Crystal structure of the alkylsulfatase AtsK: insights into the catalytic mechanism of the Fe(II) alpha-ketoglutarate-dependent dioxygenase superfamily

The alkylsulfatase AtsK from Pseudomonas putida S-313 belongs to the widespread and versatile non-heme iron(II) alpha-ketoglutarate-dependent dioxygenase superfamily and catalyzes the oxygenolytic cleavage of a variety of different alkyl sulfate esters to the corresponding aldehyde and sulfate. The enzyme is only expressed under sulfur starvation conditions, providing a selective advantage for bacterial growth in soils and rhizosphere. Here we describe the crystal structure of AtsK in the apo form and in three complexes: with the cosubstrate alpha-ketoglutarate, with alpha-ketoglutarate and iron, and finally with alpha-ketoglutarate, iron, and an alkyl sulfate ester used as substrate in catalytic studies. The overall fold of the enzyme is closely related to that of the taurine/alpha-ketoglutarate dioxygenase TauD and is similar to the fold observed for other members of the enzyme superfamily. From comparison of these structures with the crystal structure of AtsK and its complexes, we propose a general mechanism for the catalytic cycle of the alpha-ketoglutarate-dependent dioxygenase superfamily.     

The crystal structures of Zea mays and Arabidopsis 4-hydroxyphenylpyruvate dioxygenase

The transformation of 4-hydroxyphenylpyruvate to homogentisate, catalyzed by 4-hydroxyphenylpyruvate dioxygenase (HPPD), plays an important role in degrading aromatic amino acids. As the reaction product homogentisate serves as aromatic precursor for prenylquinone synthesis in plants, the enzyme is an interesting target for herbicides. In this study we report the first x-ray structures of the plant HPPDs of Zea mays and Arabidopsis in their substrate-free form at 2.0 A and 3.0 A resolution, respectively. Previous biochemical characterizations have demonstrated that eukaryotic enzymes behave as homodimers in contrast to prokaryotic HPPDs, which are homotetramers. Plant and bacterial enzymes share the overall fold but use orthogonal surfaces for oligomerization. In addition, comparison of both structures provides direct evidence that the C-terminal helix gates substrate access to the active site around a nonheme ferrous iron center. In the Z. mays HPPD structure this helix packs into the active site, sequestering it completely from the solvent. In contrast, in the Arabidopsis structure this helix tilted by about 60 degrees into the solvent and leaves the active site fully accessible. By elucidating the structure of plant HPPD enzymes we aim to provide a structural basis for the development of new herbicides.     

Succinate complex crystal structures of the alpha-ketoglutarate-dependent dioxygenase AtsK: steric aspects of enzyme self-hydroxylation

The alkylsulfatase AtsK from Pseudomonas putida S-313 is a member of the non-heme iron(II)-alpha-ketoglutarate-dependent dioxygenase superfamily. In the initial step of their catalytic cycle, enzymes belonging to this widespread and versatile family coordinate molecular oxygen to the iron center in the active site. The subsequent decarboxylation of the cosubstrate alpha-ketoglutarate yields carbon dioxide, succinate, and a highly reactive ferryl (IV) species, which is required for substrate oxidation via a complex mechanism involving the transfer of radical species. Non-productive activation of oxygen may lead to harmful side reactions; therefore, such enzymes need an effective built-in protection mechanism. One of the ways of controlling undesired side reactions is the self-hydroxylation of an aromatic side chain, which leads to an irreversibly inactivated species. Here we describe the crystal structure of the alkylsulfatase AtsK in complexes with succinate and with Fe(II)/succinate. In the crystal structure of the AtsK-Fe(II)-succinate complex, the side chain of Tyr(168) is co-ordinated to the iron, suggesting that Tyr(168) is the target of enzyme self-hydroxylation. This is the first structural study of an Fe(II)-alpha-ketoglutarate-dependent dioxygenase that presents an aromatic side chain coordinated to the metal center, thus allowing structural insight into this protective mechanism of enzyme self-inactivation.     

Crystal structure of mammalian cysteine dioxygenase. A novel mononuclear iron center for cysteine thiol oxidation

Cysteine dioxygenase is a mononuclear iron-dependent enzyme responsible for the oxidation of cysteine with molecular oxygen to form cysteine sulfinate. This reaction commits cysteine to either catabolism to sulfate and pyruvate or the taurine biosynthetic pathway. Cysteine dioxygenase is a member of the cupin superfamily of proteins. The crystal structure of recombinant rat cysteine dioxygenase has been determined to 1.5-A resolution, and these results confirm the canonical cupin beta-sandwich fold and the rare cysteinyltyrosine intramolecular cross-link (between Cys(93) and Tyr(157)) seen in the recently reported murine cysteine dioxygenase structure. In contrast to the catalytically inactive mononuclear Ni(II) metallocenter present in the murine structure, crystallization of a catalytically competent preparation of rat cysteine dioxygenase revealed a novel tetrahedrally coordinated mononuclear iron center involving three histidines (His(86), His(88), and His(140)) and a water molecule. Attempts to acquire a structure with bound ligand using either cocrystallization or soaking crystals with cysteine revealed the formation of a mixed disulfide involving Cys(164) near the active site, which may explain previously observed substrate inhibition. This work provides a framework for understanding the molecular mechanisms involved in thiol dioxygenation and sets the stage for exploration of the chemistry of both the novel mononuclear iron center and the catalytic role of the cysteinyl-tyrosine linkage.     

X-ray crystal structure of Escherichia coli taurine/alpha-ketoglutarate dioxygenase complexed to ferrous iron and substrates

Taurine/alpha-ketoglutarate dioxygenase (TauD), a non-heme Fe(II) oxygenase, catalyses the conversion of taurine (2-aminoethanesulfonate) to sulfite and aminoacetaldehyde concurrent with the conversion of alpha-ketoglutarate (alphaKG) to succinate and CO(2). The enzyme allows Escherichia coli to use taurine, widely available in the environment, as an alternative sulfur source. Here we describe the X-ray crystal structure of TauD complexed to Fe(II) and both substrates, alphaKG and taurine. The tertiary structure and fold of TauD are similar to those observed in other enzymes from the broad family of Fe(II)/alphaKG-dependent oxygenases, with closest structural similarity to clavaminate synthase. Using the TauD coordinates, a model was determined for the closely related enzyme 2,4-dichlorophenoxyacetate/alphaKG dioxygenase (TfdA), supporting predictions derived from site-directed mutagenesis and other studies of that biodegradative protein. The TauD structure and TfdA model define the metal ligands and the positions of nearby aromatic residues that undergo post-translational modifications involving self-hydroxylation reactions. The substrate binding residues of TauD were identified and those of TfdA predicted. These results, along with sequence alignment information, reveal how TauD selects a tetrahedral substrate anion in preference to the planar carboxylate selected by TfdA, providing insight into the mechanism of enzyme catalysis.     

Direct evidence for a glutamate switch necessary for substrate recognition: crystal structures of lysine epsilon-aminotransferase (Rv3290c) from Mycobacterium tuberculosis H37Rv

Lysine epsilon-aminotransferase (LAT) is a PLP-dependent enzyme that is highly up-regulated in nutrient-starved tuberculosis models. It catalyzes an overall reaction involving the transfer of the epsilon-amino group of L-lysine to alpha-ketoglutarate to yield L-glutamate and alpha-aminoadipate-delta-semialdehyde. We have cloned and characterized the enzyme from Mycobacterium tuberculosisH37Rv. We report here the crystal structures of the enzyme, the first from any source, in the unliganded form, external aldimine with L-lysine, with bound PMP and with its C5 substrate alpha-ketoglutarate. In addition to interaction details in the active site, the structures reveal a Glu243 "switch" through which the enzyme changes substrate specificities. The unique substrate L-lysine is recognized specifically when Glu243 maintains a salt-bridge with Arg422. On the other hand, the binding of the common C5 substrates L-glutamate and alpha-ketoglutarate is enabled when Glu243 switches away and unshields Arg422. The structures reported here, sequence conservation and earlier mutational studies suggest that the "glutamate switch" is an elegant solution devised by a subgroup of fold type I aminotransferases for recognition of structurally diverse substrates in the same binding site and provides for reaction specificity.     

Crystal structure of a phenol-coupling P450 monooxygenase involved in teicoplanin biosynthesis

The lipoglycopeptide antibiotic teicoplanin has proven efficacy against gram‐positive pathogens. Teicoplanin is distinguished from the vancomycin‐type glycopeptide antibiotics, by the presence of an additional cross‐link between the aromatic amino acids 1 and 3 that is catalyzed by the cytochrome P450 monooxygenase Orf6* (CYP165D3). As a goal towards understanding the mechanism of this phenol‐coupling reaction, we have characterized recombinant Orf6* and determined its crystal structure to 2.2‐Å resolution. Although the structure of Orf6* reveals the core fold common to other P450 monooxygenases, there are subtle differences in the disposition of secondary structure elements near the active site cavity necessary to accommodate its complex heptapeptide substrate. Specifically, the orientation of the F and G helices in Orf6* results in a more closed active site than found in the vancomycin oxidative enzymes OxyB and OxyC. In addition, Met226 in the I helix replaces the more typical Gly/Ala residue that is positioned above the heme porphyrin ring, where it forms a hydrogen bond with a heme iron‐bound water molecule. Sequence comparisons with other phenol‐coupling P450 monooxygenases suggest that Met226 plays a role in determining the substrate regiospecificity of Orf6*. These features provide further insights into the mechanism of the cross‐linking mechanisms that occur during glycopeptide antibiotics biosynthesis. Proteins 2011; © 2011 Wiley‐Liss, Inc.     

Analysis of the class I aldolase binding site architecture based on the crystal structure of 2-deoxyribose-5-phosphate aldolase at 0.99A resolution

The crystal structure of the bacterial (Escherichia coli) class I 2-deoxyribose-5-phosphate aldolase (DERA) has been determined by Se-Met multiple anomalous dispersion (MAD) methods at 0.99A resolution. This structure represents the highest-resolution X-ray structure of an aldolase determined to date and enables a true atomic view of the enzyme. The crystal structure shows the ubiquitous TIM alpha/beta barrel fold. The enzyme contains two lysine residues in the active site. Lys167 forms the Schiff base intermediate, whereas Lys201, which is in close vicinity to the reactive lysine residue, is responsible for the perturbed pK(a) of Lys167 and, hence, also a key residue in the reaction mechanism. DERA is the only known aldolase that is able to use aldehydes as both aldol donor and acceptor molecules in the aldol reaction and is, therefore, of particular interest as a biocatalyst in synthetic organic chemistry. The uncomplexed DERA structure enables a detailed comparison with the substrate complexes and highlights a conformational change in the phosphate-binding site. Knowledge of the enzyme active-site environment has been the basis for exploration of catalysis of non-natural substrates and of mutagenesis of the phosphate-binding site to expand substrate specificity. Detailed comparison with other class I aldolase enzymes and DERA enzymes from different organisms reveals a similar geometric arrangement of key residues and implies a potential role for water as a general base in the catalytic mechanism.     

Crystal structure of muconate lactonizing enzyme at 3 A resolution

The crystal structure of muconate lactonizing enzyme has been solved at 3 A resolution, and an unambiguous alpha-carbon backbone chain trace made. The enzyme contains three domains; the central domain is a parallel-stranded alpha-beta barrel, which has previously been reported in six other enzymes, including triose phosphate isomerase and pyruvate kinase. One novel feature of this enzyme is that its alpha-beta barrel has only seven parallel alpha-helices around the central core of eight parallel beta-strands; all other known alpha-beta barrels contain eight such helices. The N-terminal (alpha + beta) and C-terminal domains cover the cleft where the eighth helix would be. The active site of muconate lactonizing enzyme has been found by locating the manganese ion that is essential for catalytic activity, and by binding and locating an inhibitor, alpha-ketoglutarate. The active site lies in a cleft between the N-terminal and barrel domains; when the active sites of muconate lactonizing enzyme and triose phosphate isomerase are superimposed, barrel-strand 1 of triose phosphate isomerase is aligned with barrel-strand 3 of muconate lactonizing enzyme. This implies that structurally homologous active-site residues in the two enzymes are carried on different parts of the primary sequence; the ancestral gene would had to have been transposed during its evolution to the modern proteins, which seems unlikely. Therefore, these two enzymes may be related by convergent, rather than divergent, evolution.     

One protein, two enzymes revisited: a structural entropy switch interconverts the two isoforms of acireductone dioxygenase

Acireductone dioxygenase (ARD) catalyzes different reactions between O2 and 1,2-dihydroxy-3-oxo-5-(methylthio)pent-1-ene (acireductone) depending upon the metal bound in the active site. Ni2+ -ARD cleaves acireductone to formate, CO and methylthiopropionate. If Fe2+ is bound (ARD'), the same substrates yield methylthioketobutyrate and formate. The two forms differ in structure, and are chromatographically separable. Paramagnetism of Fe2+ renders the active site of ARD' inaccessible to standard NMR methods. The structure of ARD' has been determined using Fe2+ binding parameters determined by X-ray absorption spectroscopy and NMR restraints from H98S ARD, a metal-free diamagnetic protein that is isostructural with ARD'. ARD' retains the beta-sandwich fold of ARD, but a structural entropy switch increases order at one end of a two-helix system that bisects the beta-sandwich and decreases order at the other upon interconversion of ARD and ARD', causing loss of the C-terminal helix in ARD' and rearrangements of residues involved in substrate orientation in the active site.     

Comparison of the crystal structures of L2 and L8S8 Rubisco suggests a functional role for the small subunit.

Comparison of the crystal structures of the L2 and L8S8 forms of ribulose-1,5-bisphosphate carboxylase from Rhodospirillum rubrum and spinach respectively, reveals a remarkable similarity in the overall architecture of the L2 building blocks in the two enzymes. Within the L subunits, no large conformational differences such as domain-domain rotations were found. In spite of a somewhat different packing of the L subunits in the L2 dimer, the active sites of the two enzymes are highly conserved. Significant local conformational differences are, however, observed for the C-terminal part of the polypeptide chains as well as for loop 7, helix alpha 7, loop 8 and helix alpha 8 in the barrel domain. The small subunit forms extensive interactions with one of these alpha helices, alpha 8, in the spinach L8S8 enzyme. The loops are at the active site and one of them forms a phosphate binding site for the substrate. We suggest that the small subunit modulates substrate binding and, possibly, the carboxylation/oxygenation ratio by inducing conformational changes in the active site through interactions distant from this site.     

Substrate-induced conformational changes in Escherichia coli taurine/alpha-ketoglutarate dioxygenase and insight into the oligomeric structure

The enzymes in the alpha-ketoglutarate (alphaKG) dependent dioxygenase superfamily represent the largest class of non-heme iron oxidases and have important medical, ecological, and biotechnological roles. One such enzyme, taurine/alpha-ketoglutarate dioxygenase (TauD), catalyzes the conversion of 2-aminoethanesulfonate (taurine) to sulfite and aminoacetaldehyde while decomposing alphaKG to succinate and CO(2). This alphaKG dependent dioxygenase is expressed in Escherichia coli under sulfur starvation conditions and allows the cell to utilize taurine, and other similar sulfonates in the environment, as an alternative sulfur source. In this work, we report the structures of the apo and holo forms of TauD to 1.9 A resolution (R(cryst) = 21.2%, R(free) = 24.9%) and 2.5 A resolution (R(cryst) = 22.5%, R(free) = 27.8%), respectively. The models reported herein provide significant new insight into the substrate orientations at the active site and the conformational changes that are induced upon taurine binding. Furthermore, analysis of our crystallographic data coupled with reanalysis of the crystallographic model (resolution = 3.0 A, R(cryst) = 28.1, R(free) = 32.0) presented by Elkins et al. (Biochemistry (2002) 41, 5185-5192) reveals an alternative oligomeric arrangement for the enzyme that is consistent with the conserved primary and secondary structure elements of other alphaKG dependent dioxygenases.     

Crystal structure of the petal death protein from carnation flower

Expression of the PSR132 protein from Dianthus caryophyllus (carnation, clover pink) is induced in response to ethylene production associated with petal senescence, and thus the protein is named petal death protein (PDP). Recent work has established that despite the annotation of PDP in sequence databases as carboxyphosphoenolpyruvate mutase, the enzyme is actually a C-C bond cleaving lyase exhibiting a broad substrate profile. The crystal structure of PDP has been determined at 2.7 A resolution, revealing a dimer-of-dimers oligomeric association. Consistent with sequence homology, the overall alpha/beta barrel fold of PDP is the same as that of other isocitrate lyase/PEP mutase superfamily members, including a swapped eighth helix within a dimer. Moreover, Mg(2+) binds in the active site of PDP with a coordination pattern similar to that seen in other superfamily members. A compound, covalently bound to the catalytic residue, Cys144, was interpreted as a thiohemiacetal adduct resulting from the reaction of glutaraldehyde used to cross-link the crystals. The Cys144-carrying flexible loop that gates access to the active site is in the closed conformation. Models of bound substrates and comparison with the closed conformation of isocitrate lyase and 2-methylisocitrate lyase revealed the structural basis for the broad substrate profile of PDP.     

Crystal structure of human heme oxygenase-1.

Heme oxygenase catalyzes the first step in the oxidative degradation of heme. The crystal structure of heme oxygenase-1 (HO-1) reported here reveals a novel helical fold with the heme sandwiched between two helices. The proximal helix provides a heme iron ligand, His 25. Conserved glycines in the distal helix near the oxygen binding site allow close contact between the helix backbone and heme in addition to providing flexibility for substrate binding and product release. Regioselective oxygenation of the alpha-meso heme carbon is due primarily to steric influence of the distal helix.     

Structure and reaction geometry of geranylgeranyl diphosphate synthase from Sinapis alba

The crystal structure of the geranylgeranyl diphosphate synthase from Sinapis alba (mustard) has been solved in two crystal forms at 1.8 and 2.0 A resolutions. In one of these forms, the dimeric enzyme binds one molecule of the final product geranylgeranyl diphosphate in one subunit. The chainfold of the enzyme corresponds to that of other members of the farnesyl diphosphate synthase family. Whereas the binding modes of the two substrates dimethylallyl diphosphate and isopentenyl diphosphate at the allyl and isopentenyl sites, respectively, have been established with other members of the family, the complex structure presented reveals for the first time the binding mode of a reaction product at the isopentenyl site. The binding geometry of substrates and product in conjunction with the protein environment and the established chemistry of the reaction provide a clear picture of the reaction steps and atom displacements. Moreover, a comparison with a ligated homologous structure outlined an appreciable induced fit: helix alpha8 and its environment undergo a large conformational change when either the substrate dimethylallyl diphosphate or an analogue is bound to the allyl site; only a minor conformational change occurs when the other substrate isopentenyl diphosphate or the product is bound to the isopentenyl site.     

Vmax regulation through domain and subunit changes. The active form of phosphoglycerate dehydrogenase

An active conformation of phosphoglycerate dehydrogenase (PGDH) from Escherichia coli has been obtained using X-ray crystallography. The X-ray crystal structure is used to examine the potential intermediates for V(max) regulation, for the redox reaction, and for cooperative effects of serine binding. The crystal structure at 2.2 A resolution contains bound NAD(+) cofactor, either sulfate or phosphate anions, and alpha-ketoglutarate, a nonphysiological substrate. A PGDH subunit is formed from three distinct domains: regulatory (RBD), substrate (SBD), and nucleotide binding (NBD). The crystal conformation of the homotetramer points to the fact that, in the absence of serine, coordinated movement of the RBD-SBD domains occurs relative to the NBD. The result is a conformational change involving the steric relationships of both the domains and the subunits. Within the active site of each subunit is a bound molecule of alpha-ketoglutarate and the coenzyme, NAD. The catalytic or active site cleft is changed slightly although it is still solvent exposed; therefore, the catalytic reaction probably involves additional conformational changes. By comparing the inhibited with the uninhibited complex, it is possible to describe changes in conformation that are involved in the inhibitory signal transduction of serine.     

Crystal structure and mechanism of tryptophan 2,3-dioxygenase, a heme enzyme involved in tryptophan catabolism and in quinolinate biosynthesis

The structure of tryptophan 2,3-dioxygenase (TDO) from Ralstonia metallidurans was determined at 2.4 A. TDO catalyzes the irreversible oxidation of l-tryptophan to N-formyl kynurenine, which is the initial step in tryptophan catabolism. TDO is a heme-containing enzyme and is highly specific for its substrate l-tryptophan. The structure is a tetramer with a heme cofactor bound at each active site. The monomeric fold, as well as the heme binding site, is similar to that of the large domain of indoleamine 2,3-dioxygenase, an enzyme that catalyzes the same reaction except with a broader substrate tolerance. Modeling of the putative (S)-tryptophan hydroperoxide intermediate into the active site, as well as substrate analogue and mutagenesis studies, are consistent with a Criegee mechanism for the reaction.     

Structure of the polyketide cyclase SnoaL reveals a novel mechanism for enzymatic aldol condensation

SnoaL belongs to a family of small polyketide cyclases, which catalyse ring closure steps in the biosynthesis of polyketide antibiotics produced in Streptomyces. Several of these antibiotics are among the most used anti-cancer drugs currently in use. The crystal structure of SnoaL, involved in nogalamycin biosynthesis, with a bound product, has been determined to 1.35 A resolution. The fold of the subunit can be described as a distorted alpha+beta barrel, and the ligand is bound in the hydrophobic interior of the barrel. The 3D structure and site-directed mutagenesis experiments reveal that the mechanism of the intramolecular aldol condensation catalysed by SnoaL is different from that of the classical aldolases, which employ covalent Schiff base formation or a metal ion cofactor. The invariant residue Asp121 acts as an acid/base catalyst during the reaction. Stabilisation of the enol(ate) intermediate is mainly achieved by the delocalisation of the electron pair over the extended pi system of the substrate. These polyketide cyclases thus form of family of enzymes with a unique catalytic strategy for aldol condensation.     

Crystal structure of the terminal oxygenase component of cumene dioxygenase from Pseudomonas fluorescens IP01

The crystal structure of the terminal component of the cumene dioxygenase multicomponent enzyme system of Pseudomonas fluorescens IP01 (CumDO) was determined at a resolution of 2.2 A by means of molecular replacement by using the crystal structure of the terminal oxygenase component of naphthalene dioxygenase from Pseudomonas sp. strain NCIB 9816-4 (NphDO). The ligation of the two catalytic centers of CumDO (i.e., the nonheme iron and Rieske [2Fe-2S] centers) and the bridging between them in neighboring catalytic subunits by hydrogen bonds through a single amino acid residue, Asp231, are similar to those of NphDO. An unidentified external ligand, possibly dioxygen, was bound at the active site nonheme iron. The entrance to the active site of CumDO is different from the entrance to the active site of NphDO, as the two loops forming the lid exhibit great deviation. On the basis of the complex structure of NphDO, a biphenyl substrate was modeled in the substrate-binding pocket of CumDO. The residues surrounding the modeled biphenyl molecule include residues that have already been shown to be important for its substrate specificity by a number of engineering studies of biphenyl dioxygenases.     

Crystal structure of Escherichia coli malate synthase G complexed with magnesium and glyoxylate at 2.0 A resolution: mechanistic implications

The crystal structure of selenomethionine-substituted malate synthase G, an 81 kDa monomeric enzyme from Escherichia coli has been determined by MAD phasing, model building, and crystallographic refinement to a resolution of 2.0 A. The crystallographic R factor is 0.177 for 49 242 reflections observed at the incident wavelength of 1.008 A, and the model stereochemistry is satisfactory. The basic fold of the enzyme is that of a beta8/alpha8 (TIM) barrel. The barrel is centrally located, with an N-terminal alpha-helical domain flanking one side. An inserted beta-sheet domain folds against the opposite side of the barrel, and an alpha-helical C-terminal domain forms a plug which caps the active site. Malate synthase catalyzes the condensation of glyoxylate and acetyl-coenzyme A and hydrolysis of the intermediate to yield malate and coenzyme A, requiring Mg(2+). The structure reveals an enzyme-substrate complex with glyoxylate and Mg(2+) which coordinates the aldehyde and carboxylate functions of the substrate. Two strictly conserved residues, Asp631 and Arg338, are proposed to provide concerted acid-base chemistry for the generation of the enol(ate) intermediate of acetyl-coenzyme A, while main-chain hydrogen bonds and bound Mg(2+) polarize glyoxylate in preparation for nucleophilic attack. The catalytic strategy of malate synthase appears to be essentially the same as that of citrate synthase, with the electrophile activated for nucleophilic attack by nearby positive charges and hydrogen bonds, while concerted acid-base catalysis accomplishes the abstraction of a proton from the methyl group of acetyl-coenzyme A. An active site aspartate is, however, the only common feature of these two enzymes, and the active sites of these enzymes are produced by quite different protein folds. Interesting similarities in the overall folds and modes of substrate recognition are discussed in comparisons of malate synthase with pyruvate kinase and pyruvate phosphate dikinase.