Emergence of pyridoxal phosphorylation through a promiscuous ancestor during the evolution of hydroxymethyl pyrimidine kinases

In the family of ATP-dependent vitamin kinases, several bifunctional enzymes that phosphorylate hydroxymethyl pyrimidine (HMP) and pyridoxal (PL) have been described besides enzymes specific towards HMP. To determine how bifunctionality emerged, we reconstructed the sequence of three ancestors of HMP kinases, experimentally resurrected, and assayed the enzymatic activity of their last common ancestor. The latter has ∼8-fold higher specificity for HMP due to a glutamine residue (Gln44) that is a key determinant of the specificity towards HMP, although it is capable of phosphorylating both substrates. These results show how a specific enzyme with catalytic promiscuity gave rise to current bifunctional enzymes.     

Crystal structure of 4-amino-5-hydroxymethyl-2-methylpyrimidine phosphate kinase from Salmonella typhimurium at 2.3 A resolution

The crystal structures of Salmonella typhimurium 4-amino-5-hydroxymethyl-2-methylpyrimidine phosphate kinase (HMPP kinase) and its complex with substrate HMP have been determined. HMPP kinase catalyzes two separate ATP-dependent phosphorylation reactions and is an essential enzyme in the thiamin biosynthetic pathway. HMPP kinase is a homodimer with one active site per monomer and is structurally homologous to members of the ribokinase family. A comparison of the structure of HMPP kinase with other members of the ribokinase family suggests an evolutionary progression. Modeling studies suggest that HMPP kinase catalyzes both of its phosphorylation reactions using in-line displacement mechanisms. We propose that the active site accommodates the two separate reactions by providing two different binding modes for the phosphate group of HMP phosphate.     

Crystal structure of the PdxY Protein from Escherichia coli

The crystal structure of Escherichia coli PdxY, the protein product of the pdxY gene, has been determined to a 2.2-A resolution. PdxY is a member of the ribokinase superfamily of enzymes and has sequence homology with pyridoxal kinases that phosphorylate pyridoxal at the C-5' hydroxyl. The protein is a homodimer with an active site on each monomer composed of residues that come exclusively from each respective subunit. The active site is filled with a density that fits that of pyridoxal. In monomer A, the ligand appears to be covalently attached to Cys122 as a thiohemiacetal, while in monomer B it is not covalently attached but appears to be partially present as pyridoxal 5'-phosphate. The presence of pyridoxal phosphate and pyridoxal as ligands was confirmed by the activation of aposerine hydroxymethyltransferase after release of the ligand by the denaturation of PdxY. The ligand, which appears to be covalently attached to Cys122, does not dissociate after denaturation of the protein. A detailed comparison (of functional properties, sequence homology, active site and ATP-binding-site residues, and active site flap types) of PdxY with other pyridoxal kinases as well as the ribokinase superfamily in general suggested that PdxY is a member of a new subclass of the ribokinase superfamily. The structure of PdxY also permitted an interpretation of work that was previously published about this enzyme.     

A new tyrosine phosphorylation mechanism involved in signal transduction in Bacillus subtilis

A new kind of prokaryotic protein tyrosine kinase was recently discovered, utilizing a guanidino-phosphotransferase domain for its kinase activity. Guanidino kinase domains originate from eukaryotic phosphagen kinases, a family of phosphoryl transfer enzymes with no homology to the serine/threonine and tyrosine kinase superfamily. Nevertheless, this kinase, McsB, exhibits the main structural and functional properties of prokaryotic tyrosine kinases. Tyrosine phosphorylation in bacteria is predominantly described to be involved in the regulation of exopolysaccharide synthesis and is therefore required for biofilm formation and virulence. McsB on the other hand modulates together with its activator protein, McsA, the activity of the repressor of the class III heat shock genes in B. subtilis. The analogy of the kinase mechanism of McsB to tyrosine kinases implicates that tyrosine kinases may harbor various and independently evolved domains for ATP-binding/hydrolysis and the transfer of the gamma-phosphate of ATP onto tyrosine residues.     

Structural evolution of the protein kinase-like superfamily

The protein kinase family is large and important, but it is only one family in a larger superfamily of homologous kinases that phosphorylate a variety of substrates and play important roles in all three superkingdoms of life. We used a carefully constructed structural alignment of selected kinases as the basis for a study of the structural evolution of the protein kinase-like superfamily. The comparison of structures revealed a "universal core" domain consisting only of regions required for ATP binding and the phosphotransfer reaction. Remarkably, even within the universal core some kinase structures display notable changes, while still retaining essential activity. Hence, the protein kinase-like superfamily has undergone substantial structural and sequence revision over long evolutionary timescales. We constructed a phylogenetic tree for the superfamily using a novel approach that allowed for the combination of sequence and structure information into a unified quantitative analysis. When considered against the backdrop of species distribution and other metrics, our tree provides a compelling scenario for the development of the various kinase families from a shared common ancestor. We propose that most of the so-called "atypical kinases" are not intermittently derived from protein kinases, but rather diverged early in evolution to form a distinct phyletic group. Within the atypical kinases, the aminoglycoside and choline kinase families appear to share the closest relationship. These two families in turn appear to be the most closely related to the protein kinase family. In addition, our analysis suggests that the actin-fragmin kinase, an atypical protein kinase, is more closely related to the phosphoinositide-3 kinase family than to the protein kinase family. The two most divergent families, alpha-kinases and phosphatidylinositol phosphate kinases (PIPKs), appear to have distinct evolutionary histories. While the PIPKs probably have an evolutionary relationship with the rest of the kinase superfamily, the relationship appears to be very distant (and perhaps indirect). Conversely, the alpha-kinases appear to be an exception to the scenario of early divergence for the atypical kinases: they apparently arose relatively recently in eukaryotes. We present possible scenarios for the derivation of the alpha-kinases from an extant kinase fold.     

Crystal structure of brain pyridoxal kinase,a novel member of the ribokinase superfamily

The three-dimensional structures of brain pyridoxal kinase and its complex with the nucleotide ATP have been elucidated in the dimeric form at 2.1 and 2.6 A, respectively. Results have shown that pyridoxal kinase, as an enzyme obeying random sequential kinetics in catalysis, does not possess a lid shape structure common to all kinases in the ribokinase superfamily. This finding has been shown to be in line with the condition that pyridoxal kinase binds substrates with variable sizes of chemical groups at position 4 of vitamin B(6) and its derivatives. In addition, the enzyme contains a 12-residue peptide loop in the active site for the prevention of premature hydrolysis of ATP. Conserved amino acid residues Asp(118) and Tyr(127) in the peptide loop could be moved to a position covering the nucleotide after its binding so that its chance to hydrolyze in the aqueous environment of the active site was reduced. With respect to the evolutionary trend of kinase enzymes, the existence of this loop in pyridoxal kinase could be classified as an independent category in the ribokinase superfamily according to the structural feature found and mechanism followed in catalysis.     

Crystal structure of pyridoxal kinase from the Escherichia coli pdxK gene: implications for the classification of pyridoxal kinases

The pdxK and pdxY genes have been found to code for pyridoxal kinases, enzymes involved in the pyridoxal phosphate salvage pathway. Two pyridoxal kinase structures have recently been published, including Escherichia coli pyridoxal kinase 2 (ePL kinase 2) and sheep pyridoxal kinase, products of the pdxY and pdxK genes, respectively. We now report the crystal structure of E. coli pyridoxal kinase 1 (ePL kinase 1), encoded by a pdxK gene, and an isoform of ePL kinase 2. The structures were determined in the unliganded and binary complexes with either MgATP or pyridoxal to 2.1-, 2.6-, and 3.2-A resolutions, respectively. The active site of ePL kinase 1 does not show significant conformational change upon binding of either pyridoxal or MgATP. Like sheep PL kinase, ePL kinase 1 exhibits a sequential random mechanism. Unlike sheep pyridoxal kinase, ePL kinase 1 may not tolerate wide variation in the size and chemical nature of the 4' substituent on the substrate. This is the result of differences in a key residue at position 59 on a loop (loop II) that partially forms the active site. Residue 59, which is His in ePL kinase 1, interacts with the formyl group at C-4' of pyridoxal and may also determine if residues from another loop (loop I) can fill the active site in the absence of the substrate. Both loop I and loop II are suggested to play significant roles in the functions of PL kinases.     

Structural Evolution of the Protein Kinase–Like Superfamily

The protein kinase family is large and important, but it is only one family in a larger superfamily of homologous kinases that phosphorylate a variety of substrates and play important roles in all three superkingdoms of life. We used a carefully constructed structural alignment of selected kinases as the basis for a study of the structural evolution of the protein kinase–like superfamily. The comparison of structures revealed a “universal core” domain consisting only of regions required for ATP binding and the phosphotransfer reaction. Remarkably, even within the universal core some kinase structures display notable changes, while still retaining essential activity. Hence, the protein kinase–like superfamily has undergone substantial structural and sequence revision over long evolutionary timescales. We constructed a phylogenetic tree for the superfamily using a novel approach that allowed for the combination of sequence and structure information into a unified quantitative analysis. When considered against the backdrop of species distribution and other metrics, our tree provides a compelling scenario for the development of the various kinase families from a shared common ancestor. We propose that most of the so-called “atypical kinases” are not intermittently derived from protein kinases, but rather diverged early in evolution to form a distinct phyletic group. Within the atypical kinases, the aminoglycoside and choline kinase families appear to share the closest relationship. These two families in turn appear to be the most closely related to the protein kinase family. In addition, our analysis suggests that the actin-fragmin kinase, an atypical protein kinase, is more closely related to the phosphoinositide-3 kinase family than to the protein kinase family. The two most divergent families, α-kinases and phosphatidylinositol phosphate kinases (PIPKs), appear to have distinct evolutionary histories. While the PIPKs probably have an evolutionary relationship with the rest of the kinase superfamily, the relationship appears to be very distant (and perhaps indirect). Conversely, the α-kinases appear to be an exception to the scenario of early divergence for the atypical kinases: they apparently arose relatively recently in eukaryotes. We present possible scenarios for the derivation of the α-kinases from an extant kinase fold.     

Structural and functional features of an NDP kinase from the hyperthermophile crenarchaeon Pyrobaculum aerophilum

Nucleoside diphosphate (NDP) kinases are ubiquitous enzymes that transfer gamma-phosphates from nucleoside triphosphates to nucleoside diphosphates via a ping-pong mechanism. The important role of this large family of enzymes in controlling cellular functions and developmental processes along with their crystallizability has made them good candidates for structural studies. We recently determined the structure of an evolved version of an NDP kinase from Pyrobaculum aerophilum, an extreme thermophile. This NDP kinase has similarity to the 42 other NDP kinases deposited in the Protein Data Bank (PDB) but differs significantly in sequence, structure, and biophysical properties. The P. aerophilum NDP kinase sequence contains two unique segments not present in other NDP kinases, comprising residues 66-100 and 156-165. We show that deletion mutants of the P. aerophilum NDP kinase lacking either or both of these inserts have an altered substrate specificity, allowing dGTP as the phosphate donor. A structural analysis of the evolved NDP kinase in conjunction with mutagenesis experiments suggests that the substrate specificity of the P. aerophilum NDP kinase is related to the presence of these two inserts.     

Structures of human N-Acetylglucosamine kinase in two complexes with N-Acetylglucosamine and with ADP/glucose: insights into substrate specificity and regulation

N-Acetylglucosamine (GlcNAc), a major component of complex carbohydrates, is synthesized de novo or salvaged from lysosomally degraded glycoconjugates and from nutritional sources. The salvage pathway requires that GlcNAc kinase converts GlcNAc to GlcNAc-6-phosphate, a component utilized in UDP-GlcNAc biosynthesis or energy metabolism. GlcNAc kinase belongs to the sugar kinase/Hsp70/actin superfamily that catalyze phosphoryl transfer from ATP to their respective substrates, and in most cases catalysis is associated with a large conformational change in which the N-terminal small and C-terminal large domains enclose the substrates. Here we report two crystal structures of homodimeric human GlcNAc kinase, one in complex with GlcNAc and the other in complex with ADP and glucose. The active site of GlcNAc kinase is located in a deep cleft between the two domains of the V-shaped monomer. The enzyme adopts a "closed" configuration in the GlcNAc-bound complex and GlcNAc interacts with residues of both domains. In addition, the N-acetyl methyl group contacts residues of the other monomer in the homodimer, a unique feature compared to other members of the sugar kinase/Hsp70/actin superfamily. This contrasts an "open" configuration in the ADP/glucose-bound structure, where glucose cannot form these interactions, explaining its low binding affinity for GlcNAc kinase. Our results support functional implications derived from apo crystal structures of GlcNAc kinases from Chromobacter violaceum and Porphyromonas gingivalis and show that Tyr205, which is phosphorylated in thrombin-activated platelets, lines the GlcNAc binding pocket. This suggests that phosphorylation of Tyr205 may modulate GlcNAc kinase activity and/or specificity.     

Crystal structure of human pyridoxal kinase

Pyridoxal kinase, a member of the ribokinase superfamily, catalyzes the ATP-dependent phosphorylation reaction of vitamin B6 and is an essential enzyme in the formation of pyridoxal-5'-phosphate, a key cofactor for over 100 enzymes. Pyridoxal kinase is thus regarded as a potential target for pharmacological agents. In this paper, we report the 2.8 angstroms crystal structure of human pyridoxal kinase (HPLK) expressed in Escherichia coli. The diffraction data revealed unexpected merohedral perfect twinning along the crystallographic c axis. Taking perfect twinning into account, the structure in dimeric form was well refined according to the CNS program. Structure comparison reveals that the key 12-residue peptide over the active site in HPLK is a beta-strand/loop/beta-strand flap, while the corresponding peptide in sheep brain enzyme adopts a loop conformation. Moreover, HPLK possesses a more hydrophobic ATP-binding pocket. This structure will facilitate further biochemical studies and structure-based design of drugs related to pyridoxal kinase.     

Protein kinase D as a potential new target for cancer therapy

Protein kinase D is a novel family of serine/threonine kinases and diacylglycerol receptors that belongs to the calcium/calmodulin-dependent kinase superfamily. Evidence has established that specific PKD isoforms are dysregulated in several cancer types, and PKD involvement has been documented in a variety of cellular processes important to cancer development, including cell growth, apoptosis, motility, and angiogenesis. In light of this, there has been a recent surge in the development of novel chemical inhibitors of PKD. This review focuses on the potential of PKD as a chemotherapeutic target in cancer treatment and highlights important recent advances in the development of PKD inhibitors.     

Expression, purification, and kinetic constants for human and Escherichia coli pyridoxal kinases

Pyridoxal kinase is an ATP dependent enzyme that phosphorylates pyridoxal, pyridoxine, and pyridoxamine forming their respective 5'-phosphorylated esters. The kinase is a part of the salvage pathway for re-utilizing pyridoxal 5'-phosphate, which serves as a coenzyme for dozens of enzymes involved in amino acid and sugar metabolism. Clones of two pyridoxal kinases from Escherichia coli and one from human were inserted into a pET 22b plasmid and expressed in E. coli. All three enzymes were purified to near homogeneity and kinetic constants were determined for the three vitamin substrates. Previous studies had suggested that ZnATP was the preferred trinucleotide substrate, but our studies show that under physiological conditions MgATP is the preferred substrate. One of the two E. coli kinases has very low activity for pyridoxal, pyridoxine, and pyridoxamine. We conclude that in vivo this kinase may have an alternate substrate involved in another metabolic pathway and that pyridoxal has only a poor secondary activity for this kinase.     

Identification of a new subfamily of HNH nucleases and experimental characterization of a representative member, HphI restriction endonuclease

The restriction endonuclease (REase) R. HphI is a Type IIS enzyme that recognizes the asymmetric target DNA sequence 5'-GGTGA-3' and in the presence of Mg(2+) hydrolyzes phosphodiester bonds in both strands of the DNA at a distance of 8 nucleotides towards the 3' side of the target, producing a 1 nucleotide 3'-staggered cut in an unspecified sequence at this position. REases are typically ORFans that exhibit little similarity to each other and to any proteins in the database. However, bioinformatics analyses revealed that R.HphI is a member of a relatively big sequence family with a conserved C-terminal domain and a variable N-terminal domain. We predict that the C-terminal domains of proteins from this family correspond to the nuclease domain of the HNH superfamily rather than to the most common PD-(D/E)XK superfamily of nucleases. We constructed a three-dimensional model of the R.HphI catalytic domain and validated our predictions by site-directed mutagenesis and studies of DNA-binding and catalytic activities of the mutant proteins. We also analyzed the genomic neighborhood of R.HphI homologs and found that putative nucleases accompanied by a DNA methyltransferase (i.e. predicted REases) do not form a single group on a phylogenetic tree, but are dispersed among free-standing putative nucleases. This suggests that nucleases from the HNH superfamily were independently recruited to become REases in the context of RM systems multiple times in the evolution and that members of the HNH superfamily may be much more frequent among the so far unassigned REase sequences than previously thought.     

Identification of an Allosteric Signaling Network within Tec Family Kinases

The Tec family kinases are tyrosine kinases that function primarily in hematopoietic cells. The catalytic activity of the Tec kinases is positively influenced by the regulatory domains outside of the kinase domain. The current lack of a full-length Tec kinase structure leaves a void in our understanding of how these positive regulatory signals are transmitted to the kinase domain. Recently, a conserved structure within kinases, the ‘regulatory spine’, which assembles and disassembles as a kinase switches between its active and inactive states, has been identified. Here, we define the residues that comprise the regulatory spine within Tec kinases. Compared to previously characterized systems, the Tec kinases contain an extended regulatory spine that includes a conserved methionine within the C-helix and a conserved tryptophan within the Src homology 2-kinase linker of Tec kinases. This extended regulatory spine forms a conduit for transmitting the presence of the regulatory domains of Tec kinases to the catalytic domain. We further show that mutation of the gatekeeper residue at the edge of the regulatory spine stabilizes the regulatory spine, resulting in a constitutively active kinase domain. Importantly, the regulatory spine is preassembled in this gatekeeper mutant, rendering phosphorylation on the activation loop unnecessary for its activity. Moreover, we show that the disruption of the conserved electrostatic interaction between Bruton's tyrosine kinase R544 on the activation loop and Bruton's tyrosine kinase E445 on the C-helix also aids in the assembly of the regulatory spine. Thus, the extended regulatory spine is a key structure that is critical for maintaining the activity of Tec kinases.     

Crystal structure of a C-terminal deletion mutant of human protein kinase CK2 catalytic subunit

Protein kinase CK2 (formerly called: casein kinase 2) is a heterotetrameric enzyme composed of two separate catalytic chains (CK2alpha) and a stable dimer of two non-catalytic subunits (CK2beta). CK2alpha is a highly conserved member of the superfamily of eukaryotic protein kinases. The crystal structure of a C-terminal deletion mutant of human CK2alpha was solved and refined to 2.5A resolution. In the crystal the CK2alpha mutant exists as a monomer in agreement with the organization of the subunits in the CK2 holoenzyme. The refined structure shows the helix alphaC and the activation segment, two main regions of conformational plasticity and regulatory importance in eukaryotic protein kinases, in active conformations stabilized by extensive contacts to the N-terminal segment. This arrangement is in accordance with the constitutive activity of the enzyme. By structural superimposition of human CK2alpha in isolated form and embedded in the human CK2 holoenzyme the loop connecting the strands beta4 and beta5 and the ATP-binding loop were identified as elements of structural variability. This structural comparison suggests that the ATP-binding loop may be the key region by which the non-catalytic CK2beta dimer modulates the activity of CK2alpha. The beta4/beta5 loop was found in a closed conformation in contrast to the open conformation observed for the CK2alpha subunits of the CK2 holoenzyme. CK2alpha monomers with this closed beta4/beta5 loop conformation are unable to bind CK2beta dimers in the common way for sterical reasons, suggesting a mechanism to protect CK2alpha from integration into CK2 holoenzyme complexes. This observation is consistent with the growing evidence that CK2alpha monomers and CK2beta dimers can exist in vivo independently from the CK2 holoenzyme and may possess physiological roles of their own.     

Dimer–Dimer Interaction of the Bacterial Selenocysteine Synthase SelA Promotes Functional Active-Site Formation and Catalytic Specificity

The 21st amino acid, selenocysteine (Sec), is incorporated translationally into proteins and is synthesized on its specific tRNA (tRNA^Sec). In Bacteria, the selenocysteine synthase SelA converts Ser-tRNA^Sec, formed by seryl-tRNA synthetase, to Sec-tRNA^Sec. SelA, a member of the fold-type-I pyridoxal 5′-phosphate-dependent enzyme superfamily, has an exceptional homodecameric quaternary structure with a molecular mass of about 500 kDa. Our previously determined crystal structures of Aquifex aeolicus SelA complexed with tRNA^Sec revealed that the ring-shaped decamer is composed of pentamerized SelA dimers, with two SelA dimers arranged to collaboratively interact with one Ser-tRNA^Sec. The SelA catalytic site is close to the dimer–dimer interface, but the significance of the dimer pentamerization in the catalytic site formation remained elusive. In the present study, we examined the quaternary interactions and demonstrated their importance for SelA activity by systematic mutagenesis. Furthermore, we determined the crystal structures of “depentamerized” SelA variants with mutations at the dimer–dimer interface that prevent pentamerization. These dimeric SelA variants formed a distorted and inactivated catalytic site and confirmed that the pentamer interactions are essential for productive catalytic site formation. Intriguingly, the conformation of the non-functional active site of dimeric SelA shares structural features with other fold-type-I pyridoxal 5′-phosphate-dependent enzymes with native dimer or tetramer (dimer-of-dimers) quaternary structures.     

Active site structure and stereospecificity of Escherichia coli pyridoxine-5'-phosphate oxidase.

Pyridoxine-5'-phosphate oxidase catalyzes the oxidation of either the C4' alcohol group or amino group of the two substrates pyridoxine 5'-phosphate and pyridoxamine 5'-phosphate to an aldehyde, forming pyridoxal 5'-phosphate. A hydrogen atom is removed from C4' during the oxidation and a pair of electrons is transferred to tightly bound FMN. A new crystal form of the enzyme in complex with pyridoxal 5'-phosphate shows that the N-terminal segment of the protein folds over the active site to sequester the ligand from solvent during the catalytic cycle. Using (4'R)-[(3)H]PMP as substrate, nearly 100 % of the radiolabel appears in water after oxidation to pyridoxal 5'-phosphate. Thus, the enzyme is specific for removal of the proR hydrogen atom from the prochiral C4' carbon atom of pyridoxamine 5'-phosphate. Site mutants were made of all residues at the active site that interact with the oxygen atom or amine group on C4' of the substrates. Other residues that make interactions with the phosphate moiety of the substrate were mutated. The mutants showed a decrease in affinity, but exhibited considerable catalytic activity, showing that these residues are important for binding, but play a lesser role in catalysis. The exception is Arg197, which is important for both binding and catalysis. The R197 M mutant enzyme catalyzed removal of the proS hydrogen atom from (4'R)-[(3)H]PMP, showing that the guanidinium side-chain plays an important role in determining stereospecificity. The crystal structure and the stereospecificity studies suggests that the pair of electrons on C4' of the substrate are transferred to FMN as a hydride ion.     

An expanded adenylate kinase gene family in the protozoan parasite Trypanosoma cruzi

Adenylate kinases supply energy routes in cellular energetic homeostasis. In this work, we identified and characterized the adenylate kinase activity in extracts from the flagellated parasite Trypanosoma cruzi, the causative agent of Chagas' disease. Adenylate kinase activity was detected in different subcellular fractions and the cytosolic isoform was biochemically characterized. Cytosolic adenylate kinase specific activity increases continuously during the epimastigote growth and is down-regulated when other soluble phosphotransferase, arginine kinase, is overexpressed. Six different genes of adenylate kinase isoforms were identified and the mRNA expression was confirmed by RT-PCR and Northern Blot. Three open reading frames coding for different enzyme isoforms named TzADK1, TzADK2 and TzADK5 were cloned and functionally expressed in E. coli. This work reports an unusually large number of genes of adenylate kinases and suggests a coordinated regulation of phosphotransferase-mediated ATP regenerating pathways in the unicellular parasite Trypanosoma cruzi.     

X-ray structure of Escherichia coli pyridoxine 5'-phosphate oxidase complexed with pyridoxal 5'-phosphate at 2.0 A resolution.

Escherichia coli pyridoxine 5'-phosphate oxidase catalyzes the terminal step in the biosynthesis of pyridoxal 5'-phosphate by the FMN oxidation of pyridoxine 5'-phosphate forming FMNH(2) and H(2)O(2). Recent studies have shown that in addition to the active site, pyridoxine 5'-phosphate oxidase contains a non-catalytic site that binds pyridoxal 5'-phosphate tightly. The crystal structure of pyridoxine 5'-phosphate oxidase from E. coli with one or two molecules of pyridoxal 5'-phosphate bound to each monomer has been determined to 2.0 A resolution. One of the pyridoxal 5'-phosphate molecules is clearly bound at the active site with the aldehyde at C4' of pyridoxal 5'-phosphate near N5 of the bound FMN. A protein conformational change has occurred that partially closes the active site. The orientation of the bound pyridoxal 5'-phosphate suggests that the enzyme catalyzes a hydride ion transfer between C4' of pyridoxal 5'-phosphate and N5 of FMN. When the crystals are soaked with excess pyridoxal 5'-phosphate an additional molecule of this cofactor is also bound about 11 A from the active site. A possible tunnel exists between the two sites so that pyridoxal 5'-phosphate formed at the active site may transfer to the non-catalytic site without passing though the solvent.