* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Thermostable glycerol kinase from a
Gene regulatory network wikipedia , lookup
Catalytic triad wikipedia , lookup
Gene expression wikipedia , lookup
Western blot wikipedia , lookup
NADH:ubiquinone oxidoreductase (H+-translocating) wikipedia , lookup
Restriction enzyme wikipedia , lookup
Ancestral sequence reconstruction wikipedia , lookup
Citric acid cycle wikipedia , lookup
Proteolysis wikipedia , lookup
Evolution of metal ions in biological systems wikipedia , lookup
Community fingerprinting wikipedia , lookup
Oxidative phosphorylation wikipedia , lookup
Magnesium transporter wikipedia , lookup
Silencer (genetics) wikipedia , lookup
Genomic library wikipedia , lookup
Nucleic acid analogue wikipedia , lookup
Enzyme inhibitor wikipedia , lookup
Expression vector wikipedia , lookup
Genetic code wikipedia , lookup
Protein structure prediction wikipedia , lookup
Two-hybrid screening wikipedia , lookup
Point mutation wikipedia , lookup
Metalloprotein wikipedia , lookup
Biochemistry wikipedia , lookup
Artificial gene synthesis wikipedia , lookup
Protein Engineering vol.11 no.12 pp.1219–1227, 1998 Thermostable glycerol kinase from a hyperthermophilic archaeon: gene cloning and characterization of the recombinant enzyme Yuichi Koga1, Masaaki Morikawa1, Mitsuru Haruki1, Haruki Nakamura2, Tadayuki Imanaka3 and Shigenori Kanaya1,4 1Department of Material and Life Science, Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0871, 2Department of Bioinformatics, Biomolecular Engineering Research Institute, 6-2-3 Furuedai, Suita, Osaka 565-0874 and 3Department of Synthetic Chemistry and Biological Chemistry, Graduate School of Engineering, Kyoto University, Kyoto 606-8501, Japan 4To whom correspondence should be addressed. E-mail: [email protected] The Pk-glpK gene, which encodes glycerol kinase (GK) from a hyperthermophilic archaeon Pyrococcus kodakaraensis KOD1, was cloned and expressed in Escherichia coli. The amino acid sequence of this enzyme (Pk-GK) deduced from the nucleotide sequence showed 57% identity with that of E.coli GK and 47% identity with that of human GK. Pk-GK, which has a molecular weight of 55 902 (497 amino acid residues), was purified from E.coli and characterized. Despite the high sequence similarity, Pk-GK and E.coli GK are greatly divergent in structure and function from each other. Unlike E.coli GK, which exists as a tetramer, Pk-GK exists as a dimer. The preferred divalent cation for Pk-GK is Co21, instead of Mg21. The optimum pH and temperature for Pk-GK activity are 8.0 and 80°C, respectively. Pk-GK can utilize other nucleoside triphosphates than ATP as a phosphoryl donor. It is fairly resistant to an allosteric inhibitor of E.coli GK, fructose-1,6-bisphosphate. Determination of the kinetic parameters indicates that the Km value of the enzyme is 15.4 µM for ATP and 111 µM for glycerol and its kcat value is 940 s–1. The enzyme was shown to be fairly resistant to irreversible heat inactivation and still retained 50% of its enzymatic activity even after heating at 100°C for 30 min. Construction of a model for the three-dimensional structure of the enzyme suggests that the formation of extensive ion-pair networks is responsible for the high stability of this enzyme. Keywords: glycerol kinase/hyperthermophilic archaeon/ionpair networks/molecular cloning/subunit structure Introduction Glycerol kinase (EC 2.7.1.30; ATP:glycerol 3-phosphotransferase, GK) catalyzes Mg,ATP-dependent phosphorylation of glycerol to produce glycerol-3-phosphate (Lin, 1976), which is an important metabolic intermediate for both energy production (glycolysis) and phospholipid biosynthetic pathways. The enzyme is ubiquitously present in various organisms from bacteria to human and a number of the genes encoding GK have been cloned from them and sequenced. They include the glpK genes from Escherichia coli (Pettigrew et al., 1988), Bacillus subtilis (Holmberg et al., 1990), Pseudomonas aeruginosa (Schweizer et al., 1997) and Enterococcus casseliflavus (Charrier et al., 1997), the GUT1 gene from Saccharomyces © Oxford University Press cerevisiae (Pavlik et al., 1993) and the Xp gene from humans (Sargent et al., 1994). In addition, complete determination of the whole genome sequences of several organisms provides information on the amino acid sequences of GKs, such as that from Archaeoglobus fulgidus, which has been deposited in GenBank with accession number AE001044. Comparison of these sequences suggests that the amino acid sequence of GK is relatively well conserved among various GKs. For example, the identity in the amino acid sequence between E.coli GK and human GK is 51%. This is probably because GK is a key enzyme in glycerol metabolism and therefore the enzyme cannot have become structurally and functionally greatly divergent during evolution. Among various GKs, E.coli GK has been most extensively studied with respect to its functions and structures. This enzyme acts in the rate-limiting step in the glycerol utilization pathway and its catalytic activity is closely regulated by two allosteric inhibitors, IIIglc [EIIAGlc according to a new nomenclature proposed by Saier and Reizer (1992)] of the phosphotransferase system (PTS) and fructose-1,6-bisphosphate (FBP) (Thorner and Paulus, 1973; de Riel and Paulus, 1978a,b; Novotny et al., 1985; Liu et al., 1994). The enzyme is a tetrameric protein composed of identical subunits (Thorner and Paulus, 1971), but can be dissociated into dimers at low protein concentrations (de Riel and Paulus, 1978b). The interaction between the enzyme and one of the substrates, ATP, showed negative cooperativity, whereas the interaction between the enzyme and another substrate, glycerol, followed Michaelis–Menten kinetics (Thorner and Paulus, 1973; Novotny et al., 1985; Pettigrew et al., 1990). Kinetic and substrate binding studies have suggested that the dimer is the enzyme form in the assay and its subunits display different ATP binding affinities (Pettigrew et al., 1990). The crystal structure of the enzyme in a complex with the unphosphorylated form of EIIAGlc, glycerol and ADP (Hurley et al., 1993), and also that in which glycerol is replaced by glycerol 3-phosphate (Feese et al., 1994), were determined at 2.6 Å resolution. According to these structures, unphosphorylated EIIAGlc binds far from the active site of GK, in which glycerol and ADP binds, through the formation of an intermolecular Zn(II) binding site. Thus, it has been postulated that long-range conformational changes mediate the inhibition of GK by EIIAGlc. Likewise, a dimer–tetramer assembly reaction has been postulated to mediate the inhibition of GK by FBP (de Riel and Paulus, 1978a). It is well known that all organisms can be evolutionarily grouped into three classes, eukarya, bacteria and archaea (Woese and Fox, 1977; Woese et al., 1990). Among them, archaea, especially hyperthermophilic ones, are expected to retain traces of early life forms and to produce enzymes which may represent prototypes in the same protein family, because they can grow in extreme environments which are probably similar to that of the primitive earth. Therefore, it would be informative to analyze the structures and functions of GKs 1219 Y.Koga et al. from hyperthermophilic archaea, which may be different from those of GKs from bacteria and eukarya. Consequently, these studies may serve to elucidate the adaptation mechanisms of proteins to high temperature and the molecular evolution of the enzyme. However, none of the GK enzymes has been isolated from archaea. Therefore, we decided to characterize GK from a hyperthermophilic archaeon, Pyrococcus kodakaraensis KOD1. Pyrococcus kodakaraensis KOD1, which has previously been designated Pyrococcus sp. KOD1, was isolated from a solfatara at a wharf on Kodakara Island, Kagoshima, Japan (Morikawa et al., 1994). The growth temperature of this strain ranged from 65 to 100°C and the optimal temperature is 95°C. Here we report that GK from P.kodakaraensis KOD1 showed distinctive structural and enzymatic characteristics as compared with those from mesophilic sources. Like other enzymes from this strain, it is more stable than the bacterial enzymes and shows broad metal ion and NTP specificities. In addition, unlike E.coli GK, it exists in a dimeric form, is not sensitive to FBP inhibition and does not show negative cooperativity for ATP binding. We will discuss on the structure–function relationships of this enzyme, based on a model for the threedimensional structure. Materials and methods Cells and plasmids Pyrococcus kodakaraensis KOD1 was isolated from a solfataric hot spring at a wharf on Kodakara Island, Kagoshima, Japan (Morikawa et al., 1994). Escherichia coli strain JM109 [recA1, endA1, gyrA96, thi, hsdR17, SupE44, relA1, λ–, ∆(lac-ProAB)/ F9, traD36, ProA1B1, lacIqZ∆M15] and plasmid pUC19 for DNA manipulations were obtained from Toyobo. E.coli strain BL21(DE3) [F–, ompT, hsdSB(rB–mB–), gal, dcm(DE3)] and plasmid pET-8c for gene expressions were obtained from Novagen. Materials Restriction enzymes, DNA polymerase and T4 DNA ligase were obtained from Takara Shuzo. Glycerol-3-phosphate oxidase (GPO) and peroxidase (PO) were kindly provided by Toyobo. [1,3-14C]glycerol (100 mCi/mmol) was purchased from ICN. Cloning of the Pk-glpK gene A portion of the Pk-glpK gene was first amplified by PCR with a combination of forward (59-TGCTGTTTGGTACGG39) and reverse (59-AGCCAGTGCCATAGGTGTTCTTCG39) primers. The genomic DNA, which was prepared from a sarkosyl lysate of P.kodakaraensis KOD1 cells as described previously (Imanaka et al., 1981), was used as a template. The resulting 318-base pair (bp) DNA fragment was used as a probe for Southern blotting and colony hybridization to clone the entire Pk-glpK gene. PCR was performed in 30 cycles with a Perkin-Elmer thermal cycler (GeneAmp PCR System 2400) using Vent polymerase (New England Biolabs), as described previously (Kim et al., 1995). All DNA oligomers were synthesized by Sawady Technology. General DNA manipulations were carried out as described previously (Sambrook et al., 1989). The nucleotide sequence was determined by the dideoxy-chain termination method (Sanger et al., 1977) with an ABI PRISM 310 genetic analyzer (Perkin-Elmer). 1220 Overexpression and purification of Pk-GK The Pk-glpK gene was amplified by PCR with a combination of forward (59-ATATACCATGGAAAAGTTCGTTCTTTC39) and reverse (59-AATGGATCCATATCAATTTGATTTTGCACT-39) primers, where the underlines represent the NcoI and BamHI sites, respectively, and the ATG codon for the initiation of the translation and the sequence complementary to the termination codon TGA are shown in italics. The resultant 1.5-kb DNA fragment was digested with NcoI and BamHI and cloned into the NcoI–BamHI site of plasmid pET8c to create plasmid pET-pkgk. An overproducing strain for Pk-GK was constructed by transforming E.coli BL21(DE3) with this plasmid. For the overproduction of Pk-GK, an E.coli BL21(DE3) transformant with pET-pkgk was grown at 37°C in NZCYM medium (Novagen) containing 50 µg/ml ampicillin. When the absorbance at 600 nm of the culture reached ~0.6, 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG) was added to the culture medium and cultivation was continued for an additional 8 h. Cells were then harvested by centrifugation at 6000 g for 10 min and subjected to the purification procedures. Cells were suspended in 50 mM sodium phosphate buffer (pH 7.5), disrupted by sonication with a Model 450 sonifier (Branson Ultrasonic) and centrifuged at 15 000 g for 30 min. The supernatant was heated at 90°C for 30 min and then applied to a column of MonoQ (Pharmacia) equilibrated with the same buffer. The enzyme eluted from the column as a single peak at an NaCl concentration of ~0.5 M by linearly increasing the NaCl concentration from 0 to 1.0 M in the same buffer. All purification procedures, except for heat treatment, were carried out at 4°C. The production level and the purity of the enzyme were analyzed by SDS–PAGE (Laemmli, 1970). Determination of enzymatic activity The enzymatic activity of Pk-GK was determined in 800 µl of 50 mM glycine–NaOH (pH 8.0) containing 2 mM CoCl2, 4.6 mM glycerol and 3.5 mM ATP at 80°C, unless stated otherwise. After incubation for 5 min, the reaction was terminated by the addition of 16 µl of 0.5 M EDTA and the reaction mixture was kept on ice for a few minutes. The amount of glycerol-3-phosphate (G-3-P) in the reaction mixture was then determined colorimetrically at 37°C by coupled enzyme reactions catalyzed by glycerol-3-phosphate oxidase (GPO) and peroxidase (PO) as described previously (Bergmeyer et al., 1961). One unit of enzymatic activity was defined as the amount of enzyme producing 1 µmol of glycerol-3phosphate in 1 min. The specific activity was defined as the enzymatic activity per milligram of protein. For determination of the kinetic parameters the assay solution contained, in a volume of 100 µl, 50 mM Tris–HCl (pH 7.5), 2 mM CoCl2, 2 mM or appropriate concentration of [14C]glycerol, 2 mM or appropriate concentration of ATP and 0.06 µg of Pk-GK. After incubation for 5 min at 80°C, the reaction was terminated by the addition of 2 µl of 0.5 M EDTA. An aliquot of the reaction mixture was then applied to a DEAE-cellulose filter-paper (diameter 2.4 cm) (Toyo Roshi). After washing the filter-paper with 0.1 M glycerol, the amount of [14C]G-3-P bound to the paper was determined using Instant Imager (Packard). The production of G-3-P followed Michaelis–Menten kinetics and the kinetic parameters, Km and kcat, were determined from the Lineweaver–Burk plot. Protein concentration was determined from the UV absorption at 280 nm by using an A2800.1% value of 2.2 (molar Glycerol kinase from a hyperthermophilic archaeon absoptivity 1.233105 l/mol.cm at 280 nm). This value was determined by measuring the amount of the protein by amino acid analysis Biochemical characterization The molecular weight of Pk-GK was estimated by gel filtration on a column of HiLoad Superdex 200 (Pharmacia) equilibrated with 10 mM Tris–HCl (pH 8.0) containing 150 mM NaCl. Standard molecular marker (Bio-Rad) was applied to the column for calibration of molecular weight. The N-terminal amino acid sequence of the enzyme was determined with a Procise automated sequencer (Perkin-Elmer Model 491). Amino acid analysis was carried out with a Hitachi L-8500 automatic amino acid analyzer. Samples were hydrolyzed at 110°C for 24 h with 6 M HCl containing 0.2% (v/v) phenol. Circular dichroism (CD) measurement CD spectra were measured on a J-720W spectropolarimeter (Japan Spectroscopic). Far-UV (200–260 nm) and near-UV (250–320 nm) CD spectra were obtained at 30°C by using a solution containing enzyme at 0.13 and 0.25 mg/ml in 50 mM sodium phosphate buffer (pH 7.9) containing 0.5 M NaCl in a cell with an optical path of 2 and 10 mm, respectively. The mean residue ellipticity, [θ], which has the units of ° cm2/ dmol, was calculated by using an average amino acid molecular weight of 110. Construction of a three-dimensional model of Pk-GK A tertiary model of Pk-GK was constructed by homology modeling using the following programs: a loop search method for the backbone structure (Nakamura et al., 1991), a dead-end elimination method for the side-chain conformation (Tanimura et al., 1994) and conformation energy minimization for the structure refinement (Morikami et al., 1992), using the AMBER force field (Weiner et al., 1986). Results Cloning and sequencing of the Pk-glpK gene On the assumption that conserved amino acid sequences in various glycerol kinases (GKs) are also conserved in Pk-GK and the nucleotide sequence of the Pk-glpK gene is similar to that of the E.coli glpK gene in the regions where the amino acid sequences encoded are conserved, various DNA oligomers (15–25 bases) with sequences identical with those in the E.coli glpK gene were synthesized and used as a primer to amplify a portion of the Pk-glpK gene by PCR. Among various combinations of forward and reverse primers examined, a combination of the forward primer (15 bases) encompassing the sequences coding for amino acids 161–166 of E.coli GK and the reverse primer (24 bases) encompassing the sequences coding for amino acids 261–269 of E.coli GK was shown to be effective in amplifying the 318-bp DNA fragment from the KOD1 genome. Determination of the DNA sequence indicated that this DNA fragment is a portion of the Pk-glpK gene, which encodes the amino acid sequence of Pk-GK from residues 157 to 263. By using this DNA fragment as a probe for Southern hybridization, the 4-kb BamHI fragment or 2-kb HindIII fragment was shown to contain this 318-bp DNA fragment (data not shown). Determination of the DNA sequence indicated that the 4-kb BamHI fragment contains the entire Pk-glpK gene, whereas the 2-kb HindIII fragment contains a part of it, in which the 39-terminal region is truncated, because the HindIII site is located within the Pk-glpK gene. Identification of a strong ribosomal binding signal (59- AGGTGAT-39), which is complementary to the 39-terminal sequence of the 16S rRNA from KOD1 (fourth to tenth residues from the 39-terminus) (Morikawa et al., 1994), at the 59-terminal region of a long uninterrupted reading frame with ~1600 bases allowed us to choose the AUG codon, which is located six bases downstream from this ribosomal binding signal, for initiation of translation. The protein from this region would be composed of 497 amino acid residues with a calculated molecular weight of 55 902. A possible TATA-like promoter site (59-TTAAAAA-39, A-box) is located 67 base pairs upstream from the initiation codon and a possible transcription termination signal with a tandem repeat of thymine is located ~30 base pairs downstream from the termination codon. The nucleotide sequence of the Pk-glpK gene, and also its flanking sequences, are deposited in DDBJ with accession number AB012099. Comparison of amino acid sequences The amino acid sequence of Pk-GK deduced from the nucleotide sequence 1s compared with those of GKs from E.coli, Archaeoglobus fulgidus and humans in Figure 1. These sequences were chosen as representatives of those of GKs from bacteria, archaea and eukaryotes. Among various GK sequences so far available, the E.coli GK sequence gave the closest similarity to the Pk-GK sequence with an identity of 57%. The amino acid sequence identities between Archaeoglobus fulgidus GK and Pk-GK and between human GK and Pk-GK were 36 and 47%, respectively. All of the amino acid residues, except for Ile313, which have been shown to be involved in the bindings with glycerol, ADP and Mg21 for E.coli GK (Hurley et al., 1993) are fully conserved in the PkGK sequence. They are Arg82, Glu83, Trp101, Tyr132, Asp239 and Phe264 for glycerol binding, Asp9 and Asp239 for catalytic function and probably for divalent cation binding and Arg16, Gly260, Thr261, Gly303, Gln307, Ala319, Leu374, Ile377, Gly402, Ala403 and Asn406 for ADP binding. Ile313 of E.coli GK, which contributes to form a hydrophobic pocket for the binding of adenine base, is replaced by Val (Val306) in the Pk-GK sequence. Overproduction and purification In order to facilitate the preparation of Pk-GK in an amount sufficient for biochemical characterizations, we constructed plasmid pET-pkgk, in which transcription of the Pk-glpK gene is initiated by the T7 promoter. This plasmid was used to transform E.coli BL21(DE3) to ampicillin resistance. Induction of the gene by the addition of IPTG caused accumulation of the enzyme as the most abundant protein in the cells (Figure 2). The production level of the enzyme was estimated to be ~40 mg/l culture by SDS–PAGE as shown in Figure 2. Because the amount of the enzyme in the soluble fraction obtained after sonication lysis (crude lysate) was comparable to that in the whole cell extract, the enzyme should be accumulated in cells in a soluble form (data not shown). From the crude lysate, the enzyme was purified in only two steps with a yield of 30–40% (Table I). The purified enzyme gave a single band on SDS–PAGE with a molecular weight of 55 000 (Figure 2). This value was closely comparable to that calculated from its amino acid sequence. Biochemical properties To confirm the amino acid sequence of Pk-GK predicted from the DNA sequence, the NH2-terminal amino acid sequence and amino acid composition of the protein were determined. 1221 Y.Koga et al. Fig. 2. Comparison of the purity of Pk-GK by SDS–PAGE. Samples were subjected to electrophoresis on a 10% polyacrylamide gel in the presence of SDS. After electrophoresis, the gel was stained with Coomassie Brilliant Blue. Lane 1, a low molecular weight marker kit (Pharmacia LKB Biotechnology) containing phosphorylase b (94K), bovine serum albumin (67K), ovalbumin (43K), carbonic anhydrase (30K), trypsin inhibitor (20K) and α-lactalbumin (14K); lane 2, crude extract from E.coli BL21(DE3) transformant with pET-pkgk; lane 3, purified Pk-GK. Table I. Purification of recombinant Pk-GK from a 1 l culture of E.coli cells Fraction Total protein Total activity Specific activity Yield (%) (mg) (units) (units/mg) Crude extract 162 Heat treatment 36 Mono Q fraction 9.2 Fig. 1. The alignment of the Pk-GK, E.coli GK, A. fulgidus GK and human GK sequences. The amino acid residues of GKs from E.coli, A.fulgidus and humans, which are identical with those of Pk-GK, are indicated by a dash (–). For human GK, the sequence from Ala7 to Asp523 is shown. The ranges of the 16 α-helices and the 25 β-strands of E.coli GK are shown above the sequences, according to the crystal structure determined by Hurley et al. (1993). Numbers above the sequences indicate the position of the residue in Pk-GK, which starts from the initiator methionine. Numbers along the sequences indicate the position of the amino acid residue listed in the end of each line for GKs from E.coli, A.fulgidus and humans. The amino acid residues that are possibly involved in the catalytic function and the substrate-binding of Pk-GK are shown in bold face. The regions in the E.coli GK sequence, which are involved in the subunit-subunit interactions (Hurley et al., 1993), are shaded. O–X, O–Y and O–Z in parentheses represent the three types of the subunit–subunit interactions. The sequences have been deposited in Swiss-Prot with code numbers P08859 for E.coli GK (E. co) and P32189 for human GK (Hum), in EMBL/GenBank with accession number AE001044 for A.fulgidus GK (A. fu) and in DDBJ with accession number AB012099 for Pk-GK. The sequence of the NH2-terminal 10 amino acid residues was MEKFVLSLDE, which was identical with that predicted from the DNA sequence. In addition, there is good agreement between the amino acid composition predicted from the DNA 1222 13040 7160 4140 80 199 450 100 54.9 31.7 Fig. 3. Far- and near-UV CD spectra of Pk-GK. The CD spectra were measured in 50 mM sodium phosphate buffer (pH 7.9) containing 0.5 M NaCl at 30°C as described under Materials and methods. The mean residue ellipticity, [θ], is given in units of °·cm2/dmol. sequence and that determined experimentally (data not shown). The molecular weight of the enzyme in its native state was estimated to be 125 000 by gel filtration chromatography (data not shown). This result indicates that the enzyme exists in a dimeric form. The far-UV CD spectrum of the protein exhibited a trough with a minimum [θ] value of –13 600 at 210 nm, which is accompanied by a shoulder with a [θ] value of –12 500 at 220 nm (Figure 3a). The near-UV CD spectrum of the protein exhibited two broad peaks with [θ] values of 700 at 260 and 275 nm and a trough with a [θ] value of –400 at Glycerol kinase from a hyperthermophilic archaeon Fig. 4. Temperature and pH dependence of Pk-GK activity. (a) The enzymatic activity was determined in 50 mM glycine–NaOH (pH 8.0) containing 2 mM CoCl2, 4.6 mM glycerol and 3.5 mM ATP at the temperatures indicated. The enzymatic activities relative to that determined at 80°C are shown. (b) The enzymatic activity was determined in 50 mM Tris–HCl (pH 6.5–8.0) (s) or 50 mM glycine–NaOH (pH 7.5–10.0) (d) containing 2 mM CoCl2, 4.6 mM glycerol and 3.5 mM ATP at 80°C. The enzymatic activities relative to that determined in 50 mM Tris–HCl (pH 8.0) are shown. In both cases, the amount of glycerol-3-phosphate was determined colorimetrically at 37°C as described previously (Bergmeyer et al., 1961). Table II. Effect of divalent cations on Pk-GK activity Compound Concentration (mM) None MgCl2 CaCl2 ZnCl2 CoCl2 FeCl2 CuSO4 MnCl2 NiCl2 SrCl2 BaCl2 – 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 Relative activity (%) 0.6 49.3 4.0 99.3 100 ,0.1 4.0 78.0 57.3 0.3 0.2 The enzymatic activity was determined in 50 mM glycine–NaOH (pH 8.0) containing various divalent cations, 4.6 mM glycerol and 3.5 mM ATP at 80°C. The amount of glycerol-3-phosphate was determined colorimetrically at 37°C as described previously (Bergmeyer et al., 1961). The relative activities were calculated by dividing the enzymatic activities determined in the presence of the various divalent cations by that determined in the presence of the Co21 ion. 297 nm (Figure 3b). The helical content of the protein was calculated as 32.5% by the method of Wu et al. (1981). Enzymatic activity Analysis of the temperature and pH dependence of the Pk-GK activity indicated that the optimal values for activity were 80°C (Figure 4a) and pH 8.0 (Figure 4b). The enzyme exhibited little activity in the absence of any divalent cation, indicating that the enzyme requires divalent cations for activity. To identify the preferred metal ion for activity, the enzymatic activity was determined at 80°C and pH 8.0 in the presence of salts of various divalent cations, such as CaCl2, ZnCl2, CoCl2, FeCl2, CuSO4, MnCl2, NiCl2, SrCl2 and BaCl2. The results are summarized in Table II. The enzyme exhibited the highest activity in the presence of Co21 or Zn21, which was nearly twice as high as that determined in the presence of Mg21. It also exhibited relatively high activity in the presence of Mn21 and Ni21, but exhibited little or poor activity in the presence of other metal ions. When the concentration of the metal ion was varied from 0.5 to 4 mM, the enzyme exhibited the highest activity in the presence of 1–2 mM metal ion, Fig. 5. Stability against irreversible heat inactivation. The solution containing Pk-GK at 2.7 µg/ml in 50 mM glycine–NaOH (pH 8.0) was incubated at 90°C (d) or 100°C (s). At appropriate intervals, an aliquot of the enzyme solution was withdrawn and analyzed for remaining activity. The remaining activity is shown as a function of incubation time. regardless of the species of metal ion. Thus, the enzyme exhibited the highest specific activity of 450 units/mg (turnover number 840 s–1) at 80°C and pH 8.0 in the presence of 2 mM CoCl2. Because the enzymatic activity determined in the presence of 0.5 M NaCl or KCl was similar to that determined in the absence of salt, the enzyme neither requires salt for activity nor is sensitive to salt. In addition to the divalent cation specificity, the nucleoside triphosphate specificity of the enzyme was analyzed by replacing ATP in the assay buffer with other nucleoside triphosphates, such as GTP, CTP, UTP and ITP. The enzymatic activities determined in the presence of 3.3 mM GTP, CTP, UTP and ITP were 3.4, 15.9, 12.1 and 14.2% of that determined in the presence of 3.5 mM ATP, respectively. These results indicate that the enzyme can use various nucleoside triphosphates as a phosphoryl group donor. Since it has been proposed that E.coli GK in a tetrameric form is non-competitively inhibited by FBP, but the enzyme in a dimeric form is not (de Riel and Paulus, 1978a; Liu et al., 1994), Pk-GK was also tested for inhibition by FBP. The enzyme retained 71% of the activity even in the presence of 20 mM FBP, suggesting that the Pk-GK activity is not sensitive to FBP. ATP and glycerol are the substrates of Pk-GK for the production of G-3-P. Therefore, the kinetic parameters of PkGK were determined in the presence of an excess amount of either of these substrates. For example, to determine the Km and kcat values for glycerol, the concentration of ATP was fixed at 2 mM and the concentration of glycerol was varied. The Km values for ATP and glycerol were determined as 15.4 and 111 µmol, respectively. The kcat value was determined as 940 s–1. Heat inactivation The stability of Pk-GK against irreversible heat inactivation was analyzed by incubating the enzyme solution (2.7 µg/ml) in 50 mM sodium phosphate buffer (pH 7.5) at 90 or 100°C. At appropriate intervals, an aliquot of the enzyme solution was withdrawn, chilled on ice and analyzed for remaining activity. As shown in Figure 5, the enzyme is fairly resistant to thermal denaturation at 90°C. The enzyme gradually lost activity at 100°C with a half-life of ~30 min. 1223 Y.Koga et al. Discussion Amino acid sequence of recombinant Pk-GK The NH2-terminal methionine residue is not post-translationally removed from recombinant Pk-GK, but is removed from E.coli GK (Pettigrew et al., 1988). It has been proposed that the second amino acid residue determines whether the NH2terminal methionine residue is removed or not (Tsunasawa et al., 1985). According to this proposal, the NH2-terminal methionine residue is removed when the second amino acid is Ala, Ser, Gly, Pro, Thr or Val, whereas it is not removed when the second amino acid is Arg, Asn, Asp, Gln, Glu, Ile, Leu or Lys. The observation that the NH2-terminal methionine residue is not removed from Pk-GK but is removed from E.coli GK supports this proposal, because the second amino acid is Glu for Pk-GK and Thr for E.coli GK. However, it remains to be determined whether natural Pk-GK has Met at its NH2-terminus as well. Enzymatic activity of Pk-GK The Km value of E.coli GK for glycerol has been reported to be 1.3 µM (Hayashi and Lin, 1967), 7 µM (Novotny et al., 1985) or 10 µM (Thorner and Paulus, 1973). Because the interaction of this enzyme with ATP shows negative cooperativity, two Km values of the enzyme for low and high concentrations of ATP have been reported to be 60 and 900 µM (Novotny et al., 1985) or 80–100 and 400–500 µM (Thorner and Paulus, 1973), respectively. The Km value of 4 mM, which was the only one reported for ATP by Hayashi and Lin (1967), may represent that for high concentrations of ATP. Interestingly, Pk-GK did not show negative cooperativity for the interaction with ATP. The Km value of this enzyme for ATP (15.4 µmol) was much lower than those of E.coli GK, suggesting that the binding affinity of ATP to Pk-GK is much higher than that to E.coli GK. In contrast, the Km value of this enzyme for glycerol (111 µmol) was much higher than that of E.coli GK, suggesting that the binding affinity of glycerol to Pk-GK is much lower than that to E.coli GK. A difference in the subunit structure or a slight difference in the conformation of the active site may be responsible for the alteration in the binding affinities of these substrates to the enzyme. The specific activity of Pk-GK (450 units/mg) was comparable to that of E.coli GK, which has been reported to be 1004 units/mg (Thorner and Paulus, 1973), indicating that the catalytic efficiencies of Pk-GK at 80°C and E.coli GK at mild temperature are similar to each other. Because local instability usually makes the active site conformationally flexible and thereby makes the enzyme functional (Meiering et al., 1992; Imoto et al., 1994; Shoichet et al., 1995; Kanaya et al., 1996), the conformational flexibility of the active site of Pk-GK gained at high temperature might be similar to that of E.coli GK gained at mild temperature. Pk-GK exhibited a very low enzymatic activity at 40°C (Figure 4a). The active site of Pk-GK may be conformationally too stable to display sufficient flexibility at mild temperature. E.coli GK can use only ATP as a phosphoryl group donor (Hayashi and Lin, 1967). In contrast, Pk-GK can utilize other trinucleoside phosphates than ATP as a phosphoryl group donor, although the enzymatic activities determined in their presence are considerably lower than that determined in the presence of ATP. The amino acid residues that form a hydrophobic pocket for the binding of adenine base in E.coli GK are fully conserved in Pk-GK, except for Ile313, which is replaced by Val306 in Pk-GK. Therefore, this amino acid 1224 Fig. 6. A tertiary model of Pk-GK monomer and the X-ray crystal structure of E.coli GK monomer. A model of Pk-GK monomer (top) and the crystal structure of E.coli GK monomer (PDB code 1GLB) (Hurley et al., 1993) (bottom) are shown. White pipe models are the backbone structures. The blue and red lines are the basic and acidic side-chains, respectively, which are involved in the ion pairs as indicated in Tables III and IV. The orange lines indicate the ion pairs that are only observed in Pk-GK or E.coli GK in Table III. The green lines indicate the ion-pair networks in Table IV. The bound ADP and glycerol at the active site are also shown by yellow spacefilling models. The side-chains of Val306 in Pk-GK (top) and Ile313 in E.coli GK (bottom) are also indicated by green space-filling models. N and C represent the NH2 and COOH termini of the molecules. The figure was drawn with the program Insight II (Molecular Simulations). substitution may be related to a relatively poor nucleoside triphosphate specificity of Pk-GK. In addition to Pk-GK, GKs from rat liver (Bublitz and Kennedy, 1954) and Candida mycoderma (Bergmeyer et al., 1961) have been shown to be active with nucleoside triphosphates other than ATP. However, the sequence information is not available for either of these enzymes. All GKs so far examined, except for GK from Thermus flavus, which exhibits activity in the presence of 1 mM EDTA (Huang et al., 1997), require divalent cations for activity. The preferred metal ion for E.coli GK activity is Mg21 and only the Mn21 ion can be substituted for it (Hayashi and Lin, 1967). On the other hand, Pk-GK is active with Co21, Zn21, Glycerol kinase from a hyperthermophilic archaeon Table III. Intramolecular ion pairs (ø4.0 Å)a Conservedb Pk-GK E.coli GK Arg17–Asp10 Arg83–Glu303 Arg106–Asp133 Arg107–Glu46 Lys116–Glu113 Lys142–Asp146 Arg152–Asp146 Arg152–Asp208 Arg154–Glu90 Arg154–Glu159 Arg188–Glu84 Arg188–Glu303 Arg218–Glu216 Arg218–Glu222 Arg317–Asp318 Arg378–Glu382 Arg432–Glu467 Lys487–Asp390 Arg22–Glu23 Arg22–Glu450 Arg67–Glu69 Lys122–Glu123 Lys124–Asp195 Lys124–Asp197 Lys139–Glu112 Lys154–Glu150 Arg169–Glu49 Lys190–Glu123 Lys191–Asp193 His318–Glu321 Lys328–Glu367 Arg382–Glu471 Lys395–Glu396 Arg419–Glu389 Lys420–Glu396 Arg423–Asp407 Arg449–Asp447 Arg459–Asp407 Lys464–Glu462 Arg470–Glu471 Lys482–Glu479 Arg483–Glu479 Arg33–Glu62 His114–Glu110 Arg125–Glu121 Arg156–Glu153 Arg211–Asp198 Arg361–Glu497 Arg389–Glu393 Arg407–Glu431 Arg482–Glu478 aThe ion pairs formed in Pk-GK are hypothetical, because the distance between two ionic residues was calculated according to a model of the three-dimensional structure of the enzyme. The conserved ion pairs and also the ion pairs that are formed only in Pk-GK or E.coli GK are listed. bThe ion pairs formed in E.coli GK are listed as representatives of those formed in common. All of these residues, except Arg154 and Glu90, which are replaced by Lys151 and Asp89, respectively, in Pk-GK, are conserved in Pk-GK. Table IV. Intramolecular ion-pair networks (ø4.0 Å) Pk-GK E.coli GK Arg82–Glu296–Arg183–Glu83 Arg83–Glu303–Arg188–Glu84 Glu112–Lys139–Asp143–Arg149–Asp203 Lys142–Asp146–Arg152–Asp208 Asp89–Lys151–Glu156 Glu90–Arg154–Glu159 Glu23–Arg22–Glu450 Lys122–Glu123–Lys190 Arg382–Glu471–Arg470 Lys395–Glu396–Lys420 Glu458–Arg423–Asp407–Arg459 Lys482–Glu479–Arg483 The ion-pair networks that are formed in Pk-GK or E.coli GK are listed. Fig. 7. Schematic representation for the tetrameric structure of E.coli GK. The subunit–subunit interactions in the tetrameric structure of E.coli GK are shown schematically. The subunits X, Y and Z are related to the subunit O by a 2-fold rotation about the X, Y and Z axes, respectively. The active site of each subunit is designated by a solid square. The dimeric structure (O–Y dimer) shaded is expected to represent that of Pk-GK. Mg21, Ni21 or Mn21. Because the proteins isolated from P. kodakaraensis KOD1 sometimes show a wide preference for metal ions (Rashid et al., 1997a,b), it may be a characteristic of the proteins from hyperthermophilic archaea. Pk-GK was shown to be rather resistant to the FBP inhibition. Because Pk-GK seems to act as a dimer, this result is consistent with the previous reports that FBP binds and inhibits only the tetrameric form of E.coli GK (de Riel and Paulus, 1978a; Liu et al., 1994). We have not analyzed whether the Pk-GK activity is inhibited by another allosteric inhibitor, EIIAGlc, because neither the gene encoding EIIAGlc nor the EIIAGlc protein from hyperthermophilic archaea is available. We could not identify the gene encoding a homologue of EIIAGlc in the DNA sequence of the Archaeoglobus fulgidus genome. Stability of Pk-GK Because Pk-GK seems to be conformationally fairly stable and functionally well adapted to a hyperthermal environment and because it shows a high sequence similarity to E.coli GK, Pk-GK and E.coli GK may be an ideal protein pair to analyze the mechanism by which proteins from hyperthermophiles become more stable than the mesophilic counterparts. Comparison of the crystal structures of glutamate dehydrogenase (Yip and Rice, 1995; Knapp et al., 1997) and indole-3-glycerol phosphate synthase (Hennig et al., 1995) from hyperthermophilic archaea with those from mesophilic counterparts have suggested that an increase in the number of ion pairs (ion-pair networks) and a decrease in the volume of cavities are responsible for the hyperstability of these hyperthermophilic proteins. It has been noted that proteins from the thermophilic sources have extra ion-pair networks on the protein surface (Perutz, 1978; Nakamura, 1996). To examine whether the number of ion pairs or ion-pair networks increased in Pk-GK as compared with that in E.coli GK, a model for the threedimensional structure of Pk-GK was constructed and compared with that of E.coli GK (Figure 6). Because Pk-GK shows a high sequence similarity with E.coli GK, its overall backbone structure is expected to be nearly identical with that of E.coli GK. In fact, the helical content of Pk-GK (32.5%) estimated from the far-UV CD spectrum was comparable to that of E.coli GK (36.3%) determined from crystallographic studies (Hurley et al., 1993). Apparently, the number of the charged residues involved in the formation of ion pairs and ion-pair networks at the surface of the Pk-GK molecule is higher than that in the E.coli GK molecule (Figure 6). All possible intramolecular ion pairs and ion-pair networks formed only in Pk-GK or E.coli GK, which were estimated by the method of Barlow and Thornton (1983), are listed in Tables III and IV, respectively. Twenty-four ion pairs and nine ion-pair networks are formed only in Pk-GK, whereas only nine ion pairs and three ion-pair networks are formed only in E.coli GK. Eighteen ion pairs are formed in common in these proteins. Thus, an increase in the number of ion pairs and ionpair networks may contribute to the increase in the protein stability of Pk-GK. Estimation of the number of the ion pairs formed in GKs from A. fulgidus, which is an another hyperthermophilic archeon and humans also supports this prediction. The number of the ion pairs in thermophilic GK from T.flavus (Huang et al., 1997) cannot be estimated, because the sequence information is not available for this protein. We are now trying to analyze the conformational stability of these proteins thermodynamically and the effect of the introduction or elimination of ion pairs (ion-pair networks) on the protein stability. 1225 Y.Koga et al. Subunit structure of Pk-GK Gel filtration experiments on agarose beads have previously shown that E.coli GK exists in an equilibrium between a dimer and a tetramer at concentrations lower than ~0.01 mg/ml, but exists in a tetramer at concentrations of more than ~0.1 mg/ml (de Riel and Paulus, 1978b). In contrast, Pk-GK exists as a dimer at a concentration of 0.5 mg/ml. These results suggest that interactions or forces which are required for the assembly of the tetramer from the dimer are reduced in the Pk-GK structure. The crystal structure of E.coli GK (Hurley et al., 1993), which contains tetramers of GK, is shown schematically in Figure 7. There are three types of intersubunit interactions in this GK structure, namely the interactions between O and Y subunits (O–Y interface), O and X subunits (O–X interface) and O and Z subunits (O–Z interface). Because the solvent-accessible surface area of each monomer is most extensively buried by the O–Y interaction and because the side chain of Arg369 (Arg362 in Pk-GK) of the Y subunit penetrates deeply into the O subunit, it has been proposed that the dimer–tetramer equilibrium involves dissociation of the tetramer into O–Y dimers (Hurley et al., 1993). Comparison of the amino acid sequences of Pk-GK and E.coli GK (Figure 1) indicates that the residues involved in the O–Y interface (residues 341–347 and 361–367 for E.coli GK) are well conserved. In contrast, the residues involved in the O–X interface (residues 49–68 for E.coli GK) and those involved in the O–Z interface (residues 321–333 for E.coli GK) are not well conserved. Therefore, it seems likely that Pk-GK assumes the O–Y dimer form. The observation that the mutation of Ala65 to Thr or Asp72 to Asn in E.coli GK shifts the dimer–tetramer equilibrium toward the dimer supports this prediction (Liu et al., 1994). In Pk-GK, these residues are replaced by Gln (Gln64) and Asn (Asn71), respectively. The hydropathy profiles of Pk-GK and E.coli GK were determined by the method of Kyte and Doolittle (1982) (data not shown). The striking similarity in these profiles suggests that hydrophobic interactions do not determine the subunit structures of the enzyme. Role of Pk-GK It has been shown that the glycerophosphate backbone of archaeal phospholipids is glycerol-1-phosphate (G-1-P), instead of G-3-P (Koga et al., 1993). This raises the question as to whether Pk-GK catalyzes the production of G-1-P as well. However, the enzymatic activity of Pk-GK determined by the calorimetric method, in which only the amount of G3-P can be determined, was nearly identical with that determined by the filter method, in which the amount of the 14Clabeled glyceromonophosphate (G-1-P and G-3-P) can be determined by measuring the radioactivity bound to a DEAEcellulose filter (data not shown). If Pk-GK produced both G1-P and G-3-P, the enzymatic activity determined by the calorimetric method would be lower than that determined by the filter method. Therefore, it is clear that a major product of Pk-GK is G-3-P. Because Pk-GK may not catalyze the production of G-1-P, Pk-GK may be important for energy production, instead of phospholipid biosynthesis. The glp regulon In the chromosome of E.coli, five different operons constitute a glycerol regulon (glp) (Larson et al., 1992). The glpK gene forms one of these operons (glpFK) with the glpF gene, which encodes a membrane protein that facilitates diffusion 1226 of glycerol across a cytoplasmic membrane. This glpFK operon has also been found in the chromosome of Bacillus subtilis (Holmberg and Rutberg, 1989). However, determination of the upstream sequence of Pk-glpK has revealed that the glpF gene is not located at the upstream region of the Pk-glpK gene (Y.Koga, personal communication). Instead, the gene encoding a protein, that is homologous to glycerophosphodiester phosphodiesterase, is located in this region. In the E.coli and B.subtilis chromosomes, the glpQ gene encoding glycerophosphodiester phosphodiesterase forms an operon (glpTQ) with the glpT gene encoding glycerol-3-phosphate permease (Larson et al., 1992; Nilsson et al., 1994). Therefore, the organization of the genes involved in glycerol metabolism of P.kodakaraensis KOD1 seems to be different from those of E.coli and B.subtilis. Analyses for the DNA sequences of the flanking regions of Pk-glpK are now in progress. Acknowledgements We thank Mr N.Yoshiaki of Toyobo Co., Ltd, for providing glycerol-3phosphate oxidase (GPO) and peroxidase (PO) for GK assay. This work was supported in part by Grant 08455382 from the Ministry of Education, Science and Culture of Japan. References Barlow,D.W. and Thornton,J.M. (1983) J. Mol. Biol., 168, 867–885. Bergmeyer,H.U., Holz,G., Kauder,E.M., Mollering,H. and Wieland,O. (1961) Biochem. Z., 333, 471–480. Bublitz,C. and Kennedy,E.P. (1954) J. Biol. Chem., 211, 951–961. Charrier,V., Buckley,E., Parsonage,D., Galinier,A., Darbon,E., Jaquinod,M., Forest,E., Deutscher,J. and Claiborne,A. (1997) J. Biol. Chem., 272, 14166–14174. de Riel,J.K. and Paulus,H. (1978a) Biochemistry, 17, 5134–5140. de Riel,J.K. and Paulus,H. (1978b) Biochemistry, 17, 5141–5146. Feese,M., Pettigrew,D.W., Meadow,N.D., Roseman,S. and Remington,S.J. (1994) Proc. Natl Acad. Sci. USA, 91, 3544–3548. Hayashi,S. and Lin,E.C.C. (1967) J. Biol. Chem., 242, 1030–1035. Hennig,M., Darimont,B., Sterner,R., Kirschner,K. and Jansonius,J.N. (1995) Structure, 3, 1295–1306. Holmberg,C. and Rutberg,B. (1989) FEMS Microbiol. Lett., 58, 151–156. Holmberg,C., Beijer,L., Rutberg,B. and Rutberg,L. (1990) J. Gen. Microbiol., 136, 2367–2375. Huang,H.-S., Yoshida,T., Meng,Y., Kabashima,T., Ito,K., Nishiya,Y., Kawamura,Y. and Yoshimoto,T. (1997) J. Ferment. Bioengng, 83, 328–332. Hurley,J.H., Faber,H,R., Worthylake,D., Meadow,N.D., Roseman,S., Pettigrew,D.W. and Remington,S.J. (1993) Science, 259, 673–677. Imanaka,T., Tanaka,T., Tsunekawa,H. and Aiba,S. (1981) J. Bacteriol., 270, 776–786. Imoto,T., Ueda,T., Tamura,T., Isakari,Y., Abe,Y., Inoue,M., Miki,T., Kawano,K. and Yamada,H. (1994) Protein Engng, 7, 743–748. Kanaya,S., Oobatake,M. and Liu,Y. (1996) J. Biol. Chem., 271, 32729–32736. Kim,D.-J., Morikawa,M., Takagi,M. and Imanaka,T (1995) J. Ferment. Bioengng, 79, 87–94. Knapp,S., de Vos,W.M., Rice,D. and Ladenstein,R. (1997) J. Mol. Biol., 267, 976–932. Koga,Y., Nishihara,M., Morii,H. and Akagawa-Matsushita,M. (1993) Microbiol. Rev., 57, 164–182. Kyte,J. and Doolittle,R.F. (1982) J. Mol. Biol., 157, 105–132. Laemmli,U.K. (1970) Nature, 227, 680–685. Larson,T.J., Cantwell,J.S. and van Loo-Bhattacharya,A.T. (1992) J. Biol. Chem., 267, 6114–6121. Lin,E.C.C. (1976) Annu. Rev. Microbiol., 30, 535–578. Liu,W.Z., Faber,R., Feese,M., Remington,S.J. and Pettigrew,D.W. (1994) Biochemistry, 33, 10120–10126. Meiering,E.M., Serrano,L. and Fersht,A.R. (1992) J. Mol. Biol., 225, 585–589. Morikami,K., Nakai,T., Kidera,A., Saito,M. and Nakamura,H. (1992) Comput. Chem., 16, 243–248. Morikawa,M., Izawa,Y., Rashid,N., Hoaki,T. and Imanaka,T. (1994) Appl. Environ. Microbiol., 60, 4559–4566. Nakamura,H. (1996) Q. Rev. Biophys., 29, 1–90. Nakamura,H., Katayanagi,K., Morikawa,K. and Ikehara,M. (1991) Nucleic Acids Res., 19, 1817–1823. Glycerol kinase from a hyperthermophilic archaeon Nilsson,R.-P., Beijer,L. and Rutberg,B. (1994) Microbiol., 140, 7223–730. Novotny,M.J., Frederickson,W.L., Waygood,E.B. and Saier,M.H.,Jr. (1985) J. Bacteriol., 162, 810–816. Pavlik,P., Simon,M., Schuster,T. and Ruis,H. (1993) Curr. Genet., 24, 21–25. Perutz,M.F. (1978) Science, 201, 1187–1191. Pettigrew,D.W., Ma,D.-P., Conrad,C.A. and Johnson,J.R. (1988) J. Biol. Chem., 263, 135–139. Pettigrew,D.W., Yu,G.-J. and Liu,Y. (1990) Biochemistry, 29, 8620–8627. Rashid,N., Morikawa,M. and Imanaka,T. (1997a) J. Ferment. Bioengng, 83, 412–418. Rashid,N., Morikawa,M., Nagahisa,K., Kanaya,S. and Imanaka,T. (1997b) Nucleic Acids Res., 25, 719–726. Saier,M.H.,Jr and Reizer,J. (1992) J. Bacteriol., 174, 1433–1438. Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) Molecular Cloning. A Laboratory Manual. 2nd edn. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) Proc. Natl Acad. Sci. USA, 74, 5463–5467. Sargent,C.A., Young,C., Marsh,S., Ferguson-Smith,M.A. and Affara,N.A. (1994) Hum. Mol. Genet., 3, 1317–1324. Schweizer,H.P., Jump,R. and Po,C. (1997) Microbiology, 143, 1287–1297. Shoichet,B.K., Baase,W. A., Kuroki,R. and Matthews,B.W. (1995) Proc. Natl Acad. Sci. USA, 92, 452–456. Tanimura,R., Kidera,A. and Nakamura,H. (1994) Protein Sci., 3, 2358–2365. Thorner,J.W. and Paulus,H. (1971) J. Biol. Chem., 246, 3885–3894. Thorner,J.W. and Paulus,H. (1973) J. Biol. Chem., 248, 3922–3932. Tsunasawa,S., Stewart,J.W. and Sherman,F.S. (1985) J. Biol. Chem., 260, 5382–5391. Weiner,S.J., Kollman,P.A., Nguyen,D.T. and Case,D. (1986) J. Comput. Chem., 7, 230–252. Woese,C.R. and Fox,G.E. (1977) Proc. Natl Acad. Sci. USA, 74, 5088–5099. Woese,C.R., Kandler,O. and Wheelis,M.L. (1990) Proc. Natl Acad. Sci. USA, 87, 4576–4579. Wu,C.-S.C., Ikeda,K. and Yang,J.T. (1981) Biochemistry, 20, 566–570. Yip,K.S.P. and Rice,D.W. (1995) Structure, 3, 1147–1158. Received July 7, 1998; revised August 14, 1998; accepted August 18, 1998 1227