Crystal structures of the human adiponectin receptors.

Adiponectin stimulation of its receptors, AdipoR1 and AdipoR2, increases the activities of 5' AMP-activated protein kinase (AMPK) and peroxisome proliferator-activated receptor (PPAR), respectively, thereby contributing to healthy longevity as key anti-diabetic molecules. AdipoR1 and AdipoR2 were predicted to contain seven transmembrane helices with the opposite topology to G-protein-coupled receptors. Here we report the crystal structures of human AdipoR1 and AdipoR2 at 2.9 and 2.4 Å resolution, respectively, which represent a novel class of receptor structure. The seven-transmembrane helices, conformationally distinct from those of G-protein-coupled receptors, enclose a large cavity where three conserved histidine residues coordinate a zinc ion. The zinc-binding structure may have a role in the adiponectin-stimulated AMPK phosphorylation and UCP2 upregulation. Adiponectin may broadly interact with the extracellular face, rather than the carboxy-terminal tail, of the receptors. The present information will facilitate the understanding of novel structure-function relationships and the development and optimization of AdipoR agonists for the treatment of obesity-related diseases, such as type 2 diabetes.

Rotation mechanism of Enterococcus hirae V1-ATPase based on asymmetric crystal structures.

In various cellular membrane systems, vacuolar ATPases (V-ATPases) function as proton pumps, which are involved in many processes such as bone resorption and cancer metastasis, and these membrane proteins represent attractive drug targets for osteoporosis and cancer. The hydrophilic V(1) portion is known as a rotary motor, in which a central axis DF complex rotates inside a hexagonally arranged catalytic A(3)B(3) complex using ATP hydrolysis energy, but the molecular mechanism is not well defined owing to a lack of high-resolution structural information. We previously reported on the in vitro expression, purification and reconstitution of Enterococcus hirae V(1)-ATPase from the A(3)B(3) and DF complexes. Here we report the asymmetric structures of the nucleotide-free (2.8 Å) and nucleotide-bound (3.4 Å) A(3)B(3) complex that demonstrate conformational changes induced by nucleotide binding, suggesting a binding order in the right-handed rotational orientation in a cooperative manner. The crystal structures of the nucleotide-free (2.2 Å) and nucleotide-bound (2.7 Å) V(1)-ATPase are also reported. The more tightly packed nucleotide-binding site seems to be induced by DF binding, and ATP hydrolysis seems to be stimulated by the approach of a conserved arginine residue. To our knowledge, these asymmetric structures represent the first high-resolution view of the rotational mechanism of V(1)-ATPase.

SARS-CoV 3CL protease cleaves its C-terminal autoprocessing site by novel subsite cooperativity.

The 3C-like protease (3CLpro) of severe acute respiratory syndrome coronavirus (SARS-CoV) cleaves 11 sites in the polyproteins, including its own N- and C-terminal autoprocessing sites, by recognizing P4-P1 and P1'. In this study, we determined the crystal structure of 3CLpro with the C-terminal prosequence and the catalytic-site C145A mutation, in which the enzyme binds the C-terminal prosequence of another molecule. Surprisingly, Phe at the P3' position [Phe(P3')] is snugly accommodated in the S3' pocket. Mutations of Phe(P3') impaired the C-terminal autoprocessing, but did not affect N-terminal autoprocessing. This difference was ascribed to the P2 residue, Phe(P2) and Leu(P2), in the C- and N-terminal sites, as follows. The S3' subsite is formed by Phe(P2)-induced conformational changes of 3CLpro and the direct involvement of Phe(P2) itself. In contrast, the N-terminal prosequence with Leu(P2) does not cause such conformational changes for the S3' subsite formation. In fact, the mutation of Phe(P2) to Leu in the C-terminal autoprocessing site abolishes the dependence on Phe(P3'). These mechanisms explain why Phe is required at the P3' position when the P2 position is occupied by Phe rather than Leu, which reveals a type of subsite cooperativity. Moreover, the peptide consisting of P4-P1 with Leu(P2) inhibits protease activity, whereas that with Phe(P2) exhibits a much smaller inhibitory effect, because Phe(P3') is missing. Thus, this subsite cooperativity likely exists to avoid the autoinhibition of the enzyme by its mature C-terminal sequence, and to retain the efficient C-terminal autoprocessing by the use of Phe(P2).

Cell-Free Protein Synthesis Using S30 Extracts from Escherichia coli RFzero Strains for Efficient Incorporation of Non-Natural Amino Acids into Proteins.

Cell-free protein synthesis is useful for synthesizing difficult targets. The site-specific incorporation of non-natural amino acids into proteins is a powerful protein engineering method. In this study, we optimized the protocol for cell extract preparation from the Escherichia coli strain RFzero-iy, which is engineered to lack release factor 1 (RF-1). The BL21(DE3)-based RFzero-iy strain exhibited quite high cell-free protein productivity, and thus we established the protocols for its cell culture and extract preparation. In the presence of 3-iodo-l-tyrosine (IY), cell-free protein synthesis using the RFzero-iy-based S30 extract translated the UAG codon to IY at various sites with a high translation efficiency of >90%. In the absence of IY, the RFzero-iy-based cell-free system did not translate UAG to any amino acid, leaving UAG unassigned. Actually, UAG was readily reassigned to various non-natural amino acids, by supplementing them with their specific aminoacyl-tRNA synthetase variants (and their specific tRNAs) into the system. The high incorporation rate of our RFzero-iy-based cell-free system enables the incorporation of a variety of non-natural amino acids into multiple sites of proteins. The present strategy to create the RFzero strain is rapid, and thus promising for RF-1 deletions of various E. coli strains genomically engineered for specific requirements.

Parallel homodimer structures of the extracellular domains of the voltage-gated sodium channel ß4 subunit explain its role in cell-cell adhesion.

Voltage-gated sodium channels (VGSCs) are transmembrane proteins required for the generation of action potentials in excitable cells and essential for propagating electrical impulses along nerve cells. VGSCs are complexes of a pore-forming α subunit and auxiliary ß subunits, designated as ß1/ß1B-ß4 (encoded by SCN1B-4B, respectively), which also function in cell-cell adhesion. We previously reported the structural basis for the trans homophilic interaction of the ß4 subunit, which contributes to its adhesive function. Here, using crystallographic and biochemical analyses, we show that the ß4 extracellular domains directly interact with each other in a parallel manner that involves an intermolecular disulfide bond between the unpaired Cys residues (Cys58) in the loop connecting strands B and C and intermolecular hydrophobic and hydrogen-bonding interactions of the N-terminal segments (Ser30-Val35). Under reducing conditions, an N-terminally deleted ß4 mutant exhibited decreased cell adhesion compared with the wild type, indicating that the ß4 cis dimer contributes to the trans homophilic interaction of ß4 in cell-cell adhesion. Furthermore, this mutant exhibited increased association with the α subunit, indicating that the cis dimerization of ß4 affects α-ß4 complex formation. These observations provide the structural basis for the parallel dimer formation of ß4 in VGSCs and reveal its mechanism in cell-cell adhesion.

Phosphorylated and non-phosphorylated HCK kinase domains produced by cell-free protein expression.

Since phosphorylation is involved in various physiological events, kinases and interacting factors can be potential targets for drug discovery. For the development and improvement of inhibitors from the point of view of mechanistic enzymology, a cell-free protein synthesis system would be advantageous, since it could prepare mutant proteins easily. However, especially in the case of protein kinase, product solubility remains one of the major challenges. To overcome this problem, we prepared a chaperone-supplemented extract from Escherichia coli BL21 cells harboring a plasmid encoding a set of chaperone genes, dnaK, dnaJ, and grpE. We explored cell-disruption procedures and constructed an efficient protein synthesis system. Employing this system, we produced the kinase domain of human hematopoietic cell kinase (HCK) to obtain further structural information about its molecular interaction with one of its inhibitors, previously developed by our group (RK-20449). Lower reaction temperature improved the solubility, and addition of a protein phosphatase (YpoH) facilitated the homogeneous production of the non-phosphorylated kinase domain. Crystals of the purified product were obtained and the kinase-inhibitor complex structure was solved at 1.7 Šresolution. In addition, results of kinase activity measurement, using a synthetic substrate, showed that the kinase activity was facilitated by autophosphorylation at Tyr416, as confirmed by the peptide mass mapping.

A redox switch shapes the Lon protease exit pore to facultatively regulate proteolysis.

The Lon AAA+ protease degrades damaged or misfolded proteins in its intramolecular chamber. Its activity must be precisely controlled, but the mechanism by which Lon is regulated in response to different environments is not known. Facultative anaerobes in the Enterobacteriaceae family, mostly symbionts and pathogens, encounter both anaerobic and aerobic environments inside and outside the host's body, respectively. The bacteria characteristically have two cysteine residues on the Lon protease (P) domain surface that unusually form a disulfide bond. Here we show that the cysteine residues act as a redox switch of Lon. Upon disulfide bond reduction, the exit pore of the P-domain ring narrows by ∼30%, thus interrupting product passage and decreasing activity by 80%; disulfide bonding by oxidation restores the pore size and activity. The redox switch (E°' = -227 mV) is appropriately tuned to respond to variation between anaerobic and aerobic conditions, thus optimizing the cellular proteolysis level for each environment.

Binding interactions of the peripheral stalk subunit isoforms from human V-ATPase.

The mammalian peripheral stalk subunits of the vacuolar-type H(+)-ATPases (V-ATPases) possess several isoforms (C1, C2, E1, E2, G1, G2, G3, a1, a2, a3, and a4), which may play significant role in regulating ATPase assembly and disassembly in different tissues. To better understand the structure and function of V-ATPase, we expressed and purified several isoforms of the human V-ATPase peripheral stalk: E1G1, E1G2, E1G3, E2G1, E2G2, E2G3, C1, C2, H, a1NT, and a2NT. Here, we investigated and characterized the isoforms of the peripheral stalk region of human V-ATPase with respect to their affinity and kinetics in different combination. We found that different isoforms interacted in a similar manner with the isoforms of other subunits. The differences in binding affinities among isoforms were minor from our in vitro studies. However, such minor differences from the binding interaction among isoforms might provide valuable information for the future structural-functional studies of this holoenzyme.

Expression, purification, crystallization, and preliminary X-ray crystallographic studies of the human adiponectin receptors, AdipoR1 and AdipoR2.

The adiponectin receptors (AdipoR1 and AdipoR2) are membrane proteins with seven transmembrane helices. These receptors regulate glucose and fatty acid metabolism, thereby ameliorating type 2 diabetes. The full-length human AdipoR1 and a series of N-terminally truncated mutants of human AdipoR1 and AdipoR2 were expressed in insect cells. In small-scale size exclusion chromatography, the truncated mutants AdipoR1Δ88 (residues 89-375) and AdipoR2Δ99 (residues 100-386) eluted mostly in the intact monodisperse state, while the others eluted primarily as aggregates. However, gel filtration chromatography of the large-scale preparation of the tag-affinity-purified AdipoR1Δ88 revealed the presence of an excessive amount of the aggregated state over the intact state. Since aggregation due to contaminating nucleic acids may have occurred during the sample concentration step, anion-exchange column chromatography was performed immediately after affinity chromatography, to separate the intact AdipoR1Δ88 from the aggregating species. The separated intact AdipoR1Δ88 did not undergo further aggregation, and was successfully purified to homogeneity by gel filtration chromatography. The purified AdipoR1Δ88 and AdipoR2Δ99 proteins were characterized by thermostability assays with 7-diethylamino-3-(4-maleimidophenyl)-4-methyl coumarin, thin layer chromatography of bound lipids, and surface plasmon resonance analysis of ligand binding, demonstrating their structural integrities. The AdipoR1Δ88 and AdipoR2Δ99 proteins were crystallized with the anti-AdipoR1 monoclonal antibody Fv fragment, by the lipidic mesophase method. X-ray diffraction data sets were obtained at resolutions of 2.8 and 2.4 Å, respectively.

Structural basis for mutual relief of the Rac guanine nucleotide exchange factor DOCK2 and its partner ELMO1 from their autoinhibited forms.

DOCK2, a hematopoietic cell-specific, atypical guanine nucleotide exchange factor, controls lymphocyte migration through ras-related C3 botulinum toxin substrate (Rac) activation. Dedicator of cytokinesis 2-engulfment and cell motility protein 1 (DOCK2•ELMO1) complex formation is required for DOCK2-mediated Rac signaling. In this study, we identified the N-terminal 177-residue fragment and the C-terminal 196-residue fragment of human DOCK2 and ELMO1, respectively, as the mutual binding regions, and solved the crystal structure of their complex at 2.1-Šresolution. The C-terminal Pro-rich tail of ELMO1 winds around the Src-homology 3 domain of DOCK2, and an intermolecular five-helix bundle is formed. Overall, the entire regions of both DOCK2 and ELMO1 assemble to create a rigid structure, which is required for the DOCK2•ELMO1 binding, as revealed by mutagenesis. Intriguingly, the DOCK2•ELMO1 interface hydrophobically buries a residue which, when mutated, reportedly relieves DOCK180 from autoinhibition. We demonstrated that the ELMO-interacting region and the DOCK-homology region 2 guanine nucleotide exchange factor domain of DOCK2 associate with each other for the autoinhibition, and that the assembly with ELMO1 weakens the interaction, relieving DOCK2 from the autoinhibition. The interactions between the N- and C-terminal regions of ELMO1 reportedly cause its autoinhibition, and binding with a DOCK protein relieves the autoinhibition for ras homolog gene family, member G binding and membrane localization. In fact, the DOCK2•ELMO1 interface also buries the ELMO1 residues required for the autoinhibition within the hydrophobic core of the helix bundle. Therefore, the present complex structure reveals the structural basis by which DOCK2 and ELMO1 mutually relieve their autoinhibition for the activation of Rac1 for lymphocyte chemotaxis.

Basic properties of rotary dynamics of the molecular motor Enterococcus hirae V1-ATPase.

V-ATPases are rotary molecular motors that generally function as proton pumps. We recently solved the crystal structures of the V1 moiety of Enterococcus hirae V-ATPase (EhV1) and proposed a model for its rotation mechanism. Here, we characterized the rotary dynamics of EhV1 using single-molecule analysis employing a load-free probe. EhV1 rotated in a counterclockwise direction, exhibiting two distinct rotational states, namely clear and unclear, suggesting unstable interactions between the rotor and stator. The clear state was analyzed in detail to obtain kinetic parameters. The rotation rates obeyed Michaelis-Menten kinetics with a maximal rotation rate (Vmax) of 107 revolutions/s and a Michaelis constant (Km) of 154 µM at 26 °C. At all ATP concentrations tested, EhV1 showed only three pauses separated by 120°/turn, and no substeps were resolved, as was the case with Thermus thermophilus V1-ATPase (TtV1). At 10 µM ATP (<>Km), the distribution of the durations of the catalytic pause was reproduced by a consecutive reaction with two time constants of 2.6 and 0.5 ms. These kinetic parameters were similar to those of TtV1. Our results identify the common properties of rotary catalysis of V1-ATPases that are distinct from those of F1-ATPases and will further our understanding of the general mechanisms of rotary molecular motors.

Crystal structure of the central axis DF complex of the prokaryotic V-ATPase.

V-ATPases function as ATP-dependent ion pumps in various membrane systems of living organisms. ATP hydrolysis causes rotation of the central rotor complex, which is composed of the central axis D subunit and a membrane c ring that are connected by F and d subunits. Here we determined the crystal structure of the DF complex of the prokaryotic V-ATPase of Enterococcus hirae at 2.0-Å resolution. The structure of the D subunit comprised a long left-handed coiled coil with a unique short ß-hairpin region that is effective in stimulating the ATPase activity of V(1)-ATPase by twofold. The F subunit is bound to the middle portion of the D subunit. The C-terminal helix of the F subunit, which was believed to function as a regulatory region by extending into the catalytic A(3)B(3) complex, contributes to tight binding to the D subunit by forming a three-helix bundle. Both D and F subunits are necessary to bind the d subunit that links to the c ring. From these findings, we modeled the entire rotor complex (DFdc ring) of V-ATPase.

Crystallographic and mutational studies on the tRNA thiouridine synthetase TtuA.

In thermophilic bacteria, specific 2-thiolation occurs on the conserved ribothymidine at position 54 (T54) in tRNAs, which is necessary for survival at high temperatures. T54 2-thiolation is achieved by the tRNA thiouridine synthetase TtuA and sulfur-carrier proteins. TtuA has five conserved CXXC/H motifs and the signature PP motif, and belongs to the TtcA family of tRNA 2-thiolation enzymes, for which there is currently no structural information. In this study, we determined the crystal structure of a TtuA homolog from the hyperthermophilic archeon Pyrococcus horikoshii at 2.1 Å resolution. The P. horikoshii TtuA forms a homodimer, and each subunit contains a catalytic domain and unique N- and C-terminal zinc fingers. The catalytic domain has much higher structural similarity to that of another tRNA modification enzyme, TilS (tRNA(Ile)2 lysidine synthetase), than to the other type of tRNA 2-thiolation enzyme, MnmA. Three conserved cysteine residues are clustered in the putative catalytic site, which is not present in TilS. An in vivo mutational analysis in the bacterium Thermus thermophilus demonstrated that the three conserved cysteine residues and the putative ATP-binding residues in the catalytic domain are important for the TtuA activity. A positively charged surface that includes the catalytic site and the two zinc fingers is likely to provide the tRNA-binding site.

Structural basis for the dual RNA-recognition modes of human Tra2-ß RRM.

Human Transformer2-ß (hTra2-ß) is an important member of the serine/arginine-rich protein family, and contains one RNA recognition motif (RRM). It controls the alternative splicing of several pre-mRNAs, including those of the calcitonin/calcitonin gene-related peptide (CGRP), the survival motor neuron 1 (SMN1) protein and the tau protein. Accordingly, the RRM of hTra2-ß specifically binds to two types of RNA sequences [the CAA and (GAA)(2) sequences]. We determined the solution structure of the hTra2-ß RRM (spanning residues Asn110-Thr201), which not only has a canonical RRM fold, but also an unusual alignment of the aromatic amino acids on the ß-sheet surface. We then solved the complex structure of the hTra2-ß RRM with the (GAA)(2) sequence, and found that the AGAA tetra-nucleotide was specifically recognized through hydrogen-bond formation with several amino acids on the N- and C-terminal extensions, as well as stacking interactions mediated by the unusually aligned aromatic rings on the ß-sheet surface. Further NMR experiments revealed that the hTra2-ß RRM recognizes the CAA sequence when it is integrated in the stem-loop structure. This study indicates that the hTra2-ß RRM recognizes two types of RNA sequences in different RNA binding modes.

Identification of critical residues in G(alpha)13 for stimulation of p115RhoGEF activity and the structure of the G(alpha)13-p115RhoGEF regulator of G protein signaling homology (RH) domain complex.

RH-RhoGEFs are a family of guanine nucleotide exchange factors that contain a regulator of G protein signaling homology (RH) domain. The heterotrimeric G protein Gα(13) stimulates the guanine nucleotide exchange factor (GEF) activity of RH-RhoGEFs, leading to activation of RhoA. The mechanism by which Gα(13) stimulates the GEF activity of RH-RhoGEFs, such as p115RhoGEF, has not yet been fully elucidated. Here, specific residues in Gα(13) that mediate activation of p115RhoGEF are identified. Mutation of these residues significantly impairs binding of Gα(13) to p115RhoGEF as well as stimulation of GEF activity. These data suggest that the exchange activity of p115RhoGEF is stimulated allosterically by Gα(13) and not through its interaction with a secondary binding site. A crystal structure of Gα(13) bound to the RH domain of p115RhoGEF is also presented, which differs from a previously crystallized complex with a Gα(13)-Gα(i1) chimera. Taken together, these data provide new insight into the mechanism by which p115RhoGEF is activated by Gα(13).

Crystal structure of the Ca²âº/calmodulin-dependent protein kinase kinase in complex with the inhibitor STO-609.

Ca(2+)/calmodulin (CaM)-dependent protein kinase (CaMK) kinase (CaMKK) is a member of the CaMK cascade that mediates the response to intracellular Ca(2+) elevation. CaMKK phosphorylates and activates CaMKI and CaMKIV, which directly activate transcription factors. In this study, we determined the 2.4 Å crystal structure of the catalytic kinase domain of the human CaMKKß isoform complexed with its selective inhibitor, STO-609. The structure revealed that CaMKKß lacks the αD helix and that the equivalent region displays a hydrophobic molecular surface, which may reflect its unique substrate recognition and autoinhibition. Although CaMKKß lacks the activation loop phosphorylation site, the activation loop is folded in an active-state conformation, which is stabilized by a number of interactions between amino acid residues conserved among the CaMKK isoforms. An in vitro analysis of the kinase activity confirmed the intrinsic activity of the CaMKKß kinase domain. Structure and sequence analyses of the STO-609-binding site revealed amino acid replacements that may affect the inhibitor binding. Indeed, mutagenesis demonstrated that the CaMKKß residue Pro(274), which replaces the conserved acidic residue of other protein kinases, is an important determinant for the selective inhibition by STO-609. Therefore, the present structure provides a molecular basis for clarifying the known biochemical properties of CaMKKß and for designing novel inhibitors targeting CaMKKß and the related protein kinases.

Structure of the Rho-specific guanine nucleotide-exchange factor Xpln.

Xpln is a guanine nucleotide-exchange factor (GEF) for Rho GTPases. A Dbl homology (DH) domain followed by a pleckstrin homology (PH) domain is a widely adopted GEF-domain architecture. The Xpln structure solely comprises these two domains. Xpln activates RhoA and RhoB, but not RhoC, although their GTPase sequences are highly conserved. The molecular mechanism of the selectivity of Xpln for Rho GTPases is still unclear. In this study, the crystal structure of the tandemly arranged DH-PH domains of mouse Xpln, with a single molecule in the asymmetric unit, was determined at 1.79 Šresolution by the multiwavelength anomalous dispersion method. The DH-PH domains of Xpln share high structural similarity with those from neuroepithelial cell-transforming gene 1 protein, PDZ-RhoGEF, leukaemia-associated RhoGEF and intersectins 1 and 2. The crystal structure indicated that the α4-α5 loop in the DH domain is flexible and that the DH and PH domains interact with each other intramolecularly, thus suggesting that PH-domain rearrangement occurs upon RhoA binding.

Crystal structure of Sulfolobus tokodaii Sua5 complexed with L-threonine and AMPPNP.

The hypermodified nucleoside N(6)-threonylcarbamoyladenosine resides at position 37 of tRNA molecules bearing U at position 36 and maintains translational fidelity in the three kingdoms of life. The N(6)-threonylcarbamoyl moiety is composed of L-threonine and bicarbonate, and its synthesis was genetically shown to require YrdC/Sua5. YrdC/Sua5 binds to tRNA and ATP. In this study, we analyzed the L-threonine-binding mode of Sua5 from the archaeon Sulfolobus tokodaii. Isothermal titration calorimetry measurements revealed that S. tokodaii Sua5 binds L-threonine more strongly than L-serine and glycine. The Kd values of Sua5 for L-threonine and L-serine are 9.3 µM and 2.6 mM, respectively. We determined the crystal structure of S. tokodaii Sua5, complexed with AMPPNP and L-threonine, at 1.8 Å resolution. The L-threonine is bound next to AMPPNP in the same pocket of the N-terminal domain. Thr118 and two water molecules form hydrogen bonds with AMPPNP in a unique manner for adenine-specific recognition. The carboxyl group and the side-chain hydroxyl and methyl groups of L-threonine are buried deep in the pocket, whereas the amino group faces AMPPNP. The L-threonine is located in a suitable position to react together with ATP for the synthesis of N(6)-threonylcarbamoyladenosine.

Structural basis for compound C inhibition of the human AMP-activated protein kinase α2 subunit kinase domain.

AMP-activated protein kinase (AMPK) is a serine/threonine kinase that functions as a sensor to maintain energy balance at both the cellular and the whole-body levels and is therefore a potential target for drug design against metabolic syndrome, obesity and type 2 diabetes. Here, the crystal structure of the phosphorylated-state mimic T172D mutant kinase domain from the human AMPK α2 subunit is reported in the apo form and in complex with a selective inhibitor, compound C. The AMPK α2 kinase domain exhibits a typical bilobal kinase fold and exists as a monomer in the crystal. Like the wild-type apo form, the T172D mutant apo form adopts the autoinhibited structure of the `DFG-out' conformation, with the Phe residue of the DFG motif anchored within the putative ATP-binding pocket. Compound C binding dramatically alters the conformation of the activation loop, which adopts an intermediate conformation between DFG-out and DFG-in. This induced fit forms a compound-C binding pocket composed of the N-lobe, the C-lobe and the hinge of the kinase domain. The pocket partially overlaps with the putative ATP-binding pocket. These three-dimensional structures will be useful to guide drug discovery.

Inhibitor-bound structures of human pyruvate dehydrogenase kinase 4.

The mitochondrial pyruvate dehydrogenase complex (PDC) catalyzes the oxidative decarboxylation of pyruvate to acetyl-CoA. PDC activity is tightly regulated by four members of a family of pyruvate dehydrogenase kinase isoforms (PDK1-4), which phosphorylate and inactivate PDC. Recently, the development of specific inhibitors of PDK4 has become an especially important focus for the pharmaceutical management of diabetes and obesity. In this study, crystal structures of human PDK4 complexed with either AMPPNP, ADP or the inhibitor M77976 were determined. ADP-bound PDK4 has a slightly wider active-site cleft and a more disordered ATP lid compared with AMPPNP-bound PDK4, although both forms of PDK4 assume open conformations with a wider active-site cleft than that in the closed conformation of the previously reported ADP-bound PDK2 structure. M77976 binds to the ATP-binding pocket of PDK4 and causes local conformational changes with complete disordering of the ATP lid. M77976 binding also leads to a large domain rearrangement that further expands the active-site cleft of PDK4 compared with the ADP- and AMPPNP-bound forms. Biochemical analyses revealed that M77976 inhibits PDK4 with increased potency compared with the previously characterized PDK inhibitor radicicol. Thus, the present structures demonstrate for the first time the flexible and dynamic aspects of PDK4 in the open conformation and provide a basis for the development of novel inhibitors targeting the nucleotide-binding pocket of PDK4.