Association of the malate dehydrogenase-citrate synthase metabolon is modulated by intermediates of the Krebs tricarboxylic acid cycle

Mitochondrial malate dehydrogenase (MDH)-citrate synthase (CS) multi-enzyme complex is a part of the Krebs tricarboxylic acid (TCA) cycle ‘metabolon’ which is enzyme machinery catalyzing sequential reactions without diffusion of reaction intermediates into a bulk matrix. This complex is assumed to be a dynamic structure involved in the regulation of the cycle by enhancing metabolic flux. Microscale Thermophoresis analysis of the porcine heart MDH-CS complex revealed that substrates of the MDH and CS reactions, NAD+ and acetyl-CoA, enhance complex association while products of the reactions, NADH and citrate, weaken the affinity of the complex. Oxaloacetate enhanced the interaction only when it was present together with acetyl-CoA. Structural modeling using published CS structures suggested that the binding of these substrates can stabilize the closed format of CS which favors the MDH-CS association. Two other TCA cycle intermediates, ATP, and low pH also enhanced the association of the complex. These results suggest that dynamic formation of the MDH-CS multi-enzyme complex is modulated by metabolic factors responding to respiratory metabolism, and it may function in the feedback regulation of the cycle and adjacent metabolic pathways.

www.nature.com/scientificreports/ The TCA cycle is very tightly regulated by various factors in the mitochondrial matrix 5 . The cycle is regulated primarily by the mitochondrial NAD + /NADH ratio which is directly related to the mitochondrial ATP/ADP ratio that signals energy level in the cell 13 . NADH inhibits all the regulatory enzymes of the cycle, it functions as a feedback inhibitor to downregulate the TCA cycle and further the ATP production 5,13 . Additionally, ample acetyl-CoA (which is a product of oxidative pathways including glycolysis and lipid β-oxidation) upregulates flux through the cycle as an allosteric regulator causing a conformational change to the CS structure to enhance catalytic activity 14 . Thus, activities of the TCA cycle enzymes are regulated by various products and substrates of oxidative pathways and anabolic processes. Therefore, it is reasonable to assume that these allosteric regulators control not only the enzyme activities but also transient multi-enzyme complex formation through conformational changes of individual enzymes. We hypothesize that metabolites whose levels fluctuate in the mitochondrial matrix depending on the metabolic flux through the TCA cycle and adjacent metabolic pathways regulate multi-enzyme complex formation as a means of feedback regulation of these pathways. In this work, we explore the effects of metabolites and pH on the affinity of the MDH-CS multi-enzyme complex in vitro. We also computationally evaluate the effect of ligand binding on already established CS structures and accessibility of the 65Arg and 67Arg residues for multi-enzyme complex formation. The results showed the effects of the metabolic intermediates on the affinity of the MDH-CS complex, suggestive of the dynamics and function of the multienzyme complex in the regulation of the TCA cycle and adjacent pathways.

Results
To determine the effect on the binding affinity of the MDH-CS complex, we assessed the affinity of the complex in the presence of various metabolites in the mitochondrial matrix whose concentrations are expected to alter depending on the respiratory activity. The K d value of MDH-CS interaction was determined by MST as 2.29 ± 0.46 µM in the control condition when CS was used as the ligand (Fig. 1A, green curve). No binding was observed when MDH was incubated with bovine serum albumin ( Fig. S1 in Supplementary Information, orange curve).

Effects of the metabolites involved in the reactions catalyzed by MDH and CS.
The presence of 10 mM of the substrates of MDH and CS reactions led to significant changes in the K d of the complex. The addition of acetyl-CoA and NAD + significantly reduced the K d to 1.10 ± 0.23 and 0.89 ± 0.26, while malate slightly but not significantly reduced K d to 1.80 ± 0.17 µM (Fig. 1A). On the other hand, products of these reactions, NADH and citrate, significantly increased the K d values to 7.50 ± 0.60 µM and 25.1 ± 6.3 µM, respectively (Fig. 1B). The effect of CoA on the affinity of this complex was not significant (K d 2.32 ± 1.15 µM; Fig. 1B). Thus, substrates of the MDH and CS reactions increased complex association while products of the reaction decreased it. MDH-CS multienzyme complex channels oxaloacetate by electropositive interaction 2,15 . Oxaloacetate did not alter the dissociation constant of this complex (K d = 2.23 ± 0.60 µM; Fig. 1C). We evaluated the effect of oxaloacetate, in the presence of NADH or acetyl-CoA which are other substrates involved in the enzyme reactions. The addition of NADH together with oxaloacetate ( Fig. 1C) did not significantly alter the binding affinity of MDH and CS (K d = 2.01 ± 0.79 µM). However, oxaloacetate together with acetyl-CoA significantly reduced the K d to 0.52 ± 0.11 µM (Fig. 1C) in comparison to those at both the control condition and the presence of acetyl-CoA. Hence, OAA altered the affinity of the multienzyme complex association only in the presence of acetyl-CoA. Alteration of CS structure by the interaction with acetyl-CoA and OAA. To gain insight into how two substrates of the CS reaction synergistically increased the MDH-CS binding affinity, we evaluated the effect of these intermediates on the structure and conformation of CS. Two previously established porcine CS structures, which were solved in the presence of citrate and citrate with CoA referred to as the open (PDB 1cts; Fig. 2A) and closed formats (PDB 2cts; Fig. 2C), respectively 16 , were used for computational analyses. These formats are believed to be the form to accept the substrates and release the products and that to carry out catalytic steps with its active site buried deep within the protein, respectively 17 . The bottom half domain was well conserved, and the top half was inconsistent between these two conformations (Fig. 2B). The distribution of the surface charges was reallocated between the open and closed formats ( Fig. 2D and E). While the patches are negatively charged and hydrophobic on the surface near the active site in the open format (Fig. 2D), positively charged patches become dominant in the closed format (Fig. 2E). Acetyl-CoA forms hydrogen bonds with the 274His and 375Asp at the CS active site as the first step in the reaction 18 . The structural modeling of the acetyl-CoA localization in the closed format of CS suggested that acetyl-CoA also forms hydrogen bonds with A46Arg and B164Arg (A and B represent the subunits of the CS homodimer) in the bottom conserved domain, and B164Arg is connected to A366Lys in the top mobile domain through acetyl-CoA (Fig. 2B). Thus, the presence of acetyl-CoA in the catalytic site of CS likely favors the closed format of CS protein by forming a hydrogen bond network.
MDH protein (PDB 1mld) is predicted to interact with open and closed CS differently (Fig. 2F&G). ZDOC score, a relative binding energy function calculated by ZDOC, of the 65Arg residue was larger in closed format than open format, indicating that CS in closed format binds better with MDH (Table S2 in Supplementary  Information). The closed CS also has a larger accessible surface area, buried surface area after binding, and more hydrogen bonds with MDH than the open format (Table S2 in Supplementary Information). These results indicate that the CS in the closed format favors MDH binding.
Acetyl-CoA also allosterically regulates CS activity 19 . Our structural modeling predicted one potential acetyl-CoA binding site in the CS apoenzyme in the open format that locates very close to the 65Arg and 67Arg involved in the interaction between MDH and CS (Fig. 2F).

Effects of other TCA cycle intermediates.
Since the TCA cycle intermediates are known to allosterically regulate the activities of the enzymes of the cycle, we tested the effect of these intermediates on the binding affinity of the MDH-CS multi-enzyme complex. Succinate and α-ketoglutarate significantly reduced the K d of the interaction to 0.18 ± 0.41 µM and 1.14 ± 0.21 µM, respectively (Table 1). No significant effect on the MDH-CS complex affinity was observed by the addition of fumarate and succinyl-CoA into the test solution (Table 1). Thus, succinate and α-ketoglutarate increased the binding affinity of this complex, while citrate decreased binding affinity (Table 1, Fig. 1B).

Effects of the indicators of mitochondrial energy status.
The TCA cycle is an amphibolic pathway. The catabolic role of this cycle is to generate reducing equivalents that are used in the electron transport chain for oxidative phosphorylation. The energy level in the cell is closely related to respiratory activities which must be constantly regulated in response to metabolic conditions. To determine whether energy molecules in the mitochondria have any effect on the binding affinity of the MDH-CS complex, we evaluated the binding affinity of the complex in the presence of 10 mM AMP, ADP, and ATP. AMP and ADP did not significantly affect the K d values (3.10 ± 0.62 µM and 2.24 ± 0.56 µM for AMP and ADP, respectively; Fig. 3A). There was a significant increase in the binding affinity of the MDH-CS complex in the presence of ATP with the K d reduced to 0.46 ± 0.03 µM (Fig. 3A). The reduction of K d was also observed at lower concentrations of ATP with a dosedependent trend; the K d value reduced with increasing ATP concentration (Fig. 3B). The presence of 1.25, 5, 10, and 20 mM of ATP significantly reduced the K d of MDH-CS interaction to 1.06 ± 0.26, 0.77 ± 0.30, 0.76 ± 0.11, and 0.72 ± 0.15 µM, respectively, in comparison to the control condition. The pH in the mitochondrial matrix changes in response to the electron transport chain activities which are closely related to the energy status 12 .
Resting matrix pH ranges between pH 7 and 8, and when protons diffuse back into the matrix to drive ATP synthesis, the matrix pH becomes more acidic down to pH 6 20 . To assess the effect of pH in the mitochondrial matrix on the MDH-CS complex affinity, we monitored binding affinity in the pH range between 6.0 and 8.0. MDH-CS interaction in pH 8 MST buffer (control) yielded a K d of 2.29 ± 0.46 µM (Fig. 3C). At pH 6 and 7, the K d values were 0.68 ± 0.05 µM and 1.60 ± 0.45 µM, respectively, and significant differences were observed when the control group (pH 8) was compared with pH 6.

Discussion
Here, we report the effects of metabolic intermediates as regulators of MDH-CS multi-enzyme complex formation in vitro. Many of the tested intermediates related to the TCA cycle either enhanced or reduced the affinity of the enzyme complex, suggesting their functions in the dynamic regulation of the TCA cycle metabolon in response to the metabolic status in the mitochondrial matrix. We tested the effects of metabolites at a high concentration of 10 mM to capture any possible effects of compounds. This is a much higher concentration than the reported concentrations of the metabolites in the mitochondria 21 , which is calculated with an assumption of equal metabolite distribution in the mitochondria. However, a higher concentration is achievable by local metabolite accumulation in microcompartments such as a metabolon [22][23][24] . It should be noted that the effects of these metabolites on the MDH-CS complex are compound-specific as some metabolites showed no effect even at this high concentration. Additionally, ATP had significant effects on MDH-CS binding even at the reported concentration range in mammalian mitochondria 21 . Thus, the observed effects of metabolites on MDH-CS interaction are likely reproducible in living cells.  Our results indicate that these metabolites regulate the TCA cycle flux via multiple mechanisms including the regulation of metabolon formation enabling the fine-tuning of the TCA cycle flux.
The regulation of the MDH-CS association by metabolites can contribute to the diurnal oscillation of the TCA cycle. Mitochondrial metabolism including the TCA cycle activity oscillates diurnally 25 leading to the rhythmic accumulation of metabolic intermediates in various organisms 26,27 . NAD + and ATP, which enhanced MDH-CS complex association in the current study, show diurnal oscillation and play a crucial role to synchronize oxidative metabolic pathways via the circadian control of enzyme acetylation by NAD + -dependent deacetylase in mice 27 . These metabolites may also contribute to enhancing the TCA cycle flux by increasing the affinity of the MDH-CS complex. Citrate accumulates in plant leaves during the night and the concentration peaks at dawn 26,28 . Hence, accumulated citrate may dissociate the MDH-CS complex to reduce the TCA cycle flux in daytime 29,30 .
It should be noted that the metabolites that affected the MDH-CS complex affinity in this study are also involved in other metabolic pathways. Therefore, the observed effects of the intermediates may also function in balancing metabolic fluxes through the TCA cycle and the related pathways. For instance, citrate can function to balance the metabolic flux through the TCA cycle and fatty acid biosynthesis. The results in this study indicate that citrate may downregulate the CS reaction by the inhibition of MDH-CS interaction. Accumulation of citrate also likely stimulates mitochondrial fatty acid synthesis by promoting the polymerization of acetyl-CoA carboxylase to activate it 31 . α-ketoglutarate and succinate enhanced the interaction of the MDH-CS multi-enzyme complex in this study. While these metabolites are TCA cycle intermediates, α-ketoglutarate is also derived from the transamination of amino acids 32 . Succinate is a substrate for succinate dehydrogenase of the electron transport chain (complex II), which couples the TCA cycle and the respiratory chain 33 . These metabolites both enter the TCA cycle as anaplerotic intermediates. These oxidative pathways are activated to produce ATP when the energy status is low 34 . Therefore, it is reasonable that these metabolites facilitate the MDH-CS interaction to enhance the TCA cycle flux to further enhance the production of reducing equivalent and ATP. OAA is another metabolite at a branching point of metabolic pathways. OAA is a substrate for aspartate aminotransferase (AAT) which converts OAA to aspartate using glutamate as the amino donor. As the AAT reaction mediates significantly higher flux than the TCA cycle at least in some mammalian tissues 35 , the TCA cycle competes for OAA with the AAT reaction in vivo. Accumulation of acetyl-CoA likely facilitates metabolic flux through the TCA cycle over the AAT pathway at this branch point by promoting MDH-CS complex formation and metabolite channeling of OAA to limit its accessibility to AAT 6 . Interestingly, the co-presence of both acetyl-CoA and OAA further enhances the interaction in this study, which most likely strengthens the re-direction of metabolic flux to the TCA cycle. Such a multi-step mechanism may enable the fine-tuning of the metabolic flux through the pathways.
The effects of CS substrates on the MDH-CS complex affinity observed in this study are partially explained by the conformational changes taking place when compounds are bound and released from the enzymes. The binding of acetyl-CoA to the substrate-binding site puts CS in the closed format which favors MDH binding. Additionally, dominant positive charges in the closed format likely enhance electrostatic interaction with MDH. This positive electrostatic potential extending over much of the surface of the protein probably also favors metabolite channeling by the electrostatic interactions with the negatively charged substrate 36,37 . These suggest that CS is in the closed format in the MDH-CS metabolon, while direct experimental evidence is missing.
OAA alone did not affect the binding affinity of the two enzymes. This agrees with the report of Tompa et al. 38 in which 10 -4 M OAA did not affect the enzyme interaction. The insignificant effect of OAA might be due to the location of the OAA binding site deep in the cleft of the catalytic active site (Fig. 2) which would have a limited Table 1. Effects of the TCA cycle intermediates on the affinity of the MDH-CS complex Data are presented as K d ± K d confidence. Asterisks represent metabolites that showed significant differences with no K d confidence overlap with the control. The data for citrate, oxaloacetate, and malate are the same as Fig. 1 www.nature.com/scientificreports/ effect on the transition to the closed format. On the other hand, acetyl-CoA makes hydrogen bonds with residues in the mobile domain and can contribute to stabilizing the closed format. In the presence of acetyl-CoA, OAA significantly increased the affinity of the MDH-CS complex likely by further stabilizing the close format in which the CS reaction takes place. Alternatively, this effect could be a result of a conformational change that occurs when acetyl-CoA binds to the CS active site. Acetyl-CoA is considered important for the proper positioning of OAA in the active site of CS. When acetyl-CoA is absent, CS is highly flexible and unstructured, which prevents OAA from proper positioning in the active site 39 . These can explain the need for acetyl-CoA in the function of OAA to modulate the interaction of the MDH-CS complex.
The structural modeling in this study also indicates that the binding of acetyl-CoA to potential allosteric sites can facilitate MDH-CS complex formation by bringing MDH residues closer to the CS arginine residues www.nature.com/scientificreports/ which contribute to the electrostatic forces responsible for the protein-protein interaction 2 . We observed that α-ketoglutarate, which has been reported as allosteric effectors of the CS activity 40-42 also significantly increased the binding affinity of the MDH-CS complex. Our simulation indicated that this compound did not share the same binding site as acetyl-CoA (Fig. S2 in Supplementary Information), indicating that the enhancing effect of ATP and α-ketoglutarate on the multi-enzyme complex formation is likely not based on the same mechanism as that of acetyl-CoA. Mitochondrial energy status can also affect the formation of the MDH-CS complex. It is reasonable since energy production is one of the significant outcomes of respiratory metabolism. ATP plays an important role in the regulation of metabolism, and it fluctuates with changes in metabolic conditions. It has been reported as an allosteric 41 and competitive inhibitor of CS for acetyl-CoA 40 . In this study, we observed a concentrationdependent positive effect of ATP on the binding affinity of the MDH-CS complex. Variations in the pH of the mitochondrial matrix reflect the pumping of protons by the respiratory chain and back-flux of protons across the ATP synthase 12 and range between pH 6 to 8 43,44 . Protons return to the matrix 'through' ATP synthase driving the synthesis of ATP 45 , and an increase in H + concentration will lead to a slight drop in pH. The significant difference in binding affinity of MDH and CS was observed when the pH 6 condition was compared to pH 8, suggesting that the MDH-CS complex formation is likely affected by the pH changes in the mitochondrial matrix that occur with ATP synthesis. Our results indicate that the energy-rich conditions can favor the interaction of MDH-CS, which is expected to further increase ATP production, although the metabolic and physiological relevance of this regulation is still unclear. These regulations might be related to the recycling of deuterium depleted water by the hydration reaction of the CS and other hydration enzymes of the TCA cycle 46 . Deuterium depleted water is proposed to activate the electron transport through mitochondria electron transport chain 47 . The enhancement of the MDH-CS interaction at high energy status may facilitate the recycling of deuterium depleted water to coordinate the mitochondrial electron transport chain and the TCA cycle, although this should be experimentally tested.
Our study showed that substrates of individual MDH and CS enzymes, which have been reported to increase the reaction rates, favor MDH-CS multi-enzyme complex association. On the other hand, reaction products that inhibit the forward reactions favor dissociation of the complex. The results also suggest that metabolites used for anabolic processes dissociate the complex while products of catabolic processes enhance it. These results indicate that the TCA cycle metabolon would dynamically interact depending on the accumulation of metabolic intermediates and is probably involved in a switch between catabolic and anabolic processes as a mechanism underlying metabolic network regulation. The dynamics of the MDH-CS complex association and its relationship with metabolic flux distributions in central carbon metabolism and local metabolite concentrations in the mitochondrial matrix should be analyzed in a living system to evaluate the metabolic functions of the dynamic metabolon.
Sample preparation. Fluorescence labeling of the target protein (MDH) was performed by Red N-hydroxy succinimide (NHS) 2nd generation (NanoTemper Technologies GmbH, München, Germany) following the manufacturer's protocol. Ammonium sulfate suspension of the MDH was centrifuged at 23,000 × g for 1 min at 4 °C and the precipitate was resuspended in the labeling buffer to gain 90 µL of 10 µM MDH and mixed with 10 µL of 300 µM of dye (NanoTemper Technologies) prepared in dimethyl sulfoxide to create 100 µl of dye-protein solution. This was incubated in the dark for 30 min at room temperature. Unbound fluorophores were removed by a desalting column using MST buffer as the running buffer. Labeled protein was diluted with MST buffer to obtain final fluorescent counts between 500-1000 FU. Fifty microlitres of ammonium sulfate suspension of the ligand protein (CS) was centrifuged and resuspended in 25 µl of the MST buffer to gain 30 µM CS solution. A twofold CS dilution series was prepared with the MST buffer across eight tubes. A series of eight dilutions was used because there was no binding at lower concentrations, and higher concentrations of CS were difficult to be achieved with the commercially available enzyme products. No binding could be detected when CS was labeled and MDH was used as the ligand protein. Ten microliters of the diluted ligand protein solution were mixed with an equal volume of the target protein solution to be analyzed. The effects of the metabolites on the affinity of the MDH-CS complex were examined by adding the metabolites to the MST buffer to be 10 mM final concentrations. To test the effects of pH, the MST buffer was prepared with pH 6 and 7 potassium phosphate buffers.
Microscale thermophoresis (MST) assay. MST experiments were performed by a Nano Temper Monolith NT.115 device (NanoTemper Technologies) using nano-red excitation. Samples were loaded into premiumcoated capillaries (NanoTemper Technologies) and analyzed with MST power 40%, excitation power 80%, and the time windows of 5 s before, 30 s during, and 5 s after the IR-laser. The temperature of the instrument was set to 23 °C for all measurements. Data analysis. All measurements were conducted with three replicates and the results were presented as mean values ± standard deviation. After recording the MST time traces, data were analyzed using the MO Affinity Analysis software (Nano Temper Technologies) by temperature jump (T-Jump) analysis, and the values obtained were normalized and plotted against the CS concentration. The dissociation constants (K d ) with confidence intervals were determined using the K d binding model. Graph Pad Prism 9.  (Table S1 in Supplementary Information). In the figures, only the first one structure in the first ten clusters is shown. Docking experiments were performed between MDH and either open or closed CS by the ZDOCK server 50 (https:// zdock. umass med. edu). The binding was performed to fix the CS structures and rotate the MDH structure. Both CS and MDH were dimerized via symmetry operation using PyMOL. Complexes with the contact surface near 65Arg and 67Arg residues were selected from the top ten binding mates out of 2000 configurations. Surface areas and suffice interactions were evaluated using Proteins, Interfaces, Structures, and Assemblies (PDBePISA) server 51 (https:// www. ebi. ac. uk/ pdbe/ pisa/).