Approximately one million people die from malaria each year, and thus this disease is a prevalent and pressing matter. Plasmodium falciparum, the protozoan parasite responsible for malaria has been subject to intensive study with the goal of developing specific drugs for malaria treatment. Plasmodium falciparum has a distinctive Malate Dehydrogenase (MDH), being tetrameric compared to the canonical dimeric forms of MDH, yet shares the same catalytic and substrate binding site motives as human mitochondrial and cytosolic forms. To investigate the possible existence of a cryptic allosteric site in the P falciparum enzyme, we have used sequence and structural bioinformatics and identified several short sequences,(G175‐L176, Q211‐M216) each containing highly conserved (in Plasmodium and related species) amino acids that are different in the Plasmodium falciparum enzyme than the human mitochondrial or cytosolic forms. To investigate the possibility that these regions form a cryptic allosteric site in Plasmodium falciparum we expressed, purified, using NiNTA Affinity chromatography, and characterized all three forms (Plasmodium falciparum, human cytosolic and human mitochondrial malate dehydrogenases). In terms of kinetic parameters the Plasmodium falciparum enzyme resembled the mammalian mitochondrial isoform, exhibiting similar substrate inhibition with oxaloacetate and weaker affinity for NADH than the mammalian cytosolic isoform. To determine the potential that the Plasmodium falciparum enzyme had a cryptic allosteric site we designed two mutations, one in each of the two significant sequence differences between the human isoforms and the Plasmodium falciparum enzyme (.D176N and R214E ). Each mutation was constructed using Quikchange mutagenesis, transformed into XL Gold cells and expressed. The purified mutant proteins were characterized using enzyme kinetics and size exclusion chromatography and their overall secondary structures compared to the wildtype enzyme using circular dichroism spectroscopy. Stability was compared using CD thermal melts at 222nm and a Fluorescence based Thermal Shift assay. While both mutations showed altered specific activities compared to wildtype, the D176N mutant showed a 90% lower activity than wildtype Plasmodium falciparum malate dehydrogenase. Both mutations significantly lowered affinity for the cofactor NADH, >20fold decrease in affinity. Taken together these results suggest that Plasmodium falciparum malate dehydrogenase may have a highly conserved cryptic allosteric site distinct from any such site in the mammalian isoforms. This Cryptic Allosteric Site could be exploited to develop inhibitors of Plasmodium falciparum malate dehydrogenase that would not impact the human isoforms. Support or Funding Information This work was supported by NSF Grants 1726932 and 0448905.

Approximately a million people die from malaria yearly. Plasmodium falciparum, the protozoan parasite responsible, is a major target for developing specific antimalarial drugs. Plasmodium falciparum has a distinctive Malate Dehydrogenase (MDH), being tetrameric compared to the canonical dimeric forms, yet shares the same catalytic and substrate binding site motives as human mitochondrial and cytosolic forms. We identified a cryptic allosteric site on Plasmodium falciparum MDH using sequence and structural bioinformatics and identified two short sequences, (G175‐L176 (subsite 1), Q211‐M216 (subsite 2)) each containing highly conserved (in Plasmodium and related species) residues that are different from those in the human forms. We made two mutations, D176N and R214E. Both showed altered specific activities compared to wildtype, the D176N mutant showed a 90% lower activity. Plasmodium falciparum MDH had significantly lowered affinity for the cofactor NADH, (>20fold decrease). These results suggest that Plasmodium falciparum MDH may have a highly conserved cryptic allosteric site distinct from any such site in the mammalian isoforms, which could be exploited to develop inhibitors of Plasmodium falciparum MDH that would not impact the human isoforms. As a prelude to designing an allosteric ligand for this site we have investigated the impact of various mutations of each residue in the site on the pKa of the active site histidine and other titratable groups in the protein using H++ analysis. This demonstrated that 4 residues in the cryptic site contributed significantly to both the pKa of the catalytic base, H174 and other titratable groups, helping define the necessary properties of each side chain in the cryptic site that contribute to activity which could be exploited in allosteric drug design. To investigate the dimensions and surface properties of this site we used POCASA in conjunction with electrostatics analysis to detect pockets (wheat in figure), and a SwissDock approach based on Multiple Solvent Crystal Structure (MSCS) analysis to detect potential ligand interactions with the cryptic site. Using a panel of scaffolds and side chains from the ZINC data base, together with ligands from the original MSCS approach revealed a number of potential interactions with ligands in subsite 1 & 2 of the cryptic site. Specifically, adjacent to subsite 1, 8 ligands, CYPO, CHLM, ACEE, ACTO, ETHR, ETOH, Hexane & Acetamide bound, with an additional nearby site binding Isopropanol, BENZ, BROM, ETHR & ETOH while subsite 2 showed specificity for CHLM, ACET, BIPH, DPME, BENZ & ACTO with an additional nearby site binding ETOH. Comparative studies with human Cytosolic (green in figure) and Mitochondrial (blue in figure) isoforms indicated that neither isoform had a similar physical pocket (POCASA) and showed no significant binding specificities in the area of the cryptic site in the Plasmodium falciparum enzyme. In conclusion, we have demonstrated the presence of a unique cryptic allosteric site on Plasmodium falciparum Malate Dehy...

Most Malate Dehydrogenases are dimeric though there are a small number (for example the LDH like Apicomplexan Malate Dehydrogenases) with tetrameric structures. A variety of studies have shown the oligomeric structure to be important for the overall activity of the enzyme with roles from stability to allosteric interactions being suggested. To explore our overall hypothesis that regions of the subunit interface would be tailored to specific functions we initiated an examination of the cross interface interactions of malate dehydrogenases from a wide variety of organisms whose crystal structures are available. Using the Hawkdock MM/GBSA routine we determined the residues involved in the interface in each structure. For the dimer interface we found between 24 and 35 residues involved depending on isoform. Mapping these residues onto structures revealed 5 areas of interaction, I to V with regions II – V being spatially conserved in all structures (figure 1). Region I was found only in chloroplastic versions of MDH. While the general areas of contact are conserved the detailed nature of the interactions revealed significant differences including inclusion of a conserved repulsion in region II in all Isoforms. These observations led to a series of hypotheses concerning the roles of specific regions and residues in isoform families. We focused on the cross interface interactions potentially related to citrate regulation of the organelle isoforms, and identified a unique cross interface interaction between the sidechain of D87 and the helix dipole of helix T268‐286 on the opposite subunit adjacent to a D92‐Y273 interaction across the interface. The role of L269/S266 in this helix and adjacent loop which we hypothesized form a cross interface molecular switch, was explored using site directed mutagenesis which showed that mutation of either residue to alanine disrupted either oxaloacetate or NADH saturation, overall catalytic activity and citrate inhibition. These results suggest a cross interface network (figure 2) in organelle malate dehydrogenases. Comparison of interface residues from structures with and without Citrate bound show that region III and V show the largest changes induced by citrate with some residues showing greater and others lesser contributions to the interface and are consistent with the observed network of interactions and suggest that changes across the interface might be induced by citrate interacting with S266. Interestingly the tetrameric MDHs from Ignacoccus Islandicus and the apicomplexans shares many but not all of these residues, lacking in particular D92, T268 and Y273. Knock in mutants of the Ignacoccus Islandicus and the Plasmodium falciparum MDHs studied both computationally and with enzyme kinetics shed further insight on the potential roles of these residues and the evolution of cooperativity in MDH.

Enzymes complexes (metabolons) support the metabolic reactions through substrate channeling. Several enzymes in the Krebs cycle are found in protein‐protein metabolons including mitochondrial malate dehydrogenase (MDH2) and citric synthase (CS). While the weak transitory interaction between these enzymes has been demonstrated, the key residues of human MDH2 involved in the interface with CS has not been identified. The purpose of this study is to probe possible residues of MDH responsible for binding to CS. Because cytosolic MDH (MDH1) poorly binds to CS, we initially identified unique sequences that were involved in or adjacent to reported or predicted binding sites of MDH2 and CS. Four regions were selected and residues corresponding to the primary sequence of cytosolic MDH1 were substituted in place of mitochondrial MDH2. Computational models of all the monomeric constructs were made using phyre2 homology modeling, subsequent refinement of the monomers using Galaxy Refine and construction of the dimer forms using the appropriate template in PyMol. Dimers were then refined using Galaxy Refine Complex to give replicate structures of the final models. Potential metabolon formation was explored using HawkDock with the dimer versions of MDH and Citrate Synthase either with or without restraints imposed by published metabolon models. Wild‐type human genes of MDH1, MDH2, and CS were codon optimized for bacterial expression and cloned into the C‐terminus of each gene in a pET28a expression vector. Each MDH1‐MDH2 substitution (DS1‐DS4) were constructed by Gibson cloning. Resulting structure and functional impacts were determined by melting points and kinetic parameters (specific activity, Km, Vmax and Kcat) in purified proteins. Interactions between wild‐type MDH1/2 and CS were compared to MDH2 DS mutants. To qualitatively demonstrate interactions, pull‐down assays were performed in the absence and presence of crowding agents, glycerol, PEG or Ficoll 70. Finally wild‐type MDH 1 and 2 vs MDH2 mutants were subjected to competitive pull‐down assays to identify regions responsible for isozyme specific interactions. This work will show four of the potential interacting regions between MDH and CS and lead to a better understanding of dynamics of this metabolic pair.