Feedback regulation mechanisms for the control of GTP cyclohydrolase I activity.

Guanosine triphosphate (GTP) cyclohydrolase I, the rate-limiting enzyme in the biosynthesis of tetrahydrobiopterin (BH4), is subject to feedback inhibition by BH4, a cofactor for phenylalanine hydroxylase. Inhibition was found to depend specifically on BH4 and the presence of another protein (p35). The inhibition occurred through BH4-dependent complex formation between p35 protein and GTP cyclohydrolase I. Furthermore, the inhibition was specifically reversed by phenylalanine, and, in conjunction with p35, phenylalanine reduced the cooperativity of GTP cyclohydrolase I. These findings also provide a molecular basis for high plasma BH4 concentrations observed in patients with hyperphenylalaninemia caused by phenylalanine hydroxylase deficiency.

An ICAM-related neuronal glycoprotein, telencephalin, with brain segment-specific expression.

Telencephalin (TLN) is a 130 kd glycoprotein expressed exclusively in neurons of the telencephalon, the most rostral brain segment. In the neurons, TLN is localized to soma-dendritic membrane but not to axonal membrane. In this study, we have cloned cDNA encoding rabbit and mouse TLN. The cDNA-derived primary structure of TLN predicts an integral membrane protein with nine tandem immunoglobulin-like domains in an extra-cellular region, a transmembrane domain, and a short cytoplasmic tail. The distal eight immunoglobulin-like domains of TLN show highest homology with the immunoglobulin-like domains of intracellular adhesion molecules (ICAMs) 1, 2, and 3/R. The structural similarity of TLN with ICAMs provides a new and strong link between immunoglobulin superfamily molecules in the nervous and immune systems. TLN is an example of a dendrite-associated cell adhesion molecule involved in the brain's segmental organization, cell-cell interactions during dendritic development, and maintenance of functional neuronal networks.

BIG-1: a new TAG-1/F3-related member of the immunoglobulin superfamily with neurite outgrowth-promoting activity.

We have cloned a rat cDNA for a novel brain-derived immunoglobulin (Ig) superfamily molecule, BIG-1, by using PCR based on the amino acid sequences of the two closely related and well-known Ig superfamily members, rat TAG-1 and mouse F3. BIG-1 is a glycosylphosphatidylinositol-anchored membrane protein with six Ig-like domains and four fibronectin type III repeats, belonging to the TAG-1/F3 subgroup. The expression of BIG-1 mRNA is developmentally regulated with the highest level in the adult brain. It is restricted to subsets of neurons such as Purkinje cells of the cerebellum, granule cells of the dentate gyrus, and neurons in the superficial layers of the cerebral cortex. Recombinant BIG-1 protein has a neurite outgrowth-promoting activity when used as a substrate for neurons in vitro. These results suggest that BIG-1 may be involved in the formation and maintenance of neuron type-specific networks in the brain.

A genetic approach to visualization of multisynaptic neural pathways using plant lectin transgene.

The wiring patterns among various types of neurons via specific synaptic connections are the basis of functional logic employed by the brain for information processing. This study introduces a powerful method of analyzing the neuronal connectivity patterns by delivering a tracer selectively to specific types of neurons while simultaneously transsynaptically labeling their target neurons. We developed a novel genetic approach introducing cDNA for a plant lectin, wheat germ agglutinin (WGA), as a transgene under the control of specific promoter elements. Using this method, we demonstrate three examples of visualization of specific transsynaptic neural pathways: the mouse cerebellar efferent pathways, the mouse olfactory pathways, and the Drosophila visual pathways. This strategy should greatly facilitate studies on the anatomical and functional organization of the developing and mature nervous system.

Cloning and characterization of the tyrB gene from Salmonella typhimurium.

We found a gene homologous to tyrB, which encodes aromatic amino acid aminotransferase (ArAT, EC2.6.1.57) in Escherichia coli, in the genome of Salmonella typhimurium IFO 13245. The S. typhimurium tyrB product consists of 397 amino acid residues. The amino acid sequence shows 87.9% identity with that of E. coli ArAT, but shows lower identity (42.3%) with that of E. coli aspartate aminotransferase (AspAT, EC2.6.1.1). When the S. typhimurium tyrB gene was expressed in an E. coli mutant whose intrinsic tyrB gene had been inactivated, the activity of transaminating tyrosine and phenylalanine could be recovered, indicating that the S. typhimurium tyrB gene product possesses transamination activities similar to those of the E. coli ArAT. Elucidation of the molecular features of a new ArAT may be helpful for structural and functional analyses of ArAT and AspAT with regard to the different but overlapping substrate specificity of the two enzymes.

Crystal structures of Paracoccus denitrificans aromatic amino acid aminotransferase: a substrate recognition site constructed by rearrangement of hydrogen bond network.

Aminotransferase reversibly catalyzes the transamination reaction by a ping-pong bi-bi mechanism with pyridoxal 5'-phosphate (PLP) as a cofactor. Various kinds of aminotransferases developing into catalysts for particular substrates have been reported. Among the aminotransferases, aromatic amino acid aminotransferase (EC 2.6.1. 57) catalyzes the transamination reaction with both acidic substrates and aromatic substrates. To elucidate the multiple substrate recognition mechanism, we determined the crystal structures of aromatic amino acid aminotransferase from Paracoccus denitrificans (pdAroAT): unliganded pdAroAT, pdAroAT in a complex with maleate as an acidic substrate analog, and pdAroAT in a complex with 3-phenylpropionate as an aromatic substrate analog at 2.33 A, 2. 50 A and 2.30 A resolution, respectively. The pdAroAT molecule is a homo-dimer. Each subunit has 394 amino acids and one PLP and is divided into small and large domains. The overall structure of pdAroAT is essentially identical to that of aspartate aminotransferase (AspAT) which catalyzes the transamination reaction with only an acidic amino acid. On binding the acidic substrate analog, arginine 292 and 386 form end-on salt bridges with carboxylates of the analog. Furthermore, binding of the substrate induces the domain movement to close the active site. The recognition mechanism for the acidic substrate analog in pdAroAT is identical to that observed in AspAT. Binding of the aromatic substrate analog causes reorientation of the side-chain of the residues, lysine 16, asparagine 142, arginine 292* and serine 296*, and changes in the position of water molecules in the active site to form a new hydrogen bond network in contrast to the active site structure of pdAroAT in the complex with an acidic substrate analog. Consequently, the rearrangement of the hydrogen bond network can form recognition sites for both acidic and aromatic side-chains of the substrate without a conformational change in the backbone structure in pdAroAT.

Role of an active site residue analyzed by combination of mutagenesis and coenzyme analog.

Asp222 of aspartate aminotransferase is an active-site residue which interacts with the pyridine nitrogen of the coenzyme, pyridoxal 5'-phosphate (PLP). The roles of Asp222 in the catalytic mechanism of Escherichia coli aspartate aminotransferase have previously been explored by site-directed mutagenesis. These studies confirmed that a negatively charged residue at position 222 is essential for catalysis, but the reason for this remained speculative. In the present studies, the roles of Asp222 were clarified experimentally by analyzing the mutant D222A enzyme (Asp222 replaced by Ala) reconstituted with the coenzyme analog N(1)-methylated PLP (N-MePLP). Spectroscopic and kinetic analyses showed that Asp222 stabilizes the protonated N(1) of PLP, raising the pKa value of N(1) by more than five units, in the active site of AspAT. The positive charge at N(1) accelerates abstraction of the alpha-proton from the amino acid substrate, stabilizing the transition state by 1.4 to 4.5 kcal.mol-1 in the reaction with aspartate. X-ray crystallographic (2.0 A resolution) and CD spectroscopic studies suggest that the coenzyme analog is not held in a proper orientation within the active site of D222A (N-MePLP). This may account for the finding that the catalytic activity was recovered only partially by the reconstitution of D222A with N-MePLP. These results fully support the following postulated role of Asp222: the negative charge of Asp222 stabilizes the positive charge at N(1) of PLP and thereby enhances the function of PLP as an electron sink.

Aromatic L-amino acid decarboxylase: conformational change in the flexible region around Arg334 is required during the transaldimination process.

Aromatic L-amino acid decarboxylase (AADC) catalytic mechanism has been proposed to proceed through two consecutive intermediates (i.e., Michaelis complex and the external aldimine). Limited proteolysis of AADC that preferentially digested at the C-terminal side of Arg334 was slightly retarded in the presence of dihydroxyphenyl acetate that formed a stable Michaelis complex. On the contrary, AADC was scarcely digested in the presence of L-dopa methyl ester that formed a stable external aldimine. Similar protection by the substrate analogs was observed in the chemical modification experiment. From these results, we concluded that the region around Arg334 must be exposed and flexible in the unliganded state, and forming the Michaelis complex generated a subtle conformational change, then underwent marked conformational change during the subsequent transaldimination process prerequisite to forming the external aldimine. For further analyses, we constructed a mutant gene encoding in tandem the two peptides of AADC cleaved at the Asn327-Met328 bond inside the putative flexible region. The gene product, fragmentary AADC, was still active with L-dopa as substrate, but its k(cat) value was decreased 57-fold, and the Km value was increased 9-fold compared with those of the wild-type AADC. The absorption spectra of the fragmentary AADC in the presence of L-dopa methyl ester showed shift in the equilibrium of the transaldimination from the external aldimine to the Michaelis complex. Tryptic digestion of the fragmentary AADC removed seven amino acid residues, Met328-Arg334, and resulted in complete inactivation. Susceptibility of the fragmentary enzyme to trypsin was not changed by L-dopa methyl ester revealing the loss of appropriate conformational change in the flexible region induced by substrate binding. From these results we propose that the conformational change in the flexible region is required during the transaldimination process.

A large compressibility change of protein induced by a single amino acid substitution.

The adiabatic compressibility (beta s) was determined, by means of the precise sound velocity and density measurements, for a series of single amino acid substituted mutant enzymes of Escherichia coli dihydrofolate reductase (DHFR) and aspartate aminotransferase (AspAT). Interestingly, the beta s values of both DHFR and AspAT were influenced markedly by the mutations at glycine-121 and valine-39, respectively, in which the magnitude of the change was proportional to the enzyme activity. This result demonstrates that the local change of the primary structure plays an important role in atomic packing and protein dynamics, which leads to the modified stability and enzymatic function. This is the first report on the compressibility of mutant proteins.

The nucleotide sequence surrounding the promoter region of colicin E1 gene.

The nucleotide sequence of 570 bp, covering the N-terminal portion of the colicin E1 gene, was determined. The sequence of the N-terminal four amino acids of the colicin E1 protein, determined by manual Edman degradation, agreed with that predicted from the nucleotide sequence. From analysis of the 5'-terminal sequences of RNAs synthesized in vitro, the promoter and operator regions of the colicin E1 gene were assigned. These data indicate the existence of two promoters, one of which is located in the coding region for colicin E1. DNA sequence homology of 16 bp was found between the putative operator regions of the colicin E1 and recA genes.

Disruption of Thermus thermophilus genes by homologous recombination using a thermostable kanamycin-resistant marker.

Genes of an extremely thermophilic bacterium, Thermus thermophilus, were disrupted by homologous recombination using a recently developed, thermostable kanamycin-resistant marker. First, the trpE gene was disrupted with various constructions of DNA. The transformation efficiency was exponentially increased as the length of the homologous regions flanking the marker gene increased above the minimum length (200-300 bp). We then disrupted five genes of the nucleotide excision repair system and examined their phenotypes. The convenience and high reliability of this method should prompt its application to the high-throughput systematic disruption of the genes of this thermophilic bacterium.

cDNA cloning and characterization of mouse nifS-like protein, m-Nfs1: mitochondrial localization of eukaryotic NifS-like proteins.

We have isolated a mouse cDNA which shows significant sequence similarity to the yeast nifS-like gene (y-NFS1), and termed it m-Nfs1. The deduced protein sequence (459 amino acids long) has several characteristic features common to those of bacterial NifS proteins, but distinct from them by its amino-terminal extension which contains a typical mitochondrial targeting presequence. m-Nfs1 was found to be a soluble 47-kDa protein in the matrix fraction of mouse liver mitochondria. The m-Nfs1 gene was ubiquitously expressed in most tissues, suggesting its housekeeping function in vivo. We also found that the gamma-NFS1 protein was localized in the mitochondrial matrix in yeast cells. These results suggest that both eukaryotic NifS-like proteins may play some roles in mitochondrial functions.

Activation of protein phosphatase 2A by cAMP-dependent protein kinase-catalyzed phosphorylation of the 74-kDa B'' (delta) regulatory subunit in vitro and identification of the phosphorylation sites.

Human erythrocyte protein phosphatase 2A, which comprises a 34-kDa catalytic C subunit, a 63-kDa regulatory A subunit and a 74-kDa regulatory B'' (delta) subunit, was phosphorylated at serine residues of B'' in vitro by cAMP-dependent protein kinase (A-kinase). In the presence and absence of 0.5 microM okadaic acid (OA), A-kinase gave maximal incorporation of 1.7 and 1.0 mol of phosphate per mol of B'', respectively. The Km value of A-kinase for CAB'' was 0.17 +/- 0.01 microM in the presence of OA. The major in vitro phosphorylation sites of B'' were identified as Ser-60, -75 and -573 in the presence of OA, and Ser-75 and -573 in the absence of OA. Phosphorylation of B'' did not dissociate B'' from CA, and stimulated the molecular activity of CAB'' toward phosphorylated H1 and H2B histones, 3.8- and 1.4-fold, respectively, but not toward phosphorylase a.

Molecular cloning of a 74-kDa regulatory subunit (B" or delta) of human protein phosphatase 2A.

Based on amino acid sequence data of a 74-kDa regulatory subunit (B" or delta) of a human heterotrimeric protein phosphatase 2A, a cDNA encoding the subunit was isolated from a human cerebral cortex library. The cDNA had an open reading frame encoding an M(r) 66,138 protein of 570 amino acids. Bacterial expression of the cDNA yielded a protein immunoreactive with antisera specific to the 74-kDa subunit. The predicted primary structure of the subunit had no similarity to already reported sequences of PP2A regulatory subunits including A, B, and PR72. Potential phosphorylation sites for protein kinases A and C, a bipartite motif of putative nuclear localization signal, and SH3 accessible proline-rich domain, and a unique PQ repeat were found in the sequence. The subunit mRNA of about 2.9 kb was ubiquitously expressed in rat tissues.

Replacement of an interdomain residue Val39 of Escherichia coli aspartate aminotransferase affects the catalytic competence without altering the substrate specificity of the enzyme.

Three mutant Escherichia coli aspartate aminotransferases in which Val39 was changed to Ala, Leu, and Phe by site-directed mutagenesis were prepared and characterized. Among the three mutant and the wild-type enzymes, the Leu39 enzyme had the lowest Km values for dicarboxylic substrates. The Km values of the Ala39 enzyme for dicarboxylates were essentially the same as those of the wild-type (Val39) enzyme. These two mutant enzymes showed essentially the same kcat values for dicarboxylic substrates as did the wild-type enzyme. On the other hand, incorporation of a bulky side-chain at position 39 (Phe39 enzyme) decreased both the affinity (1/Km) and catalytic ability (kcat) toward dicarboxylic substrates. These results show that the position 39 residue is involved in the modulation of both the binding of dicarboxylic substrates to enzyme and the catalytic ability of the enzyme. Although the replacement of Val39 with other residues altered both the kcat and Km values toward various substrates including dicarboxylic and aromatic amino acids and the corresponding oxo acids, it did not alter the ratio of the kcat/Km value of the enzyme toward a dicarboxylic substrate to that for an aromatic substrate. The affinity for aromatic substrates was not affected by changing the residue at position 39. These data indicate that, although the side chain bulkiness of the residue at position 39 correlates well with the activity toward aromatic substrates in the sequence alignment of several aminotransferases [Seville, M., Vincent M.G., & Hahn, K. (1988) Biochemistry 27, 8344-8349], the residue does not seem to be involved in the recognition of aromatic substrates.

Protonation state of the active-site Schiff base of aromatic amino acid aminotransferase: modulation by binding of ligands and implications for its role in catalysis.

The Schiff base formed between Lys258 of Escherichia coli aromatic amino acid aminotransferase (ArAT) and the coenzyme pyridoxal 5'-phosphate (PLP) has a pKa value of 6.65. The pH dependency of the kinetic parameters was consistent with a mechanism by which the enzymatic form with the nonprotonated Schiff base productively binds aspartate, phenylalanine, and tryptophan. The Schiff base pKa value rose by 1.7-2.1 unit on binding of substrate analogs, and this strongly suggested protonation of the Schiff base upon formation of the Michaelis complex with substrates. The protonated "internal" Schiff base in the Michaelis complex is supposed to be attacked by the deprotonated substrate amino group, and this explains excellently the mechanism of transaldimination to form the PLP-substrate Schiff base. Phenylpropionate and indolepropionate caused similar increases in the pKa value to maleate. [Arg292-->Ala] ArAT showed the same pKa value as the wild-type enzyme. Therefore, neutralization of Arg292 by omega-carboxylate of dicarboxylic ligands, which had been well documented in aspartate aminotransferase to increase the Schiff base pKa, has little effect on the protonation of the Schiff base in ArAT. Thus the structure of ArAT is deliberately organized so that the Schiff base pKa is effectively modulated by substrates having only one carboxylate group.

The complete amino acid sequence of the beta-subunit of protocatechuate 3,4-dioxygenase from Pseudomonas aeruginosa.

The complete amono aicd sequence of the beta-subunit of protocatechuate 3,4-dioxygenase is presented. The beta-subunit contained 237 amino acid residues, 4 of which were methionines. Accordingly, cyanogen bromide cleavage of the S-carboxymethylated beta-subunit produced five peptides. The sequences of these peptides were determined by analyses of the peptides obtained by tryptic, staphyloccal protease and thermolysin digestions. The alignment of the cyanogen bromide peptides was deduced by the use of overlapping peptides containing methionine which were obtained by tryptic digestion of the S-carboxymethylated beta-subunit. The calculated molecular weight was 26,588, which is close to the value estimated by acrylamide gel electrophoresis in the presence of sodium dodecyl sulfate.

Aspartate: 2-oxoglutarate aminotransferase from bakers' yeast: crystallization and characterization.

Aspartate: 2-oxoglutarate aminotransferase [EC 2.6.1.1] was purified and crystallized from bakers' yeast. The crystalline preparation gave a single band on polyacrylamide disc gel electrophoresis in the presence of sodium dodecyl sulfate. However, in the absence of sodium dodecyl sulfate, the preparation gave one major band with two faint bands, all of which showed the same specific activity, molecular weight and serological properties. These faint bands appeared to be modified forms produced from the major band during the purification. The enzyme showed a molecular weight of 90,000 +/- 8,000 and 92,000 +/- 8,000 by gel filtration and sedimentation equilibrium analysis, respectively. The molecular weight of a subunit was estimated to be 45,000 by sodium dodecyl sulfate slab gel electrophoresis. Each subunit bound approximately 1 mol of pyridoxal 5'-phosphate. The bound pyridoxal 5'-phosphate showed an absorption maximum at 360 nm (epsilon M: 11,500) and 430 nm (epsilon M: 8,200) in alkaline and acidic conditions, respectively. Its proteolytic pK was pH 6.3. The enzyme showed an optimum pH of 8.0-9.0, and fairly high amino donor and acceptor specificities; aromatic amino acids and their corresponding 2-oxoacids were catalyzed at rates of 0.2-0.8% of those for aspartate and oxalacetate, respectively. Michaelis constants for various substrate were: L-aspartate (0.11 mM), L-glutamate (20.0 mM), oxalacetate (0.006 mM), and 2-oxoglutarate (0.16 mM). The antiserum against yeast aspartate aminotransferase did not form precipitin bands with homogeneous aspartate aminotransferases from pig heart cytosol, pig heart mitochondria or Escherichia coli B.

Branched-chain amino acid aminotransferase of Escherichia coli: nucleotide sequence of the ilvE gene and the deduced amino acid sequence.

The ilvE gene of the Escherichia coli K-12 ilvGEDA operon, which encodes branched-chain amino acid aminotransferase [EC 2.6.1.42], was cloned. The nucleotide sequence of 1.5 kilobase pairs containing the gene was determined. The coding region of the ilvE gene contained 927 nucleotide residues and could encode 309 amino acid residues. The predicted molecular weight, amino acid composition and the sequence of the N-terminal 15 residues agreed with the enzyme data reported previously (Lee-Peng, F.-C., et al. (1979) J. Bacteriol. 139, 339-345). From the deduced amino acid sequence, the secondary structure was predicted.

An anomalous side reaction of the Lys303 mutant aromatic L-amino acid decarboxylase unravels the role of the residue in catalysis.

Aromatic L-amino acid decarboxylase (AADC) in which the pyridoxal 5'-phosphate (PLP)-binding residue Lys303 was replaced by an alanine residue is virtually inactive as a catalyst. On reaction of the normal substrate L-dopa with this mutant AADC, the absorption at around 330 nm gradually increased with concomitant decrease of the absorption of the free PLP molecule at 390 nm. Analysis of the 330-nm absorbing species on HPLC and spectrophotometry showed that it is a 1:1 adduct of PLP and dopamine, probably the Pictet-Spengler type adduct formed from the PLP-dopamine Schiff base. The product dopamine was not found outside of the enzyme, showing that the PLP-dopamine Schiff base undergoes adduct formation rather than yields product dopamine. The Pictet-Spengler adduct of PLP with L-dopa was not detected, whereas the corresponding adduct was formed when the carboxylate group of L-dopa was esterified with a methyl group to block the decarboxylation reaction. This suggests that the PLP-L-dopa Schiff base may undergo the Pictet-Spengler reaction, but its rate is much smaller than that of decarboxylation. These results indicate that Lys303 is not essential for the decarboxylation step, but has an important role in the product release, and possibly the transaldimination process.