumarate can inhibit Jumonji C-domain–containing histone demethylases and TET enzymes, driving a widespread hypermethylation phenotype that disrupts normal cell differentiation (
). Consistently, heterozygous germline mutations in genes encoding the TCA cycle enzymes involved in the breakdown of these metabolites—SDH and FH—have been shown to facilitate the pathogenesis of certain rare cancers (e.g., paragangliomas) and inherited cancer-predisposition syndromes (
). In addition to affecting the activity of chromatin-associated αKG-dependent dioxygenases, succinate and fumarate accumulation can also inhibit αKG-dependent prolyl hydroxylases, leading to stabilization of HIF-1 and concomitant activation of the HIF/hypoxia transcriptional program (
). These tumors provide examples of how tumors can exploit TCA cycle metabolism to promote epigenetic and transcriptional remodeling that facilitates tumorigenesis.
Acetyl-CoA and histone acetylation
The TCA cycle provides another critical input to chromatin regulation in the form of acetyl-CoA. As discussed previously, several substrates feed into the TCA cycle at the level of acetyl-CoA, which condenses with OAA to form citrate. Acetyl-CoA itself cannot transit across the mitochondrial membrane. However, citrate export to the cytosol and subsequent cleavage by ACL provides a major source of cytosolic acetyl-CoA that is the obligate substrate for both de novo lipid synthesis and protein acetylation reactions (
). These findings suggest that activation of signaling pathways stimulates transcriptional remodeling in part by altering levels of TCA cycle metabolites.
TCA cycle metabolites in paracrine signaling
Beyond their intracellular roles, TCA cycle metabolites can also function as messengers facilitating interorgan crosstalk. To date, succinate represents the most established example of a TCA cycle intermediate that serves as a paracrine signal. During exercise, skeletal muscle responds to paracrine factors to undergo remodeling and tissue adaptation. Notably, both mouse and human muscle cells selectively release succinate during physical activity (
). In this system, two metabolic shifts—namely exercise-mediated acidification of muscle cells and succinate accumulation—coordinate to initiate a paracrine signaling cascade during physical activity. Succinate may also affect tissue function by directly controlling intracellular processes. For example, intracellular succinate accumulates rapidly upon cold temperature–mediated activation of thermogenesis in brown adipose tissue—in part because brown adipocytes begin sequestering circulating succinate more effectively (
). Succinate pools are directly oxidized by SDH, resulting in ROS production that stimulates uncoupling protein 1 activity, which in turn uncouples the mitochondrial proton gradient from respiration, thereby producing heat (
). Future work should continue to elucidate how succinate and other intracellular TCA cycle metabolites are converted to paracrine messengers that signal to the local environment and/or the rest of the organism.
TCA cycle heterogeneity
TCA cycle heterogeneity in cancer
Beyond inherent flexibility in reactions of the TCA cycle, different cell states and contexts also display variability in overall TCA cycle behavior, including preferred substrates for oxidation. Most mammalian cells growing in culture rely on both glucose- and glutamine-derived carbon to fuel the TCA cycle. However, isotope tracing studies in both lung and glioblastoma tumors suggest that glutamine is a relatively minor source of TCA cycle carbon in some cancer cells growing in vivo, indicating that environmental context can drive fuel preferences (
). Recent tracing studies have also revealed that TCA cycle metabolism in vivo is fueled by substrates beyond glucose and glutamine. As described previously, the contribution of circulating lactate to TCA cycle intermediates exceeds that of glucose in most tissues and tumors (
Beyond being driven by environmental differences between in vitro versus in vivo growth, TCA cycle substrate preferences are also determined by cancer tissue type and driver genetic mutations. For example, liver tumors with aberrant overexpression of MYC display enhanced glutamine oxidation and sensitivity to therapies that target glutamine catabolism relative to liver tumors driven by overexpression of MET (
). Clear cell renal carcinomas exhibit a tumor type–specific metabolic profile: these tumors do not oxidize glucose-derived carbon in the TCA cycle and instead preferentially engage in aerobic glycolysis (
). Altogether, advances in in vivo metabolic tracing are uncovering a wide array of metabolic strategies employed by tumors, and future work should continue to elucidate how oncogenes, lineage, environment, and cell state collectively determine metabolic phenotypes and liabilities of cancer cells.
TCA cycle heterogeneity in immune cells
The field of immunometabolism was ignited following the discovery that activated T-cells significantly upregulate glucose uptake and glycolysis to support the energetic and biosynthetic demands of T-cell expansion (
). Since this seminal finding, it has become increasingly appreciated that the TCA cycle also undergoes significant remodeling during immune cell activation and that this remodeling supports distinct immune cell functions and fates. For example, activated macrophages increase expression of the enzyme aconitate decarboxylase 1 (previously known as IRG1), which produces itaconate from the decarboxylation of TCA cycle–derived aconitate, an intermediate in the conversion of citrate to isocitrate (
). In response to T-cell receptor stimulation, 2HG accumulates to millimolar levels in CD8+ T cells and enhances effector differentiation through a mechanism involving altered histone and DNA methylation (
). In natural killer cells, cytokine activation induces Srebp-dependent metabolic reprogramming that drives metabolism of glucose-derived citrate through the citrate–malate shuttle; accordingly, blocking the citrate–malate shuttle or Srebp activation inhibits natural killer cell effector function (
). These are just a few examples of the enormous potential of TCA cycle reprogramming to reinforce immune cell function and alter immune cell fate. More examples of TCA cycle–driven immunomodulation can be found in recent reviews (
Adult tissue stem cells maintain organ homeostasis by balancing self-renewal and differentiation into mature cell types. Intriguingly, this functional plasticity may be accompanied by notable metabolic plasticity. In multiple tissue contexts, adult stem cells have been found to primarily engage in aerobic glycolysis. Upon lineage commitment and differentiation, however, cells increase their oxidation of pyruvate in the TCA cycle and display enhanced OXPHOS (
). For example, intestinal stem cells (ISCs) express lower levels of MPC than their differentiated progeny, and blocking mitochondrial pyruvate entry via MPC inhibition drives increased ISC numbers and proliferative potential (
). Collectively, these studies illustrate that TCA cycle substrate preference and behavior can vary within the same lineage in adult stem cells.
Pluripotent stem cells are emerging as a powerful system in which to study metabolism during cell-state transitions. While pluripotency, or the ability to give rise to all embryonic germ layers, is a state that exists only transiently during early mammalian development, it can be modeled indefinitely in vitro using ESCs derived from the inner cell mass of preimplantation blastocysts (
). Mouse and human ESCs can be cultured under conditions that drive a spectrum of pluripotent states, ranging from those that closely mimic the preimplantation state (otherwise known as the “naïve” state) to those that resemble the postimplantation epiblast. Increasing work is identifying metabolic alterations that correlate with developmental stages in both mouse and human, including variability in preferred substrate for oxidation in the TCA cycle and changes in steady-state levels of TCA cycle metabolites. While most cultured cell lines require glutamine supplementation to support cell survival and proliferation, naïve mouse ESCs can grow in the absence of exogenous glutamine. The glutamine-independent phenotype exhibited by naïve mouse ESCs is coupled with a reduction in the contribution of glutamine-derived carbons to TCA cycle intermediates (