Two structures of the catalytic domain of phosphorylase kinase: an active protein kinase complexed with substrate analogue and product

Activation mechanism of the MAP kinase ERK2 by dual phosphorylation

The structure of the active form of the MAP kinase ERK2 has been solved, phosphorylated on a threonine and a tyrosine residue within the phosphorylation lip. The lip is refolded, bringing the phosphothreonine and phosphotyrosine into alignment with surface arginine-rich binding sites. Conformational changes occur in the lip and neighboring structures, including the P+1 site, the MAP kinase insertion, the C-terminal extension, and helix C. Domain rotation and remodeling of the proline-directed P+1 specificity pocket account for the activation. The conformation of the P+1 pocket is similar to a second proline-directed kinase, CDK2-CyclinA, thus permitting the origin of this specificity to be defined. Conformational changes outside the lip provide loci at which the state of phosphorylation can be felt by other cellular components.

A binary complex of the catalytic subunit of cAMP-dependent protein kinase and adenosine further defines conformational flexibility

BACKGROUND

cAMP-dependent protein kinase (cAPK), a ubiquitous protein in eukaryotic cells, is one of the simplest members of the protein kinase family. It was the first protein kinase to be crystallized and continues to serve as a biochemical and structural prototype for this family of enzymes. To further understand the conformational changes that occur in different liganded and unliganded states of cAPK, the catalytic subunit of cAPK was crystallized in the absence of peptide inhibitor.

RESULTS

The crystal structure of the catalytic subunit of mouse recombinant cAPK (rC) complexed with adenosine was solved at 2.6 A resolution and refined to a crystallographic R factor of 21.9% with good stereochemical parameters. This is the first structure of the rC subunit that lacks a bound inhibitor or substrate peptide. The structure was solved by molecular replacement and comprises two lobes (large and small) which contain a number of conserved loops.

CONCLUSIONS

The binary complex of rC and adenosine adopts an 'intermediate' conformation relative to the previously described 'closed' and 'open' conformations of other rC complexes. Based on a comparison of these structures, the induced fit that is necessary for catalysis and closing of the active-site cleft appears to be confined to the small lobe, as in the absence of the peptide the conformation of the large lobe, including the peptide-docking surface, does not change. Three specific components contribute to the closing of the cleft: rotation of the small lobe; movement of the C-terminal tail; and closing of the so-called glycine-rich loop. There is no induced fit in the large lobe to accommodate the peptide and the closing of the cleft. A portion of the C-terminal tail, residues 315-334, serves as a gate for the entry or exit of the nucleotide into the hydrophobic active-site cleft.

Structure of an enzyme required for aminoglycoside antibiotic resistance reveals homology to eukaryotic protein kinases

Bacterial resistance to aminoglycoside antibiotics is almost exclusively accomplished through either phosphorylation, adenylylation, or acetylation of the antibacterial agent. The aminoglycoside kinase, APH(3')-IIIa, catalyzes the phosphorylation of a broad spectrum of aminoglycoside antibiotics. The crystal structure of this enzyme complexed with ADP was determined at 2.2 A. resolution. The three-dimensional fold of APH(3')-IIIa reveals a striking similarity to eukaryotic protein kinases despite a virtually complete lack of sequence homology. Nearly half of the APH(3')-IIIa sequence adopts a conformation identical to that seen in these kinases. Substantial differences are found in the location and conformation of residues presumably responsible for second-substrate specificity. These results indicate that APH(3') enzymes and eukaryotic-type protein kinases share a common ancestor.

Structural trees for protein superfamilies

Structural trees for large protein superfamilies, such as beta proteins with the aligned beta sheet packing, beta proteins with the orthogonal packing of alpha helices, two-layer and three-layer alpha/beta proteins, have been constructed. The structural motifs having unique overall folds and a unique handedness are taken as root structures of the trees. The larger protein structures of each superfamily are obtained by a stepwise addition of alpha helices and/or beta strands to the corresponding root motif, taking into account a restricted set of rules inferred from known principles of the protein structure. Among these rules, prohibition of crossing connections, attention to handedness and compactness, and a requirement for alpha helices to be packed in alpha-helical layers and beta strands in beta layers are the most important. Proteins and domains whose structures can be obtained by stepwise addition of alpha helices and/or beta strands to the same root motif can be grouped into one structural class or a superfamily. Proteins and domains found within branches of a structural tree can be grouped into subclasses or subfamilies. Levels of structural similarity between different proteins can easily be observed by visual inspection. Within one branch, protein structures having a higher position in the tree include the structures located lower. Proteins and domains of different branches have the structure located in the branching point as the common fold.

A highly specific inhibitor of human p38 MAP kinase binds in the ATP pocket

The crystal structure of human p38 mitogen-activated protein (MAP) kinase in complex with a potent and highly specific pyridinyl-imidazole inhibitor has been determined at 2.0 A resolution. The structure of the kinase, which is in its unphosphorylated state, is similar to that of the closely-related ERK2. The inhibitor molecule is bound in the ATP pocket. A hydrogen bond is made between the pyridyl nitrogen of the inhibitor and the main chain amido nitrogen of residue 109, analogous to the interaction from the N1 atom of ATP. The crystal structure provides possible explanations for the specificity of this class of inhibitors. Other protein kinase inhibitors may achieve their specificity through a similar mechanism. The structure also reveals a possible second binding site for this inhibitor, with currently unknown function.

Importance of the A-helix of the catalytic subunit of cAMP-dependent protein kinase for stability and for orienting subdomains at the cleft interface

All eukaryotic protein kinases share a conserved catalytic core. In the catalytic (C) subunit of cAMP-dependent protein kinase (cAPK) this core is preceded by a myristylation motif followed by a long helix with Trp 30 at the end of this A-helix filling a hydrophobic cavity between the two lobes of the core. To understand the importance of the A-helix, the myristylation motif (delta 1-14) as well as the entire N-terminal segment (delta 1 -39) were deleted. In addition, Trp 30 was replaced with both Tyr and Ala. All proteins were overexpressed in E. coli and purified to homogeneity. rC(delta 1-14), rC(W30Y), and rC(W30A) all had reduced thermostability, but were catalytically indistinguishable from wild-type C. Based on Surface Plasmon Resonance, all three also formed stable holoenzyme complexes with the RI-subunit, although the appKds were reduced by more than 10-fold due to decrease in the association rate. Surprisingly, however, the holoenzymes were even more thermostable than wild-type holoenzyme. To obtain active enzyme, it was necessary to purify rC(delta 1-39) as a fusion protein with glutathione-S-transferase (GST-rC(delta 1-39), although its thermostability (Tm) was decreased by 12.5 degrees C, was catalytically similar to wild-type C and was inhibited by both the type I and II R-subunits and the heat-stable protein kinase inhibitor (PKI). The Tm for holoenzyme II formed with GST-rC(delta 1-39) was 16.5 degrees C greater than the Tm for free GST-rC(delta 1-39), and the Ka(cAMP) was increased nearly 10-fold. These mutants point out striking and unanticipated differences in how the RI and RII subunits associate with the C-subunit to form a stable holoenzyme and indicate, furthermore, that this N-terminal segment, far from the active site cleft, influences those interactions. The importance of the A-helix and Trp 30 for stability correlates with its location at the cleft interface where it orients the C-helix in the small lobe and the activation loop in the large so that these subdomains are aligned in a way that allows for correct configuration of residues at the active site. This extensive network of contacts that links the A-helix directly to the active site in cAPK is compared to other kinases whose crystal structures have been solved.

Proximal regions of the catalytic gamma and regulatory beta subunits on the interior lobe face of phosphorylase kinase are structurally coupled to each other and with enzyme activation

Phosphorylase kinase from skeletal muscle is a hexadecameric enzyme with the subunit composition (alphabeta gammadelta)4 and a mass of 1.3 x 10(6) Da. The catalytic gamma subunit and the remaining regulatory subunits are packed as a tetrahedral structure composed of two elongated, opposing (alphabeta gammadelta)2 octameric lobes. We show by immunoelectron microscopy with subunit-specific monoclonal antibodies that a portion of the beta subunit occurs on the interior face of the lobes at a region of inter-lobal interactions, and that at a proximal position slightly more central and distal on the interior lobe face lies the base (residues 277 to 290) of the helical domain of the catalytic core of the gamma subunit. Activation of the kinase by a variety of means caused similar increases in the binding to the holoenzyme of the monoclonal antibodies against these two regions of the beta and gamma subunits. Moreover, monovalent fragments of the antibodies against both regions stimulated the activity of the non-activated holoenzyme. Thus, the epitopes of the beta and gamma subunits recognized by the monoclonal antibodies are structurally coupled to each other and with the activation of phosphorylase kinase. Activation of the holoenzyme apparently involves the repositioning of the base of the catalytic domain of the gamma subunit and a proximal region of the beta subunit within the identified area on the interior face of the lobes of the tetrahedral phosphorylase kinase molecule.

Control by phosphorylation

The two examples of phospho and dephospho proteins for which structural data were previously available (glycogen phosphorylase and isocitrate dehydrogenase) demonstrated two different mechanisms for control. In glycogen phosphorylase, activation by phosphorylation results in long-range allosteric changes. In isocitrate dehydrogenase, inhibition by phosphorylation is achieved by an electrostatic blocking mechanism with no conformational changes. During the past year, the structures of the phospho and dephospho forms of two more proteins, the cell cycle protein kinase CDK2 and yeast glycogen phosphorylase, have been determined. The new results highlight the importance of the phosphoamino acids both in the organization of local regions of protein structure through phosphate-arginine interactions and in the promotion of long-range conformational responses.

Mutations in the testis/liver isoform of the phosphorylase kinase gamma subunit (PHKG2) cause autosomal liver glycogenosis in the gsd rat and in humans

Heritable deficiency of phosphorylase kinase (Phk), a regulatory enzyme of glycogen metabolism, is responsible for 25% of all cases of glycogen storage disease and occurs with a frequency of -1 in 100,000 births. It is genetically and clinically heterogeneous, occurring in X-linked and autosomal-recessive forms and exhibiting various patterns of principally affected tissues (liver only, muscle only, liver and muscle, liver and kidney, heart only). This heterogeneity is thought to reflect the enzyme's structural complexity [subunit composition, (alpha beta gamma delta)4] and isoform diversity. Two isoforms encoded by separate genes are known for the subunits alpha (muscle [alpha M] and liver [alpha L isoforms) and gamma (muscle [gamma M] and testis [gamma T] isoforms), whereas only one gene appears to exist for the subunit beta. The subunit delta is calmodulin; identical calmodulins are expressed from three different human genes. Additional isoform diversity arises by differential mRNA splicing of the alpha M, alpha L and beta subunits. Mutations responsible for the various forms of Phk deficiency are sought in those subunit/isoform genes with a matching chromosomal location and tissue-specificity of expression. We report here that autosomal liver-specific Phk deficiency is associated with mutations in the gene encoding the testis/liver isoform of the catalytic gamma subunit (PHKG2). We found homozygous PHKG2 mutations in three human patients of consanguineous parentage and in the gsd (glycogen storage disease) rat strain, which is thus identified as an animal model for the human disorder. One human mutation is a single base-pair insertion in codon 89 that causes a frameshift and premature chain termination. The three other mutations result in non-conservative replacements of amino acid residues (V106E, G189E, D215N) that are highly conserved within the catalytic core regions of all protein kinases. These are the first mutations to be reported for an autosomal form of Phk deficiency. The findings suggest that the PHKG2 gene product is the predominant isoform of the catalytic gamma subunit of Phk not only in testis but also in liver, erythrocytes and, possibly, other non-muscle tissues.

Crystal structures of catalytic subunit of cAMP-dependent protein kinase in complex with isoquinolinesulfonyl protein kinase inhibitors H7, H8, and H89. Structural implications for selectivity

The discovery of several hundred different protein kinases involved in highly diverse cellular signaling pathways is in stark contrast to the much smaller number of known modulators of cell signaling. Of these, the H series protein kinase inhibitors (1-(5-isoquinolinesulfonyl)-2-methylpiperazine (H7), N-[2-(methylamino)ethyl]-5-isoquinolinesulfonamide (H8) N-[2-(p-Bromocinnamylamino)ethyl]-5-isoquinolinesulfonamide (H89)) are frequently used to block signaling pathways in studies of cellular regulation. To elucidate inhibition mechanisms at atomic resolution and to enable structure-based drug design of potential therapeutic modulators of signaling pathways, we determined the crystal structures of corresponding complexes with the cAPK catalytic subunit. Complexes with H7 and H8 (2.2 A) and with H89 (2.3 A) define the binding mode of the isoquinoline-sulfonamide derivatives in the ATP-binding site while demonstrating effects of ligand-induced structural change. Specific interactions between the enzyme and the inhibitors include the isoquinoline ring nitrogen ligating to backbone amide of Val-123 and an inhibitor side chain amide bonding to the backbone carbonyl of Glu-170. The conservation of the ATP-binding site of protein kinases allows evaluation of factors governing general selectivity of these inhibitors among kinases. These results should assist efforts in the design of protein kinase inhibitors with specific properties.

Structural aspects of the functional modules in human protein kinase-C alpha deduced from comparative analyses

Three-dimensional models of the five functional modules in human protein kinase C alpha (PKC alpha) have been generated on the basis of known related structures. The catalytic region at the C-terminus of the sequence and the N-terminal auto-inhibitory pseudo-substrate have been modeled using the crystal structure complex of cAMP-dependent protein kinase (cAPK) and PKI peptide. While the N-terminal helix of the catalytic region of PKC alpha is predicted to be in a different location compared with cAPK, the C-terminal extension is modeled like that in the cAPK. The predicted permissive phosphorylation site of PKC alpha, Thr 497, is found to be entirely consistent with the mutagenesis studies. Basic Lys and Arg residues in the pseudo-substrate make several specific interactions with acidic residues in the catalytic region and may interact with the permissive phosphorylation site. Models of the two zinc-binding modules of PKC alpha are based on nuclear magnetic resonance and crystal structures of such modules in other PKC isoforms while the calcium phospholipid binding module (C2) is based on the crystal structure of a repeating unit in synaptotagmin I. Phorbol ester binding regions in zinc-binding modules and the calcium binding region in the C2 domain are similar to those in the basis structures. A hypothetical model of the relative positions of all five modules has the putative lipid binding ends of the C2 and the two zinc-binding domains pointing in the same direction and may serve as a basis for further experiments.

Modern developments in molecular replacement

Molecular replacement is a possible route to obtaining initial phasing for an unknown structure from a known, structurally related molecule. Recent years have seen an explosive growth in the number of protein structures solved using this technique. Automated packages can make the application quite straightforward. Progress has been made in the placing of fragments of complexes, and in the use of imprecise models from NMR or homology modelling. Such models have necessitated the development of new approaches to rebuilding and refinement.

Activation and inhibition of phosphorylase kinase by monospecific antibodies raised against peptides from the regulatory domain of the gamma-subunit

The C terminus of the catalytic gamma-subunit of phosphorylase kinase comprises a regulatory domain that contains regions important for subunit interactions and autoinhibitory functions. Monospecific antibodies raised against four synthetic peptides from this region, PhK1 (362-386), PhK5 (342-366), PhK9 (322-346), and PhK13 (302-326), were found to have significant effects on the catalytic activities of phosphorylase kinase holoenzyme and the gamma delta complex. Antibodies raised against the very C terminus of the gamma-subunit, anti-PhK1 and anti-PhK5, markedly activated both holoenzyme and the gamma delta complex, in the presence and absence of Ca2+. In the presence of Ca2+ at pH 8.2, anti-PhK1 activated the holoenzyme more than 11-fold and activated the gamma delta complex 2.5-fold. Activation of the holoenzyme and the gamma delta complex by anti-PhK5 was 50-70% of that observed with anti-PhK1. Prior phosphorylation of the holoenzyme by the cAMP-dependent protein kinase blocked activation by both anti-PhK1 and anti-PhK5. Antibodies raised against the peptides from the N terminus of the regulatory domain, anti-PhK9 and anti-PhK13, were inhibitory, with their greatest effects on the gamma delta complex. These data demonstrate that the binding of antibodies to specific regions within the regulatory domain of the gamma-subunit can augment or inhibit structural changes and subunit interactions important in regulating phosphorylase kinase activity.

In vivo activation of recombinant cAPK catalytic subunit active site mutants by coexpression of the wild-type enzyme, evidence for intermolecular cotranslational phosphorylation

The catalytic subunit of cAMP dependent protein kinase (cAPK) carries two stable autophosphorylated residues. One of them, Thr197, resides in the so-called protein kinase activation segment, and needs to be phosphorylated for full activity and protein kinase inhibitor binding of the enzyme. While wild-type recombinant mammalian C-subunit, expressed in E. coli, can fully autoactivate itself by phosphorylation at Thr197, many active site mutants lack this autophosphorylation activity, so that the primary effects of the mutations become obscured. Two active site mutants of bovine C-subunit, defective in protein kinase inhibitor peptide binding, were activated by wild-type enzyme in vivo, but could not be activated in vitro, demonstrating intermolecular and presumably cotranslational autophosphorylation. The results may delineate strategies for the expression and mutagenesis of other protein kinases with requirements for activation segment phosphorylation.

Structural basis for selectivity of the isoquinoline sulfonamide family of protein kinase inhibitors

A large family of isoquinoline sulfonamide compounds inhibits protein kinases by competing with adenosine triphosphates(ATP), yet interferes little with the activity of other ATP-using enzymes such as ATPases and adenylate cyclases. One such compound, N-(2-aminoethyl)-5-chloroisoquinoline-8-sulfonamide (CK17), is selective for casein kinase-1 isolated from a variety of sources. Here we report the crystal structure of the catalytic domain of Schizosaccharomyces pombe casein kinase-1 complexed with CK17, refined to a crystallographic R-factor of 17.8% at 2.5 angstrom resolution. The structure provides new insights into the mechanism of the ATP-competing inhibition and the origin of their selectivity toward different protein kinases. Selectivity for protein kinases versus other enzymes is achieved by hydrophobic contacts and the hydrogen bond with isoquinoline ring. We propose that the hydrogen bond involving the ring nitrogen-2 atom of the isoquinoline must be preserved, but that the ring can flip depending on the chemical substituents at ring positions 5 and 8. Selectivity for individual members of the protein kinase family is achieved primarily by interactions with these substituents.

Modelling study of protein kinase inhibitors: binding mode of staurosporine and origin of the selectivity of CGP 52411

A model for the binding mode of the potent protein kinase inhibitor staurosporine is proposed. Using the information provided by the crystal structure of the cyclic-AMP-dependent protein kinase, it is suggested that staurosporine, despite a seemingly unrelated chemical structure, exploits the same key hydrogen-bond interactions as ATP, the cofactor of the protein kinases, in its binding mode. The structure-activity relationship of the inhibitor and a docking analysis give strong support to this hypothesis. The selectivity of the dianilinophthalimide inhibitor CGP 52411 towards the EGF-receptor protein tyrosine kinase is rationalized on the basis of the model. It is proposed that this selectivity originates in the occupancy, by one of the anilino moieties of the inhibitor, of the region of the enzyme cleft that normally binds the ribose ring of ATP, which appears to possess a marked lipophilic character in this kinase.

The crystal structure of cyclin A

BACKGROUND

Eukaryotic cell cycle progression is regulated by cyclin dependent protein kinases (CDKs) whose activity is regulated by association with cyclins and by reversible phosphorylation. Cyclins also determine the subcellular location and substrate specificity of CDKs. Cyclins exhibit diverse sequences but all share homology over a region of approximately 100 amino acids, termed the cyclin box. From the determination of the structure of cyclin A, together with results from biochemical and genetic analyses, we can identify which parts of the cyclin molecular may contribute to cyclin A structure and function.

RESULTS

We have solved the crystal structure, at 2.0 A resolution, of an active recombinant fragment of bovine cyclin A, cyclin A-3, corresponding to residues 171-432 of human cyclin A. The cyclin box has an alpha-helical fold comprising five alpha helices. This fold is repeated in the C-terminal region, although this region shares negligible sequence similarity with the cyclin box.

CONCLUSIONS

Analysis of residues that are conserved throughout the A, B, and E cyclins identifies two exposed clusters of residues, one of which has recently been shown to be involved in the association with human CDK2. The second cluster may identify another site of cyclin A-protein interaction. Comparison of the structure of the unbound cyclin with the structure of cyclin A complexed with CDK2 reveals that cyclin A does not undergo any significant conformational changes on complex formation. Threading analysis shows that the cyclin-box fold is consistent with the sequences of the transcription factor TFIIB and other functionally related proteins. The structural results indicate a role for the cyclin-box fold both as a template for the cyclin family and as a generalised adaptor molecule in the regulation of transcription.

Mapping the transition state for ATP hydrolysis: implications for enzymatic catalysis

BACKGROUND

Phosphoryl transfer, typically involving high energy phosphate donors such as ATP, is the most common class of biological reactions. Despite this, the transition state for phosphoryl transfer from ATP in solution has not been systematically investigated. Characterization of the transition state for the uncatalyzed hydrolysis of ATP would provide a starting point for dissection of enzyme-catalyzed reactions.

RESULTS

We examined phosphoryl transfer from ATP, GTP and pyrophosphate to a series of alcohols; these reactions are analogous to the phosphorylation of sugars and other biological alcohols and to the hydrolysis of ATP. The Brønsted beta(nucleophile) value of 0.07 is small, indicating that there is little bond formation between the incoming nucleophile and the electrophilic phosphoryl group in the transition state. Coordination of Mg2+ has no measurable effect on this value. The Brønsted beta(leaving group) value of -1.1 for phosphoryl transfer to water from a series of phosphoanhydrides is large and negative, suggesting that the bond between phosphorous and the leaving group oxygen is largely broken in the transition state.

CONCLUSIONS

Uncatalyzed hydrolysis of ATP in solution occurs via a dissociative, metaphosphate-like transition state, with little bond formation between nucleophile and ATP and substantial cleavage of the bond between the gamma-phosphoryl moiety and the ADP leaving group. Bound Mg2+ does not perturb the dissociative nature of the transition state, contrary to proposals that enzyme-bound metal ions alter this structure. The simplest expectation for phosphoryl transfer at the active site of enzymes thus entails a dissociative transition state. These results provide a basis for analyzing catalytic mechanisms for phosphoryl transfer.