Characterization and cloning of an extremely thermostable, Pyrococcus furiosus-type 4Fe ferredoxin from Thermococcus profundus.

An extremely thermostable [4Fe-4S] ferredoxin was isolated under anaerobic conditions from a hyperthermophilic archaeon Thermococcus profundus, and the ferredoxin gene was cloned and sequenced. The nucleotide sequence of the ferredoxin gene shows the ferredoxin to comprise 62 amino acid residues with a sequence similar to those of many bacterial and archaeal 4Fe (3Fe) ferredoxins. The unusual Fe-S cluster type, which was identified in the resonance Raman and EPR spectra, has three cysteines and one aspartate as the cluster ligands, as in the Pyrococcus furiosus 4Fe ferredoxin. Under aerobic conditions, a ferredoxin was purified that contains a [3Fe-4S] cluster as the major Fe-S cluster and a small amount of the [4Fe-4S] cluster. Its N-terminal amino acid sequence is the same as that of the anaerobically-purified ferredoxin up to the 26th residue. These results indicate that the 4Fe ferredoxin was degraded to 3Fe ferredoxin during aerobic purification. The aerobically-purified ferredoxin was reversibly converted back to the [4Fe-4S] ferredoxin by the addition of ferrous ions under reducing conditions. The anaerobically-purified [4Fe-4S] ferredoxin is quite stable; little degradtion was observed over 20 h at 100 degrees C, while the half-life of the aerobically-purified ferredoxin is 10 h at 100 degrees C. Both the anaerobically- and aerobically-purified ferredoxins were found to function as electron acceptors for the pyruvate-ferredoxin oxidoreductase purified from the same archaeon.

Human interleukin-1 receptor antagonist: large-scale expression in Bacillus brevis 47-5Q.

Interleukin-1 receptor antagonist (IL-1RA) has been used as a tool to study the biologic activity of IL-1 and as a possible therapeutic substance for inflammatory disease. To perform in vivo study, however, large quantities of IL-1RA are required. Bacillus brevis strains secrete large amounts of protein but little protease into the medium. Using B. brevis 47-5Q, we developed a large-scale expression system of human IL-1RA (HuIL-1RA). The bacteria secreted HuIL-1RA into the culture medium at very high levels, approximately 200 mg/L. The protein was isolated in one-step purification with monoclonal antibody (mAb) against HuIL-1RA. The IL-1RA molecule was determined to be functionally active by the inhibiting assay of HuIL-1-induced cell proliferation in a mouse T cell line, D10N4M.

Existence of a bioactive lipid, cyclic phosphatidic acid, bound to human serum albumin.

A novel bioactive lipid, cyclic phosphatidic acid (cPA), was identified in lipids bound to human serum albumin. A cPA fraction was extracted and purified from human serum albumin by use of a combination of preparative TLC and HPLC. Electrospray ionization mass spectrometry of the purified fraction showed molecular ions corresponding to cPA, which was composed of some different fatty acid species. The most abundant component was identified as palmitoyl-cPA by tandem mass spectrometry using collision-induced dissociation. These data have established that cPA is a naturally occurring lipid bound to human serum albumin.

Binding of salicylhydroxamic acid and several aromatic donor molecules to Arthromyces ramosus peroxidase, investigated by X-ray crystallography, optical difference spectroscopy, NMR relaxation, molecular dynamics, and kinetics.

The X-ray crystal structure of the complex of salicylhydroxamic acid (SHA) with Arthromyces ramosus peroxidase (ARP) has been determined at 1.9 A resolution. The position of SHA in the active site of ARP is similar to that of the complex of benzhydroxamic acid (BHA) with ARP [Itakura, H., et al. (1997) FEBS Lett. 412, 107-110]. The aromatic ring of SHA binds to a hydrophobic region at the opening of the distal pocket, and the hydroxamic acid moiety forms hydrogen bonds with the His56, Arg52, and Pro154 residues but is not asscoiated with the heme iron. X-ray analyses of ARP-resorcinol and ARP-p-cresol complexes failed to identify the aromatic donor molecules, most likely due to the very low affinities of these aromatic donors for ARP. Therefore, we examined the locations of these and other aromatic donors on ARP by the molecular dynamics method and found that the benzene rings are trapped similarly by hydrophobic interactions with the Ala92, Pro156, Leu192, and Phe230 residues at the entrance of the heme pocket, but the dihedral angles between the benzene rings and the heme plane vary from donor to donor. The distances between the heme iron and protons of SHA and resorcinol are similar to those obtained by NMR relaxation. Although SHA and BHA are usually considered potent inhibitors for peroxidase, they were found to reduce compound I and compound II of ARP and horseradish peroxidase C in the same manner as p-cresol and resorcinol. The aforementioned spatial relationships of these aromatic donors to the heme iron in ARP are discussed with respect to the quantum chemical mechanism of electron transfer in peroxidase reactions.

[Development from actions of bacterial phospholipases C on eucaryotic plasma membranes to molecular biology of GPI-anchored proteins].

Bacterial phospholipases C are known to act on biomembranes, since they can cleave the phosphodiester linkage between the polar head and the hydrophobic moiety of each phospholipid in these membranes. These enzymes have been classified into three groups; phosphatidylcholine (PC)-, sphingomyelin (SM)- and phosphatidylinositol (PI)-degrading phospholipases C. Enzymatic properties and toxicities of these phospholipases C are reviewed, in relation to author's research. Studies on the hemolytic phospholipases of Clostridium sp., Bacillus cereus etc., revealed that hydrolysis of choline-containing phospholipids such as PC and SM was responsible for the hemolysis of mammalian erythrocytes by these enzymes in the presence of Ca2+ and/or Mg2+. Also, the studies on a structure-activity correlation of SM-hydrolyzing phospholipase C from B. cereus disclosed the similarity of active sites between this enzyme and bovine pancreatic DNase I. By action of PI-degrading phospholipases C, several membrane proteins such as alkaline phosphatase, 5'-nucleotidase, VSG (protozoal surface glycoprotein) etc., were shown to be released from the plasma membranes of eucaryotic cells. From structural analysis, these proteins have been revealed to be glycosylphosphatidylinositol (GPI)-anchored proteins bound to the plasma membranes with carboxyl terminal-attached glycolipid. Biochemistry and molecular biology of GPI-anchored proteins, including the structures and biosynthetic routes of GPI glycolipids as well as the process of GPI attachment to proteins, requirements of C-terminal signal peptide for the protein modification by GPI, and distribution of GPI-anchored proteins in living world, are described in relation to our studies.

Application of electrospray ionization MS/MS and matrix-assisted laser desorption/ionization-time of flight mass spectrometry to structural analysis of the glycosyl-phosphatidylinositol-anchored protein.

We applied the improved sensitivity and soft ionization characteristics of electrospray Ionization (ESI)-MS/MS and matrix-assisted laser desorption/ionization(MALDI)-time of flight (TOF) mass spectrometry (MS) to analysis of the GPI-anchored C-terminal peptide derived from 5'-nucleotidase. ESI-MS/MS analysis was applied to the core structure (MW, 2,743). In the collision-induced dissociation (CID) spectrum, single-charged ions such as m/z 162 (glucosamine), 286 (mannose-phosphate-ethanolamine), and 447 ([mannose-phosphate-ethanolamine]-glucosamine) were clearly detected as characteristic fragment ions of the GPI-anchored peptide. On MALDI-TOF-MS analysis, heterogeneous peaks of GPI-anchored peptides were detected as single-charged ions in the positive mode. Product ions were obtained by post-source decay (PSD) of m/z 2,905 using curved field reflectron of TOF-MS. Most of the expected product ions derived from the GPI-anchored peptide, containing the core structure and an additional mannose side chain, were successively obtained. Thus, ESI-MS/MS and MALDI-TOF-PSD-MS proved to be effective and sensitive methods for analyzing the GPI-anchored peptide structure with less than 10 pmol of sample. These characteristic fragments or fragmentation patterns seem to be very useful for identification of GPI-anchored C-terminal peptides derived from any kind of GPI-anchored protein.

Roles of asp126 and asp156 in the enzyme function of sphingomyelinase from Bacillus cereus.

To elucidate the roles of conserved Asp residues of Bacillus cereus sphingomyelinase (SMase) in the kinetic and binding properties of the enzyme toward various substrates and Mg2+, the kinetic data on mutant SMases (D126G and D156G) were compared with those of wild type (WT) enzyme. The stereoselectivity of the enzyme in the hydrolysis of monodispersed short-chain sphingomyelin (SM) analogs and the binding of Mg2+ to the enzyme were not affected by the replacement of Asp126 or Asp156. The pH-dependence curves of kinetic parameters (1/Km and kcat) for D156G-catalyzed hydrolysis of micellar SM mixed with Triton X-100 (1:10) and of micellar 2-hexadecanoylamino-4-nitrophenylphosphocholine (HNP) were similar in shape to those for WT enzyme-catalyzed hydrolysis. On the other hand, the curves for D126G lacked the transition observed for D156G and WT enzymes. Comparison of the values and the shape of pH-dependence curves of kinetic parameters indicated that Asp126 of WT SMase enhances the enzyme's catalytic activity toward both substrates and its binding of HNP but not SM. The deprotonation of Asp126 enhances the substrate binding and slightly suppresses the catalytic activity toward both substrates. Asp156 of WT SMase acts to decrease the binding of both substrates and the catalytic activity to HNP but not SM. From the present study and the predicted three-dimensional structure of B. cereus SMase, Asp126 was thought to be located close to the active site, and its ionization was shown to affect the catalytic activity and substrate binding.

Identification of a new glycosylphosphatidylinositol-anchored 42-kDa protein and its C-terminal peptides from bovine erythrocytes by gas chromatography-, time-of-flight-, and electrospray-ionization-mass spectrometry.

By treatment with phosphatidylinositol-specific phospholipase C (PIPLC), we obtained several candidates of glycosylphosphatidylinositol (GPI)-anchored proteins such as 55, 42, 40, and 30 kDa from bovine erythrocyte membrane, in addition to the well-known GPI-anchored protein acetylcholinesterase. In these proteins, the presence of myo-inositol was confirmed by gas chromatography (GC)-mass spectrometry. Among them, the 42-kDa protein was further analyzed by electrospray-ionization (ESI)-mass spectrometry (MS) after hydrolysis by lysyl endoprotease. By liquid chromatography (LC)-ESI-MS analysis, C-terminal peptides bearing the products of GPI (Ct. GPI-peptides) were effectively detected by combination with in-source collision and multifunctional scanning for the several characteristic fragment ions from the GPI-anchor structure. Existence of microheterogeneity was also observed in the Ct. GPI-peptides from the 42-kDa protein. This result was confirmed by analysis with time-of-flight (TOF)-MS. Furthermore, one of the Ct. GPI-peptides was analyzed in ESI-MS-MS mode. Characteristic fragment ions were effectively detected by collision-induced decay. By the result of MS-MS analysis, this GPI-anchor structure was revealed to contain additional N-acetyl hexosamine. By the above-mentioned method, the C-terminal GPI-anchor structure can be easily identified from the target protein even if its amino acid sequence data are not available.

Mg2+ binding and catalytic function of sphingomyelinase from Bacillus cereus.

The modes of Mg2+ binding to SMase from Bacillus cereus were studied on the basis of the changes in the tryptophyl fluorescence intensity. This enzyme was shown to possess at least two binding sites for Mg2+ with low and high affinities. The effects of Mg2+ binding on the enzymatic activity and structural stability of the enzyme molecule were also studied. The results indicated that the binding of Mg2+ to the low-affinity site was essential for the catalysis, but was independent of the substrate binding to the enzyme. It was also indicated that the alkaline denaturation of the enzyme was partly prevented by the Mg2+ binding, whereas no significant protective effect was observed against the denaturation by urea. The pH dependence of the kinetic parameters for the hydrolysis of micellar HNP and mixed micellar SM with Triton X-100 (1:10), catalyzed by SMase from B. cereus, was studied in the presence of a large amount of Mg2+ to saturate both the low- and high-affinity sites. The pH dependence curves of the logarithm of 1/Km for these two kinds of substrates were similar in shape to each other, and showed a single transition. On the other hand, the shapes of the pH dependence curves of the logarithm of kcat for these two kinds of substrates were different from each other. The pH dependence curve for micellar HNP showed three transitions and, counting from the acidic end of the pH region, the first and third transitions having tangent lines with slopes of +1 and -1, respectively. On the other hand, the curve for mixed micellar SM with Triton X-100 showed one large transition with a slope of +1 (the first transition) and a very small transition (the third transition). On the basis of the present results and the three-dimensional structure of bovine pancreatic DNase I, which has a primary structure similar to that of B. cereus SMase, we proposed a catalytic mechanism for B. cereus SMase based on general-base catalysis.

Molecular cloning of a GPI-anchored aminopeptidase N from Bombyx mori midgut: a putative receptor for Bacillus thuringiensis CryIA toxin.

An aminopeptidase N (APN) with a molecular weight of 110kDa was released from the midgut membrane of Bombyx mori by phosphatidylinositol-specific phospholipase C (PI-PLC), and purified to a homogeneous state. This 110-kDa APN was different from the 100-kDa APN that we previously reported, in chromatographic behaviors, substrate specificity, and N-terminal and internal amino acid sequences. However, the N-terminal sequence of 110-kDa APN, DPAFRLPTTTRPRHYQVTLT, was highly homologous with those of Manduca sexta and Heliothis virescens APNs, which were identified as a receptor for an insecticidal toxin of Bacillus thuringiensis. From a B. mori midgut cDNA library, we cloned the 110-kDa APN cDNA that possessed a 2958-bp open reading frame encoding a 111573-Da polypeptide of 986 residues. The sequence of the eicosa-peptide Asp42Thr61 deduced from the cDNA was completely matched with the N-terminal sequence of the mature 110-kDa APN. One potential N-glycosylation site, HEXXHXW zinc-binding motif and characteristic proline-rich repeats were observed in the ORF. Moreover, the primary sequence contained two hydrophobic peptides on N- and C-termini. The N-terminal peptide sequence showed characteristics of leader peptide for secretion and the C-terminal peptide contained a possible glycosylphosphatidylinositol (GPI) anchoring site. Taken together, the deduced amino acid sequence suggests that the 110-kDa APN is a GPI-anchored protein and a specific receptor protein for B. thuringiensis CryIA delta-endotoxin.

Characterization of aminopeptidase N from the brush border membrane of the larvae midgut of silkworm, Bombyx mori as a zinc enzyme.

Three GPI-anchored proteins, aminopeptidase N, alkaline phosphatase and alkaline phosphodiesterase I were released from the midgut brush border membrane of Bombyx mori by phosphatidylinositol-specific phopholipase C and the aminopeptidase N was purified to a homogeneous state. N-terminus and 6 internal sequences, one of which possessed part of zinc-binding motif, showed homology with those from other species. The zinc content in purified aminopeptidase N was estimated as approximately 0.72 mol/mol of the protein and 1,10-phenanthroline completely inhibited the enzyme activity, suggesting zinc requirement for the activity. The aminopeptidase N activity was inhibited not only by probestin and actinonin, but also strongly depressed by amastatin, while leuhistin and bestatin were less inhibitory. These suggest that the active site of aminopeptidase N might be structurally different from those of mammals. Calcium and magnesium ions stimulated the aminopeptidase N activity, but copper ion was rather inhibitory. Zinc ion showed bi-modal effect on the activity, i.e., stimulatory at low concentration, but inhibitory at higher than 100 microM. This inhibition was completely restored by EDTA. These results suggest that the aminopeptidase N possesses two zinc ion-binding sites with high and low affinity as essential and inhibitory one, as well as some regulatory metal-binding sites.

Two distinct phosphatidylinositol-specific phospholipase Cs from Streptomyces antibioticus.

Two phosphatidylinositol-specific phospholipase C (PI-PLC) genes from Streptomyces antibioticus were cloned by a shotgun method using Streptomyces lividans TK24 as a host. The genes of the two PI-PLCs (named as PLC1 and PLC2) were adjoined and opposite in the direction of transcription/translation. Both of them were confirmed to be expressed in S. antibioticus. The two enzymes were different in the following properties. (i) PLC2 had considerable sequence similarity to other bacterial PI-PLCs, while PLC1 had a short stretch that was similar to PI-PLCs of eukaryotes rather than the other bacterial enzymes. (ii) PLC1 was Ca2+-dependent, whereas PLC2 was not. (iii) PLC1 generated myo-inositol-1-phosphate and myo-inositol-1:2-cyclic phosphate simultaneously from PI, but PLC2 showed sequential formation of them. (iv) PLC2 has GPI-anchor-degrading activity while PLC1 does not have. Both enzymes did not hydrolyze phosphatidylcholine, phosphatidylinositol-4-monophosphate and phosphatidylinositol-4,5-bisphosphate. Both PLC1 and PLC2 contained two histidine residues that might be catalytic residues. PLC1 has residues that possibly form a Ca2+-binding site. Then it was suggested that both PLC1 and PLC2 act according to the catalytic mechanism using the two histidine residues as proposed in both eukaryotic and prokaryotic enzymes, but that PLC1 has a more 'eukaryotic' mechanism in which Ca2+ participates than that of the Ca2+-independent bacterial enzymes. Thus, we propose that PLC2 is a conventional 'bacteria-type' enzyme, while PLC1 is more closely related to the eukaryotic enzymes rather than the bacterial enzymes.

Cloning and sequence analysis of the aminopeptidase N isozyme (APN2) from Bombyx mori midgut.

An aminopeptidase N (APN) isozyme having the molecular weight of 90 kDa, was released by phosphatidylinositol-specific phospholipase C (PI-PLC) and purified homogeneously, from the brush border membrane of Bombyx mori. From the result of cDNA cloning, the primary structure of 90 kDa APN proved to consist of 948 amino acid residues, containing a typical metalloprotease-specific zinc-binding motif in the deduced sequence. Moreover, the primary sequence contained two hydrophobic segments on N- and C-termini. The N-terminal one showed characteristics of leader peptide for secretion and the C-terminal one contained a possible glycosylphosphatidylinositol (GPI) anchoring site, suggesting that the APN encoded by the cDNA is not only a zinc-binding enzyme, but also a GPI-anchored protein. The primary sequence is significantly homologous with those of insect and mammalian APNs, and contains four conserved segments around the zinc-binding motif, two potential N-glycosylation sites and four conserved Cys residues. The deduced primary sequence had 30.7% identity with that of B. mori 110 kDa APN, and did not contain the N-terminal and internal amino acid sequences of B. mori 100 kDa APN, revealing B. mori 90 kDa APN to be the third isozyme on the midgut brush border membrane. On the other hand, the primary sequence of 90 kDa APN showed high homology with Manduca sexta APN2 (65.1% identity) and Plutella xylostella APN2 (63.8% identity). It appears that the B. mori 90 kDa APN should be classified in the insect apn2 cluster and differentiated from insect apn1 and mammalian apn clusters by phylogenetic analysis. These results suggest that 90 kDa APN isozyme encoded by the cDNA is a product of B. mori apn2 gene.

Chiral discrimination of 2'-deoxy-L-cytidine and L-nucleotides by mouse deoxycytidine kinase: low stereospecificities for substrates and effectors.

The effects of four kinds of 2'-deoxy-L-nucleoside 5'-triphosphates and L-ATP, which are enantiomers of natural D-dNTPs and D-ATP, on deoxycytidine kinase (dCK) partially purified from mouse leukemic P388 cells were investigated. Only L-dCTP did not act as a phosphate donor while other L-dNTPs and L-ATP showed 15-30% of the activity of the corresponding D-dNTP or D-ATP. L-dCTP inhibited dCK non-competitively with 2'-deoxycytidine (D-dCyd) and competitively with phosphate donor D-ATP. These inhibitory effects of L-dCTP on dCK were similar to the results of earlier studies using D-dCTP. Thus, L-dCTP was shown to be capable of serving as a feedback inhibitor for dCK instead of D-dCTP. Mouse dCK was also able to phosphorylate L-dCyd, as demonstrated in the case of human dCK. The present results suggest that the chirality of not only dCyd as the substrate but also nucleotides as the substrate or effector is not strictly discriminated by dCK.

Mutational analysis of the C-terminal signal peptide of bovine liver 5'-nucleotidase for GPI anchoring: a study on the significance of the hydrophilic spacer region.

Bovine liver 5'-nucleotidase is a GPI-anchored protein whose Ser523 attaches to GPI as the omega-site. For GPI-modification, pro-protein of the enzyme possesses a signal peptide at the C-terminus, comprising a hydrophilic spacer sequence of 8 amino acid residues and the following hydrophobic region of 17 amino acid residues. The C-terminal signal peptide is replaced by GPI on a luminal leaflet of endoplasmic reticulum. To characterize the C-terminal signal peptide for GPI modification, we constructed a series of deletion and elongation mutant genes, altering length of the hydrophilic spacer sequence by site-directed mutagenesis. Systematic deletion and Ala insertion of the sequence showed that the sequence of 6-14 residues were compatible for GPI modification. For GPI transfer to the pro-protein, the optimum length of spacer sequence would be 8, being consistent with natural selection. The spacer sequence may play a role for leading the omega-residue correctly to the active site of putative GPI transamidase. The elongation of the spacer is more permissible than deletion. Nevertheless, the length of the spacer sequence may influence efficiency of GPI modification by its positive or negative control.

Identification of chiro-inositol and its formation by isomerization of myo-inositol during hydrolysis of glycosylphosphatidylinositol-anchored proteins.

myo-Inositol has been believed to be a sole inositol isomer existing in phosphatidylinositol (PI) and related derivatives. In this experiment, chiro-inositol, an inositol isomer other than myo-inositol, was identified in hydrolytic products from several GPI-anchored proteins. The chiro-inositol contents in several different GPI-anchored proteins including 5'-nucleotidase of bovine liver and alkaline phosphatase of mouse NS-1 varied with hydrolytic conditions of these GPI anchor. Isomerization of 20-60% of myo-inositol occurred on the hydrolysis in 6 N HCl solution. Under the hydrolytic conditions of a HCl gas stream in place of solution, however, isomerization was very low (less than 0.1%). Even in the hydrolysis under HCl gas stream, existence of CNBr accelerated the isomerization of inositol in GPI up to 70-95%. In the hydrolysis of phosphatidylinositol or myo-inositol 1-phosphate, however, a significant amount of chiro-inositol was not detected in 6 N HCl solution or in the existence of CNBr under the HCl stream. These facts indicated that isomerization occurred during the hydrolysis of the GPI anchor, when myo-inositol is substituted by glucosamine at 6-OH and is substituted by phosphate at 1-OH. It also suggested that the former identification of chiro-inositol in GPI structure in the various reports might be due to isomerization.

Growth inhibition, morphological change, and ectoenzyme release of LLC-PK1 cells by phosphatidylinositol-specific phospholipase C of Bacillus thuringiensis.

Phosphatidylinositol-specific phospholipase C (PI-PLC) from Bacillus thuringiensis added to a culture of LLC-PK1 cells inhibited cell growth by 40%. In contrast with normal cells, the cells cultured in the presence of PI-PLC showed needle-like appendages which seemed to have been formed due to portions of the cell remaining adhered to the culture dish as the cell shrank. When LLC-PK1 cells were treated with PI-PLC, significant amounts of alkaline phosphatase and alkaline phosphodiesterase I were released specifically from the apical surface of the LLC-PK1 cells. Furthermore, PI-PLC treatment caused a delay of enzyme production and dome formation. These data indicate that glycosyl-phosphatidylinositol (GPI)-anchored proteins on the surface of LLC-PK1 cells are important in cell growth and differentiation. Also, the combined use of LLC-PK1 cells and PI-PLC of B. thuringiensis is effective for investigating the function of GPI-anchor proteins.