Search

Crystal structures of 5-aminoimidazole ribonucleotide (AIR) synthetase, also known as PurM, from Thermus thermophilus (Tt) and Geobacillus kaustophilus (Gk) were determined. For TtPurM, the maximum resolution was 2.2 Å and the space group was P21212 with four dimers in an asymmetric unit. For GkPurM, the maximum resolution was 2.2 Å and the space group was P21212 with one monomer in asymmetric unit. The biological unit is dimer for both TtPurM and GkPurM and the dimer structures were similar to previously determined structures of PurM in general. For TtPurM, ∼50 residues at the amino terminal were disordered in the crystal structure whereas, for GkPurM, the corresponding region covered the ATP-binding site forming an α helix in part, suggesting that the N-terminal region of PurM changes its conformation upon binding of ligands. FGAM binding site was predicted by the docking simulation followed by the MD simulation based on the SO4 (2-) binding site found in the crystal structure of TtPurM., (© The Authors 2015. Published by Oxford University Press on behalf of the Japanese Biochemical Society. All rights reserved.)

The crystal structures of glycinamide ribonucleotide transformylases (PurNs) from Aquifex aeolicus (Aa), Geobacillus kaustophilus (Gk) and Symbiobacterium toebii (St), and of formyltetrahydrofolate hydrolase (PurU) from Thermus thermophilus (Tt) were determined. The monomer structures of the determined PurN and PurU were very similar to the known structure of PurN, but oligomeric states were different; AaPurN and StPurN formed dimers, GkPurN formed monomer and PurU formed tetramer in the crystals. PurU had a regulatory ACT domain in its N-terminal side. So far several structures of PurUs have been determined, yet, the mechanisms of the catalysis and the regulation of PurU have not been elucidated. We, therefore, modelled ligand-bound structures of PurN and PurU, and performed molecular dynamics simulations to elucidate the reaction mechanisms. The evolutionary relationship of the two enzymes is discussed based on the comparisons of the structures and the catalytic mechanisms.

The genome sequence of an extremely thermophilic bacterium, Thermus thermophilus HB8, revealed that TTHA0350 is a tandem-type universal stress protein (Usp) consisting of two Usp domains. Usp proteins, which are characterized by a conserved domain consisting of 130-160 amino acids, are inducibly expressed under a large number of stress conditions. The N-terminal domain of TTHA0350 contains a motif similar to the consensus ATP-binding one (G-2 x-G-9x-G-(S/T)), but the C-terminal one seems to lack the consensus motif. In order to determine its structural properties, we determined the crystal structures of TTHA0350 in the unliganded form and TTHA0350•2ATP at 2.50 and 1.70 Å resolution, respectively. This is the first structure determination of a Usp family protein in both unliganded and ATP-liganded forms. TTHA0350 is folded into a fan-shaped structure which is similar to that of tandem-type Usp protein Rv2623 from Mycobacterium tuberculosis. However, the dimer assembly with C2-symmetry in TTHA0350 is quite different from that with D2-symmetry in Rv2623. The X-ray structure showed that not only the N-terminal but also the C-terminal domain binds one ATP, although the ATP-binding motif could not be detected in the C-terminal domain. The loop interacting with ATP in the C-terminal domain is in a conformation quite different from that in the N-terminal domain.

Alkylation is a type of stress that is fatal to cells. However, cells have various responses to alkylation. Alkyltransferase-like (ATL) protein is a novel protein involved in the repair of alkylated DNA; however, its repair mechanism at the molecular level is unclear. DNA microarray analysis revealed that the upregulation of 71 genes because of treatment with an alkylating agent N-methyl-N'-nitro-N-nitrosoguanidine was related to the presence of TTHA1564, the ATL protein from Thermus thermophilus HB8. Affinity chromatography showed a direct interaction of purified TTHA1564 with purified RNA polymerase holoenzyme. The amino acid sequence of TTHA1564 is homologous to that of the C-terminal domain of Ada protein, which acts as a transcriptional activator. These results suggest that TTHA1564 might act as a transcriptional regulator. The results of DNA microarray analysis also implied that the alkylating agent induced oxidation stress in addition to alkylation stress.

The major damage to DNA caused by alkylating agents involves the formation of O6-methylguanine (O6-meG). Almost all species possess O6-methylguanine-DNA-methyltransferase (Ogt) to repair such damage. Ogt repairs O6-meG lesions in DNA by stoichiometric transfer of the methyl group to a cysteine residue in its active site (PCHR). Thermus thermophilus HB8 has an Ogt homologue, TTHA1564, but in this case an alanine residue replaces cysteine in the putative active site. To reveal the possible function of TTHA1564 in processing O6-meG-containing DNA, we characterized the biochemical properties of TTHA1564. No methyltransferase activity for synthetic O6-meG-containing DNA could be detected, indicating TTHA1564 is an alkyltransferase-like protein. Nevertheless, gel shift assays showed that TTHA1564 can bind to DNA containing O6-meG with higher affinity (9-fold) than normal (unmethylated) DNA. Experiments using a fluorescent oligonucleotide suggested that TTHA1564 recognizes O6-meG in DNA using the same mechanism as other Ogts. We then investigated whether TTHA1564 functions as a damage sensor. Pull-down assays identified 20 proteins, including a nucleotide excision repair protein UvrA, which interacts with TTHA1564. Interaction of TTHA1564 with UvrA was confirmed using a surface plasmon resonance assay. These results suggest the possible involvement of TTHA1564 in DNA repair pathways.

Pyruvate dehydrogenase (PDH), branched-chain 2-oxo acid dehydrogenase (BCDH) and 2-oxoglutarate dehydrogenase (OGDH) are multienzyme complexes that play crucial roles in several common metabolic pathways. These enzymes belong to a family of 2-oxo acid dehydrogenase complexes that contain multiple copies of three different components (E1, E2 and E3). For the Thermus thermophilus enzymes, depending on its substrate specificity (pyruvate, branched-chain 2-oxo acid or 2-oxoglutarate), each complex has distinctive E1 (E1p, E1b or E1o) and E2 (E2p, E2b or E2o) components and one of the two possible E3 components (E3b and E3o). (The suffixes, p, b and o identify their respective enzymes, PDH, BCDH and OGDH.) Our biochemical characterization demonstrates that only three specific E3*E2 complexes can form (E3b*E2p, E3b*E2b and E3o*E2o). X-ray analyses of complexes formed between the E3 components and the peripheral subunit-binding domains (PSBDs), derived from the corresponding E2-binding partners, reveal that E3b interacts with E2p and E2b in essentially the same manner as observed for Geobacillus stearothermophilus E3*E2p, whereas E3o interacts with E2o in a novel fashion. The buried intermolecular surfaces of the E3b*PSBDp/b and E3o*PSBDo complexes differ in size, shape and charge distribution and thus, these differences presumably confer the binding specificities for the complexes.

In bacterial RNA metabolism, mRNA degradation is an important process for gene expression. Recently, a novel ribonuclease (RNase), belonging to the beta-CASP family within the metallo-beta-lactamase superfamily, was identified as a functional homologue of RNase E, a major component for mRNA degradation in Escherichia coli. Here, we have determined the crystal structure of TTHA0252 from Thermus thermophilus HB8, which represents the first report of the tertiary structure of a beta-CASP family protein. TTHA0252 comprises two separate domains: a metallo-beta-lactamase domain and a "clamp" domain. The active site of the enzyme is located in a cleft between the two domains, which includes two zinc ions coordinated by seven conserved residues. Although this configuration is similar to those of other beta-lactamases, TTHA0252 has one conserved His residue characteristic of the beta-CASP family as a ligand. We also detected nuclease activity of TTHA0252 against rRNAs of T. thermophilus. Our results reveal structural and functional aspects of novel RNase E-like enzymes with a beta-CASP fold.

Small protein B, SmpB, is a tmRNA-specific binding protein essential for trans-translation. We examined the interaction between SmpB and tmRNA from Thermus thermophilus, using biochemical and NMR methods. Chemical footprinting analyses using full-length tmRNA demonstrated that the sites protected upon SmpB binding are located exclusively in the tRNA-like domain (TLD) of tmRNA. To clarify the SmpB binding sites, we constructed several segments derived from TLD. Optical biosensor interaction analyses and melting profile analyses with mutational studies showed that SmpB efficiently binds to only a 30-nt segment that forms a stem and loop, with the 5' and 3' extensions composed of the D-loop and variable-loop analogues. The conserved sequences, 16UCGA and 319GAC, in the extensions are responsible for the SmpB binding. These results agree with the those visualized by the cocrystal structure of TLD and SmpB from Aquifex aeolicus. In addition, NMR chemical shift mapping analyses, using the 30-nt segment and (15)N-labeled SmpB, revealed the characteristic RNA binding mode. The hydrogen bond pattern around beta2 changes, with the Gly in beta2, which acts as a hinge, showing the largest chemical shift change. It appears that SmpB undergoes structural changes indicating an induced fit upon binding to the specific region of TLD.

GTP cyclohydrolase I (GTPCH1) catalyzes the conversion of GTP to dihydroneopterin 3'-triphosphate. We found that an 8-oxoguanine derivative of GTP (8-oxo-GTP) strongly bound to GTPCH1 from Thermus thermophilus HB8 (tGTPCH1) as a competitive inhibitor. The affinity of 8-oxo-GTP was three orders of magnitude greater than that of GTP. These results suggest that 8-oxo-GTP is a transition state analogue of GTPCH1. We have solved the X-ray crystal structures of tGTPCH1 complexed with 8-oxo-GTP and 8-oxo-dGTP at 2.0 and 1.8 A resolution, respectively, as well as the free form of the enzyme at 2.2 A resolution. In the structure of tGTPCH1 complexed with 8-oxo-GTP or 8-oxo-dGTP, the oxygen atoms at O8 of the 8-oxoguanine groups, together with residues Cys108, His111 and Cys179, are coordinated to the zinc ion. The water molecule between Ndelta1 of His177 and N7 of 8-oxoguanine is conserved in both structures. These structural data are in accordance with one of the proposed transition states. Superimpositioning of the structures indicates the imidazole ring of His110 is rotated, implying concomitant proton transfer to the ribose ring O4'. Based on these structural data we propose a novel reaction mechanism for GTPCH1.

L-Cysteine is an important amino acid in terms of its industrial applications. The biosynthesis of L-cysteine in enteric bacteria is regulated through the feedback inhibition by L-cysteine of L-serine O-acetyltransferase (SAT), a key enzyme in L-cysteine biosynthesis. We recently found that L-cysteine is overproduced in Escherichia coli strains expressing a gene encoding feedback inhibition-insensitive SAT. Further improvements in L-cysteine production are expected by the use of SAT with high stability. We report here the sat1 gene encoding SAT of an extreme thermophile, Thermus thermophilus HB8. The sat1 gene was cloned and overexpressed in E. coli cells based on the genome sequence in T. thermophilus HB8. The predicted amino acid sequence consists of 295 amino acids and is homologous to other O-acetyltransferase members. In particular, the carboxyl-terminal region shares approximately 30% identities with SATs found in bacteria and plants, despite showing only about 15% identity in the overall sequence. Enzymatic analysis and an atomic absorption study of the purified recombinant proteins revealed that the enzyme is highly activated by Co(2+) or Ni(2+), and contains Zn(2+) and Fe(2+). These results indicate that the T. thermophilus SAT is a novel type of enzyme different from other members of this protein family.

The HD domain motif is found in a superfamily of proteins in bacteria, archaea and eukaryotes. A few of these proteins are known to have metal-dependant phosphohydrolase activity, but the others are functionally unknown. Here we have characterized an HD domain-containing protein, TT1383, from Thermus thermophilus HB8. This protein has sequence similarity to Escherichia coli dGTP triphosphohydrolase, however, no dGTP hydrolytic activity was detected. The hydrolytic activity of the protein was determined in the presence of more than two kinds of deoxyribonucleoside triphosphates (dNTPs), which were hydrolyzed to their respective deoxyribonucleosides and triphosphates, and was found to be strictly specific for dNTPs in the following order of relative activity: dCTP > dGTP > dTTP > dATP. Interestingly, this dNTP triphosphohydrolase (dNTPase) activity requires the presence of dATP or dTTP in the dNTP mixture. dADP, dTDP, dAMP, and dTMP, which themselves were not hydrolyzed, were nonetheless able to stimulate the hydrolysis of dCTP. These results suggest the existence of binding sites specific for dATP and dTTP as positive modulators, distinct from the dNTPase catalytic site. This is, to our knowledge, the first report of a non-specific dNTPase that is activated by dNTP itself.

The mismatch repair system (MMR) recognizes and corrects mismatched or unpaired bases caused mainly by DNA polymerase, and contributes to the fidelity of DNA replication in living cells. In Escherichia coli, the MutHLS system is known to function in MMR, and homologues of MutS and MutL are widely conserved in almost all organisms. However, the MutH endonuclease has not been found in the majority of organisms. Such organisms, including Thermus thermophilus HB8, often possess the so-called MutS2 protein, which is highly homologous to MutS but contains an extra C-terminal stretch. To elucidate the function of MutS2, we overexpressed and purified T. thermophilus MutS2 (ttMutS2). ttMutS2 demonstrated the ability to bind double-stranded (ds) DNA, but, unlike ttMutS, ttMutS2 showed no specificity for mismatched duplexes. ttMutS2 ATPase activity was also detected and was stimulated by dsDNA. Our results also showed that ttMutS2 incises dsDNA. ttMutS2 incises not only oligo dsDNA but also plasmid DNA, suggesting that ttMutS2 possesses an endonuclease activity. At low concentrations, the incision activity was not retained, but was promoted by T. thermophilus MutL.

Aspartate aminotransferase from an extremely thermophilic bacterium, Thermus thermophilus HB8 (ttAspAT), has been believed to be specific for an acidic substrate. However, stepwise introduction of mutations in the active-site residues finally changed its substrate specificity to that of a dual-substrate enzyme. The final mutant, [S15D, T17V, K109S, S292R] ttAspAT, is active toward both acidic and hydrophobic substrates. During the course of stepwise mutation, the activities toward acidic and hydrophobic substrates changed independently. The introduction of a mobile Arg292* residue into ttAspAT was the key step in the change to a "dual-substrate" enzyme. The substrate recognition mechanism of this thermostable "dual-substrate" enzyme was confirmed by X-ray crystallography. This work together with previous studies on various enzymes suggest that this unique "dual-substrate recognition" mechanism is a feature of not only aminotransferases but also other enzymes.

Do non-active-site residues participate in protein function in a more direct way than just by holding the static framework of the protein molecule? If so, how important are they? As a model to answer these questions, ATB17, which is a mutant of aspartate aminotransferase created by directed evolution, is an ideal system because it shows a 10(6)-fold increase in the catalytic efficiency for valine but most of its 17 mutated residues are non-active-site residues. To analyze the roles of the mutations in the altered function, we divided the mutations into four groups, namely, three clusters and the remainder, based on their locations in the three-dimensional structure. Mutants with various combinations of the clusters were constructed and analyzed, and the data were interpreted in the context of the structure-function relationship of this enzyme. Each cluster shows characteristic effects: for example, one cluster appears to enhance the catalytic efficiency by fixing the conformation of the enzyme to that of the substrate-bound form. The effects of the clusters are largely additive and independent of each other. The present results illustrate how a protein function is dramatically modified by the accumulation of many seemingly inert mutations of non-active-site residues.

We determined the crystal structure of the liganded form of alpha-aminotransferase from a hyperthermophile, Pyrococcus horikoshii. This hyperthermophilic enzyme did not show domain movement upon binding of an acidic substrate, glutamate, except for a small movement of the alpha-helix from Glu16 to Ala25. The omega-carboxyl group of the acidic substrate was recognized by Tyr70* without its side-chain movement, but not by positively charged Arg or Lys. Compared with the homologous enzymes from Thermus thermophilus HB8 and Escherichia coli, it was suggested that the more thermophilic the enzyme is, the smaller the domain movement is. This rule seems to be applicable to many other enzymes already reported.

MutM protein, which removes the oxidatively damaged DNA base product, 8-oxoguanine (GO), has been crystallized by means of a hanging-drop vapor-diffusion procedure using polyethyleneglycol monomethylether 2000 as a precipitant in 2-(cyclohexylamino) ethanesulfonic acid (CHES) buffer, pH 9.8. The diffraction data derived from oscillation photographs indicate that the crystals belong to the monoclinic system and space group P2(1). The crystals have unit-cell dimensions of a = 45.4 A, b = 62.0 A, c = 99.7 A, and beta = 90.8 degrees. Assuming that the asymmetric unit contains two molecules, the Vm value was calculated to be 2.35 A(3).Da(-1). The crystals diffracted X-rays to at least 2.1 A resolution and were suitable for high-resolution X-ray crystal structure determination.

In the nucleotide excision repair system, UvrB plays a central role in damage recognition and DNA incision by interacting with UvrA and UvrC. We have determined the crystal structure of Thermus thermophilus HB8 UvrB at 1.9 A resolution. UvrB comprises four domains, two of which have an alpha/beta structure resembling the core domains of DNA and RNA helicases. Additionally, UvrB has an alpha-helical domain and a domain consisting of antiparallel beta-sheets (beta-domain). The sequence similarity suggests that the beta-domain interacts with UvrA. Based on the distribution of the conserved regions and the structure of the PcrA-DNA complex, a model for the UvrB-DNA complex is proposed.

The whole-genome sequencing of an extreme thermophile, Thermus thermophilus, is now in progress. Like other genome projects, major concern is shifting from the sequence itself to post-sequencing research such as functional or structural genomics. Under such circumstances, the demand for convenient genetic-engineering tools is increasing. In this study we have increased the thermostability of a kanamycin-resistance gene product using strategies based on directed evolution in T. thermophilus to the upper limit of its growth temperature. The most thermostable mutant has 19 amino-acid substitutions, whereby the thermostability is increased by 20 degrees C, but the enzymatic activity is not significantly changed. Most of the mutated residues are located on the surface of the protein molecule, and, interestingly, five of the 19 substitutions are those to proline residues. The evolved kanamycin-resistance gene products could be used as selection markers at the optimum growth temperature of T. thermophilus. The development of such a convenient genetic-engineering tool would facilitate post-sequencing research on T. thermophilus.

The region containing the RNA polymerase alpha subunit (RNAPalpha) gene (rpoA) and the ribosomal protein genes of a thermophilic eubacterial strain, Thermus thermophilus (Tt) HB8, was cloned from a genomic DNA library by Southern hybridization. The gene order in this region is rpl36-rps13-rps11-rps4-rpoA-rpl17, which is identical to that in some other eubacteria. The rpoA gene encodes a 315 amino acid residue protein with a molecular weight of 35,013, the amino acid sequence showing 42% identity to that of Escherichia coli (Ec). From the results of comparison of the amino acid sequence and the predicted secondary structure of the C-terminal domain of Tt RNAPalpha (Tt alphaCTD) with those of Ec, the overall folding is expected to be similar. However, amino acid residues Asn268 and Cys269 in Ec alphaCTD, which are essential for its interaction with DNA or regulatory proteins, were replaced by His and Ser, respectively, in Tt alphaCTD. By means of a T7-based expression system in Ec cells, Tt RNAPalpha was overexpressed and purified. The high thermostability of Tt RNAPalpha was demonstrated by the CD spectra.

The single-stranded DNA-dependent ATPase activity of Escherichia coli RecA protein, especially its cooperativity for ATP, was investigated. To measure the ATPase activity in detail, the methods and reaction conditions for the ATPase assay were reexamined. Under conditions where RecA protein always showed a maximal rate of ATP hydrolysis, its poly(dT)-dependent ATPase activity was measured. At 25 degrees C, increasing the concentration of RecA protein from 0.3 to 1.0 microM increased the turnover number (kcat) from 0.16 to 0.19 s-1 and the Hill coefficient (nH) for ATP from 9.3 to 11.6. At 0.5 microM RecA protein, increasing the temperature from 25 to 37 degrees C increased kcat from 0.18 to 0.35 s-1 but decreased nH from 9.8 to 6.6. Interestingly, the ATPase activity of RecA protein measured in this study showed much higher cooperativity for ATP than those reported to date. Furthermore, the nH value of 11.6 for ATP obtained here was the highest of any ATPase reported so far. These results suggest that the binding of an ATP molecule to a RecA molecule within a nucleoprotein helical filament causes structural change of many other neighboring RecA molecules. This implies that ATP binding induces structural change of the whole nucleoprotein helical filament. Finally, we demonstrated that analysis of cooperativity is useful for revealing how a protein composed of many subunits functions as a whole.

The mechanism of recognition of hydrophobic substrates was investigated using Escherichia coli aspartate aminotransferase (AspAT), E. coli aromatic amino acid aminotransferase (AroAT), and their chimeric enzyme (DY18). Surprisingly, broad substrate specificity was observed in the reaction of aminotransferases with hydrophobic substrates. The catalytic efficiency increased with an increase in the side chain length of straight or branched-terminal aliphatic substrates. The straight-chain substrates catalysed with maximal efficiency were the 7-carbon substrate in the case of AspAT and the 8-carbon substrate for AroAT and DY18. Consecutive addition of single methylene groups to the substrate had a constant effect on the stabilization energy of the transition state relative to the unbound state. The dependency of binding energy on each methylene group is usually interpreted as indicating hydrophobicity of the active site. However, we observed that AroAT and DY18 had different dependencies although both enzymes have the same residues in the substrate-binding pocket. For substrates with more than 7 carbons, the aminotransferases did not strictly distinguish between substrates with straight and branched side chains. These results suggest that the recognition of manifold hydrophobic substrates of different shapes might require not only the hydrophobicity of the active site but also enzyme flexibility.

The aspartate aminotransferase gene (AspAT, EC 2.6.1.1) of an extremely thermophilic bacterium, Thermus thermophilus HB8, was cloned and sequenced, and its gene product was overproduced. The purified T. thermophilus AspAT was stable up to about 80 degrees C at neutral pH. T. thermophilus AspAT was strictly specific for acidic amino acid substrates, such as aspartate, glutamate, and the respective keto acids. The gene coding for T. thermophilus AspAT showed that it comprised 1,155 bp with a high G+C content (70 mol%), and encoded a 385-residue protein with a molecular weight of 42,050. The amino acid sequence of T. thermophilus AspAT deduced from its gene showed about 15, 46, and 29% homology with those from Escherichia coli, Bacillus sp. YM-2, and Sulfolobus solfataricus, respectively. When the amino acid sequence of T. thermophilus AspAT was compared with that of E. coli AspAT, the number of Cys was found to have decreased from 5 to 1, that of Asn from 23 to 9, that of Gln from 16 to 8, and that of Asp from 20 to 13, all of which are known to be relatively labile at high temperatures. Conversely, the number of Pro was increased from 15 to 25, Arg from 22 to 32, and Glu 27 to 37. As shown by the E. coli AspAT structure, there was a marked tendency for the extra prolyl residues to be located around the surface of the molecule. This was quite different from that in the case of RecA protein, which shows an increased number of prolyl residues in the interior of its molecule. Different strategies of different proteins as to prolyl contribution to thermostability have been suggested. Despite the high degree of conservation of active-site residues, Arg292 in E. coli AspAT, which interacts with the distal carboxylate of the substrate, was not found in T. thermophilus AspAT. Arg89 may complement the function of Arg292.

Interactions of Escherichia coli RecA protein with ATP and its analogues in the absence of DNA were studied by circular dichroic (CD) spectroscopy. The binding of RecA protein to ATP increased the CD band of ATP at around 260 nm. The positive CD band of the RecA protein-ATP complex suggested that the bound ATP was in the anti conformation, in accord with X-ray crystallographic data [Story, R.M. and Steitz, T.A. (1992) Nature 355, 374-376]. At pH 7.5 and at 25 degrees C the dissociation constant (Kd) and thermodynamic parameters for the binding of ATP to RecA protein were 18 microM (delta G = -6.5 kcal.mol-1), delta H = 0 kcal.mol-1, and delta S = 22 cal.mol-1.K-1. A non-hydrolyzable ATP analogue, adenosine 5'-O-(3-thiotriphosphate) (ATP gamma S), gave a spectral change similar to that of ATP. The Kd for this analogue, 22 microM, was very close to the Km of ATP. These results in the absence of single-stranded DNA were different from those obtained by kinetic analysis [Weinstock, G.M. et al. (1981) J. Biol. Chem. 256, 8850-8855], which indicated that the inhibition constant of ATP gamma S was much smaller than the Km of ATP in the presence of DNA. For other ATP analogues (dATP, ADP, and dADP), similar spectral changes were observed, and their Kd values ranged from 19 to 54 microM. UTP, dUTP, and TTP also gave CD spectral changes, but not AMP, GTP, dGTP, CTP, and dCTP.(ABSTRACT TRUNCATED AT 250 WORDS)

We determined the three-dimensional structures of aspartate aminotransferase (AspAT) from Escherichia coli and its complex with inhibitor (2-methyl-L-aspartate) at 1.8A resolution. This enzyme reversibly catalyzes the transamination reaction and is a dimer of two identical subunits. Each subunit has 396 amino acid residues and one pyridoxal 5'-phosphate as a cofactor, and is divided into two domains, one large and the other small. Upon binding of the inhibitor, the small domain rotates by 5 degrees toward the large domain to close the active site. This domain movement is caused mainly by small but important main-chain conformational changes in the residues located over the domain interface of the small domain. In chicken mitochondrial AspAT, the domain movement was larger, with a rotational angle of 13 degrees. By comparison of these two structures, the difference in the rotational angles was found to be caused by the larger opening of the domain in the open form of chicken mitochondrial AspAT. Although the overall structures of these two enzymes were almost identical, the surface area of the domain interface in the E. coli enzyme was larger than that in mitochondrial AspAT, suggesting that the structure of the domain interface is responsible for the degree of movement of the small domain.

Escherichia coli aspartate aminotransferase (AspAT) and E. coli aromatic amino acid aminotransferase (AroAT) have almost identical and high activities toward acidic amino acid substrates. AroAT also has high activity toward aromatic amino acid substrates. The two proteins have 44% amino acid sequence homology. In order to study the mechanism responsible for the different substrate specificities of these aminotransferases, chimeric enzymes of AspAT and AroAT were constructed using homologous recombination in E. coli cells. Five chimeric enzymes were obtained, even though the nucleotide sequence homology between the two parent enzymes was as low as about 50%. The yields of the legitimate chimeric genes were related to the lengths of the homologous region between the two parent genes. Homologous recombination occurred in the region where more than eight nucleotides out of ten were identical. The substrate specificity of the chimeric enzymes suggest that not only the amino acid residues in the active site but also those distant from the active site contribute to the substrate specificity of the parental aminotransferases.

The active-site lysine residue of thermostable aspartate aminotransferase, Lys-239, to which the cofactor, pyridoxal 5'-phosphate (PLP), is bound, has been converted to Cys by site-directed mutagenesis. The thiol group of Cys-239 was chemically aminoethylated with ethylenimine. Amino acid analysis of the modified enzyme showed that it contained about 1 mol of S-(2-aminoethyl)cysteine (SAEC) per mol subunit. The activity of the mutant enzyme (K239SAEC) was about 14% of that of the wild-type enzyme. No significant difference in thermostability was found between the wild-type and K239SAEC enzymes. The UV-visible spectrum of K239SAEC showed a peak (lambda max 380 nm), due to absorption by the cofactor, at a 20 nm longer wavelength than that of the wild-type enzyme. The circular dichroism band due to the bound cofactor of K239SAEC also shifted toward a 20 nm longer wavelength. We determined kinetic parameters (rate constants, kmax, and dissociation constants, Kd, for the substrates) for each half transamination catalyzed by the wild-type and K239SAEC mutant enzymes by the stopped-flow method. The kmax values for the mutant enzyme reactions were 2.6-24 times lower than those for the wild-type enzyme ones. The two enzymes showed similar Kd values for the same substrates except glutamate; the mutant enzyme showed higher affinity for glutamate than the wild-type enzyme.

The recA gene of a thermophilic eubacterial strain, Thermus thermophilus (T.th.) HB8, was cloned from a genomic DNA library by Southern hybridization using a gene-internal fragment amplified by the polymerase chain reaction (PCR) method as the probe. The gene encoded a 36 kDa polypeptide whose amino acid sequence showed 61% identity with that of the Escherichia coli RecA protein. Characteristic amino acid changes between the two RecA proteins were found. In the amino acid composition of the T.th. RecA protein, the number of Pro residues was increased, the number of Cys residues was decreased, and Lys residues were replaced by Arg, Asp by Glu, Thr by Val, and Ile by Val or Leu. These changes are supposed to stabilize the native protein conformation against heat denaturation. The amino acid residues in the nucleotide binding site of the protein and in the protein-protein interaction site responsible for the oligomer formation were well conserved. The T.th. recA gene has the ability to complement the ultraviolet light (UV) sensitivity of a E. coli recA deletion mutant. Thus, the thermophilic bacterium has a RecA protein whose function will be common to the E. coli RecA protein.

Aspartate aminotransferase (AspAT) [EC 2.6.1.1] of thermophilic methanogen was further characterized with the enzyme from Methanobacterium thermoautotrophicum strain FTF-INRA as well as M. thermoformicicum strain SF-4. AspAT of strain FTF-INRA was similar in the amino donor specificity to the enzyme of M. thermoformicicum strain SF-4, in that it was active on L-cysteine and L-cysteine sulfinate in addition to L-glutamate and L-aspartate. The enzymes gave similar absorption spectra having maxima at around 326 and 415 nm with no pH-dependent shift but were found to contain 1 mol of tightly bound pyridoxal 5'-phosphate (PLP) per subunit. Reconstitution of each apoenzyme with added PLP resulted in partial recovery of the original enzymatic activity, suggesting a significant conformational change of the active site region upon removal of the cofactor. Polyacrylamide gel electrophoresis (PAGE) and gel filtration analyses revealed a tetrameric structure (180 kDa) of identical subunits with a molecular mass of 43 kDa for each of these enzymes. Electric current was found to affect the interaction or affinity of each subunit, promoting dissociation of the native enzyme into the monomeric form. Alkaline treatment was effective only for dissociation of the enzyme from strain SF-4. They were distinguishable by the more rapid reassociation of the monomer to the native aggregated form in the enzyme of strain FTF-INRA.

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.

Tyr225 in the active site of Escherichia coli aspartate aminotransferase (AspAT) was replaced by phenylalanine or arginine by site-directed mutagenesis. X-ray crystallographic analysis of Y225F AspAT showed that the benzene ring of Phe225 was situated at the same position as the phenol ring of Tyr225 in wild-type AspAT. The mutations resulted in a great decrease in the rate of the transamination reaction, suggesting that Tyr225 is important for efficient catalysis. The kinetic analysis of half-transamination reactions of Y225F AspAT with four substrates (aspartate, glutamate, oxalacetate, and 2-oxoglutarate) and some analogues (2-methylaspartate, succinate, and glutarate) revealed a considerable increase in the affinities for all these compounds. In contrast, affinity for the amino acid substrates was decreased by mutation to arginine, but affinities for the keto acid substrates and the two dicarboxylates (succinate and glutarate) were increased. The electrostatic interaction between O(3') of the coenzyme [pyridoxal 5'-phosphate (PLP)] and the residue at position 225 affected the pKa value of the Schiff base, which is formed between the epsilon-amino group of Lys258 and the aldehyde group of PLP; based on the spectrophotometric titration the pKa values were determined to be 6.8 for wild-type AspAT, 8.5 for Y225F AspAT, and 6.1 for Y225R AspAT in the absence of substrate. The absorption spectra of the three AspATs were almost identical in the acidic pH region, but the spectrum of Y225F AspAT differed from that of wild-type or Y225R AspAT in the alkaline pH region.(ABSTRACT TRUNCATED AT 250 WORDS)

The structure of Escherichia coli aspartate aminotransferase complex with the inhibitor 2-methylaspartate, and that of the mutant enzyme in which an arginine was substituted for a lysine residue thereby forming a Schiff base with the coenzyme pyridoxal 5'-phosphate, were determined at 2.5 A resolution, by the molecular replacement method using the known structure of pig cytosolic aspartate aminotransferase. The enzyme catalyzes the reversible transamination between L-aspartate and alpha-ketoglutarate, and forms a dimeric structure of two identical subunits. Each subunit comprises two domains, a small and a large one. Although, in general, the overall and secondary structure of E. coli enzyme are similar to those of higher animals, some differences of enzymatic action between the enzyme from E. coli and those from higher animals could be explained on the basis of the X-ray structures and molecular mechanics calculation based on them.

The nucleotide sequence of the aspartate aminotransferase [EC 2.6.1.1] structural gene, aspC, of Escherichia coli K-12 was determined. The coding region of the aspC gene contained 1,188 nucleotide residues and encoded 396 amino acid residues. The amino acid sequence deduced from the nucleotide sequence agreed perfectly with that of the protein recently determined for the aspartate aminotransferase of E. coli B (Kondo, K., Wakabayashi, S., Yagi, T., & Kagamiyama, H. (1984) Biochem. Biophys. Res. Commun. 122, 62-67).

Pure recA protein was easily obtained from Escherichia coli harboring plasmid pTM-2 which carried the recA gene by two chromatographic steps on phosphocellulose and DEAE-cellulose. RecA protein was stable in the pH range of 6 to 9 at 25 degrees C. RecA protein was found to aggregate highly under these conditions. Lowering of the protein concentration, the presence of glycerol, and lowering of the pH in the pH stability region diminished the extent of aggregation. The spectroscopic properties of recA protein were measured in the presence of 10% (v/v) glycerol. RecA protein had an absorption maximum at 278 nm. The value of a1% 1cm at 278 nm was determined to be 5.7. The tryptophyl fluorescence spectrum excited at 295 nm had an emission maximum at 340 nm and the quantum efficiency of recA protein relative to N-acetyl-L-tryptophanamide was determined to be 0.65. The CD spectrum of recA protein had negative double maxima at 210 and 220 nm. The alpha-helical content was estimated to be about 40% from the CD spectrum in the region of 200 to 250 nm. All three cysteinyl residues of recA protein were reacted with 5,5'-dithiobis(2-nitrobenzoic acid), and recA protein was found to have neither intramolecular nor intermolecular disulfide bond. The reactivities of the SH groups were changed by the presence of ATP or ADP. The denaturation of recA protein by guanidine hydrochloride was studied by measuring CD at 220 nm and tryptophyl fluorescence. The denaturation curve obtained by CD measurement consisted of two stages, one of which lies between 0 and 1.8 M and the other above 1.8 M guanidine hydrochloride. On the other hand, the denaturation curve obtained by fluorescence measurement consisted of a single transition in the concentration range of about 1 to 2.3 M guanidine hydrochloride.

The pH difference absorption spectra of human lysozyme [EC 3.2.1.17] were measured. The difference spectra in the acidic region had a peak at 300 nm, as observed for hen and turkey lysozymes. The pH dependence curve of the extinction difference at 300 nm was well interpreted in terms of the pK values of the catalytic groups (3.4 for Asp 52 and 6.8 for Glu 35 at 0.1 ionic strength and 25 degrees C) determined from the pH dependence of the circular dichroism at 303.5 nm (Kuramitsu et al. (1974) J. Biochem. 76, 671--683) and the fluorescence excited at 305 nm (Kuramitsu et al. (1978) J. Biochem. 83, 159--170). The difference spectra of human lysozyme in the alkaline pH region were characteristic of tyrosyl ionization. The perturbation of tryptophyl residues, which had been observed for hen and turkey lysozymes (Kuramitsu & Hamaguchi (1979) J. Biochem. 85, 443--456), was not observed for human lysozyme. On the basis of the pH dependence curves of the extinction difference at 245 and and 295 nm, we roughly estimated the apparent pK values of the six tyrosyl residues as 9.2 9.2, 10.5, 10.9, 12.4, and 12.5. A time-dependent spectral change observed above pH 11 was not due to the exposure of buried tyrosyl residues on alkali denaturation but was due mainly to disulfide cleavage and exposure of buried tryptophyl residues.

ilvE gene of Escherichia coli was inserted into the region downstream of the tac promotor. As a result, the branched-chain amino acid aminotransferase was overproduced by about a hundred-fold in E. coli W3110. The overproduced aminotransferase was purified from cell extracts about 40-fold to homogeneity. Chemical and physicochemical analyses confirmed that it was a product of the ilvE gene. The enzyme existed in a hexamer with a subunit molecular weight of 34,000; the double trimer model of the enzyme presumed by the previous chemical cross-linking experiments (Lee-Peng, F.-C. et al. (1979) J. bacteriol. 139, 339-345) was supported by electron micrographs. The circular dichroic (CD) spectrum of branch-chain amino acid aminotransferase had double negative maxima at 210 and 220 nm. The alpha-helical content was estimated to be about 40% from the CD spectrum in the region of 200 to 250 nm. The absorption spectrum of the enzyme showed two peaks at 330 and 410 nm. There was no pH-dependent spectral shift. The CD spectrum of the coenzyme, pyridoxal 5'-phosphate, had negative peaks at 330 and 410 nm. These spectral properties of branched-chain amino acid aminotransferase were quite different from those of E. coli aspartate aminotransferase. Each subunit bound approximately 1 mol of pyridoxal 5'-phosphate. A lysyl residue, which forms a Schiff base with the aldehyde group of the pyridoxal 5'-phosphate, was identified in the primary structure of the enzyme.

Binding of N-acetyl-chitotetraose, 4-methylumbelliferyl chitotrioside ((GlcNAc)3-MeU), and 4-methylumbelliferyl chitotetraoside ((GlcNAc)4-MeU) to human lysozyme [EC 3.2.1.17] was studied by fluorescence measurement. Hydrolysis of (GlcNAc)3-MeU and (GlcNAc)4-MeU catalyzed by human lysozyme was studied by measuring the release of 4-methylumbelliferone fluorimetrically and the kinetic constants were determined in the pH range of 2 to 8 at 0.1 ionic strength and 42 degrees C. On the basis of binding and kinetic data, it was shown that (GlcNAc)3-MeU binds mainly to subsites A to D with the terminal MeU group bound to subsite D (nonproductive binding) and that (GlcNAc)4-MeU binds to subsites A to E (productive binding) and subsites A to D with the nonreducing sugar residue extending beyond subsite A (nonproductive binding). The pH dependences of the kinetic constants for hydrolysis of (GlcNAc)3-MeU and (GlcNAc)4-MeU were analyzed assuming that nonproductive binding occurs competitively, that an ionizable group in addition to the catalytic groups (Asp 52 and Glu 35) participates in the catalysis, and that the molecular species with ionized Asp 52 and protonated Glu 35 is active. Analyses of the kinetic constants for (GlcNAc)3-MeU and (GlcNAc)4-MeU both gave the same pK values of the catalytic groups (pK52 = 3.6(3) and pK35 = 6.6(8) at 0.1 ionic strength and 42 degrees C). These pK values were very close to the values determined previously by spectroscopic methods in our laboratory (Kuramitsu et al. (1974) J. Biochem. 76, 671-683; (1978) ibid. 83, 159-170; (1980) ibid. 87, 771-778).

The pH dependence of the binding constant of (GlcNAc)3 to Asp 52-esterified lysozyme was determined by the fluorescence technique. The pK values of Asp 101 in the modified lysozyme and its complex with (GlcNAc)3 were determined to be 4.5 and 3.6, respectively, at 25 degrees C and 0.1 ionic strength. This result is different from that obtained by Parsons and Raftery ((1972) Biochemistry 11, 1633--1638), who observed no pK shift of Asp 101. The macroscopic pK value of Asp 52 in intact lysozyme determined by them using the pH difference titration data of Asp 52-esterified lysozyme relative to intact lysozyme ((1972) Biochemistry 11, 1623--1629) was 4.5, which is higher by about one pH unit than the pK value determined by our group (Kuramitsu et al. (1974) J. Biochem. 76, 671--683; (1977) ibid. 82, 585--597; (1978) ibid. 83, 159--170. We found that their pH difference titration data in the absence and presence of saccharides can be consistently interpreted in terms of our pK values of Asp 52, Glu 35, and Asp 101, if we assume that the pK value of another ionizable group (probably Asp 48) is perturbed on esterification of Asp 52.

The interactions of the substrate analogues, GlcNAc, beta-methyl GlcNAc, (GlcNAc)2, and (GlcNAc)3, with turkey egg-white lysozyme [ED 3.2.1.17], in which the Asp 101 of hen lysozyme is replaced by Gly, were studied at various pH values by measuring changes in the circular dichroic (CD) band at 295 nm. Results were compared with those for hen egg-white lysozyme. The modes of binding of these substrate analogues to turkey lysozyme were very similar to those hen lysozyme except for the participation of Asp 101 in hen lysozyme. The ionization constants of the catalytic carboxyls, Glu 35 and Asp 52, in the turkey lysozyme-(GlcNAc)3 complex were determined by measuring the pH dependence of the CD band at 304 nm, which originates from Trp 108 near the catalytic carboxyls. The ionization behavior of the catalytic carboxyls of turkey lysozyme in the presence and absence of (GlcNAc)3 was essentially the same as that for hen lysozyme. The pH dependence of the binding constant of (GlcNAc)3 to hen lysozyme was compared with that to turkey lysozyme between pH 2 and 8. The pH dependence of the binding constant for (GlcNAc)3 to turkey lysozyme could be interpreted entirely in terms of perturbation of catalytic carboxyls. In the case of hen lysozyme, it was interpreted in terms of perturbation of the catalytic carboxyls and Asp 101 in the substrate-binding site. The pK values of Asp 101 in hen lysozyme and the hen lysozyme-(GLcNAc)3 complex were 4.5 and 3.4, respectively. The binding constants of (GlcNAc)3 to lysozyme molecules with different microscopic protonation forms, with respect to the catalytic carboxyls, were estimated. The binding constant of lysozyme, in which Asp 52 and Glu 35 are deprotonated, to (GlcNAc)3 was the smallest. The other three species had similar binding constant to (GlcNAc)3.

The difference absorption spectra of hen and turkey lysozymes in the alkaline pH region had three maxima at around 245, 292, and 300 nm and had no isosbestic points. The ratio of the extinction difference at 245 nm to that at 295 nm changed with pH. These spectral features are quite different from those observed when only tyrosyl residues are ionized, and it was impossible to determine precisely the pK values of the tyrosyl residues in lysozyme by spectrophotometric titration. A time-dependent spectral change was observed above about pH 12. This is not due to exposure of a buried tyrosyl residue on alkali denaturation. The disulfide bonds and the peptide bonds in the lysozyme molecule were cleaved by alkali above about pH 11. The intrinsic pK value of Tyr 23 of hen lysozyme was determined to be 10.24 (apparent pK 9.8) at 0.1 ionic strength and 25 degrees C from the CD titration data. Comparison of the CD titration of turkey lysozyme with that of hen lysozyme suggested that Tyr 3 and Tyr 23 in turkey lysozyme have apparent pK values of 11.9 and 9.8, respectively.

Cytosolic and mitochondrial isozymes of aspartate aminotransferase (L-aspartate:2-oxoglutarate aminotransferase [EC 2.6.1.1] ) were purified to homogeneity from rabbit liver. The rabbit liver isozymes were closely similar to the corresponding isozymes from other sources, including human heart, pig heart, chicken heart, and rat liver, in their molecular weights, absorption spectra, amino acid compositions, isoelectric points, and Michaelis constants for the substrates. The NH2-terminal amino acid sequences of rabbit liver isozymes were identified up to 30 residues, and showed some differences from those of the corresponding isozymes obtained from other animals so far studied.

The interactions of beta-methyl-GlcNAc, (GlcNAc)2, and (GlcNAc)3 with hen egg-white lysozyme [EC 3.2.1.17] in which Trp 62 is modified to kynurenine (Kyn 62-lysozyme) were studied by measuring the changes in the CD band and fluorescence due to the kynurenine at various pH values. The pH profiles of the binding constants of these saccharides to Kyn 62-lysozyme were very similar to those to intact lysozyme, although the binding constants were lower for the modified lysozyme than for intact lysozyme. The pK values of Asp 52, Glu 35, Asp 48, and Asp 101 in Kyn 62-lysozyme and in its complexes with beta-methyl-GlcNAc and with (GlcNAc)2 were in agreement with those of intact lysozyme and its complexes. The pK values of Asp 52 and Glu 35 in the modified lysozyme-(GlcNAc)3 complex were also in agreement with those of the complex with intact lysozyme, but the pK shift of Asp 101 was smaller for Kyn 62-lysozyme than for intact lysozyme. The significance of the decreased binding constants to Kyn 62-lysozyme is discussed. The pH dependence of the CD band due to the kynureinine in Kyn 62-lysozyme was interpreted in terms of the participation of the catalytic groups, Asp 52 (apparent pK 3.5) and Glu 35 (apparent pK 6.0), and the amino group of the kynurenine (apparent pK 0.75). This indicates that the ionization of the catalytic groups affects the state around Trp 62 and supports the previous proposal that there is a relation between the state around Trp 62 and the ionization of Glu 35 (Ikeda, K. & Hamaguchi, K. (1973) J. Biochem. 74, 221--230; (1975) ibid. 77, 1--16; Nakae et al. (1975) J. Biochem. 77, 993--1006). The intrinsic pK value of the amino group of the kynurenine at position 62 shifted from 2.1 to 0.8 on complexing with (GlcNAc)3, indicating between the kynurenine and the sugar residue at subsite B.

The binding constants of N-acetylglucosamine (G1cNAc) and its methyl alpha- and beta- glycosides to hen and turkey egg-white lysozymes [EC 3.2.1.17], in the latter of which Asp 101 is replaced by Gly, were determined at various pH values by measuring changes in the circular dichroic (DC) band at 295 nm. The binding of beta-methyl-G1cNAc to turkey and hen lysozymes perturbed the pK value of Glu 35 from 6.0 to 6.5, the pK value of Asp 52 from 3.5 to 3.9, and the pK value of Asp 66 from 1.3 to 0.7. In addition, perturbation of the pK value of Asp 101 from 4.4 to 4.0 was observed in the binding of this saccharide to hen lysozyme. The binding of alpha-methyl-GlcNAc to hen and turkey lysozymes perturbed the pK value of Glu 35 to the alkaline side by about 0.5 pH unit, the pK value of Asp 66 to the acidic side by about 0.5 pH unit, and the pK value (4.4) of an ionizable group to the acidic side by about 0.6 pH unit. The last ionizable group was tentatively assigned to Asp 48. The pK value of Asp 52 was not perturbed by the binding of this saccharide. The pH dependence curves for the binding of GlcNAc to hen and turkey lysozymes were very similar and it was suggested that Asp 48, in addition to Asp 66, Asp 52, and Glu 35, is perturbed by the binding of GlcNAc.

The binding constants of alpha- and beta-GlcNAc to hen and turkey lysozymes [EC 3.2.1.17] were determined at various pH's using the method proposed by Ikeda and Hamaguchi (1975) J. Biochem. 77, 1-16). The pH dependence of the binding of beta-GlcNAc to hen lysozyme was essentially the same as that for turkey lysozyme. The pH dependence curves of the binding constants of beta-GlcNAc to hen and turkey lysozymes were interpreted in terms of the participation of Glu 35 (pK 6.0), Asp 52 (pK 3.5), Asp 48 (pK 4.5), and Asp 66 (pK 1.5). The binding constants of alpha-GlcNAc to hen and turkey lysozymes were the same below pH 3.5 but were different above this pH. The main participant residues in the binding of alpha-GlcNAc were Glu 35, Asp 48, and Asp 66 for hen lysozyme and Glu 35 and Asp 66 for turkey lysozyme. The results obtained here were well explained by the following assumptions: (1) above about pH 4, alpha-GlcNAc binds to hen lysozyme in both alpha- and beta-modes, which correspond to the binding orientation of alpha-GlcNAc and that of beta-GlcNAc, respectively, as determined by X-ray crystallographic studies, but it binds predominantly in the beta-mode below about pH 4, (2) beta-GlcNAc binds to hen and turkey lysozymes predominantly in the beta-mode above about pH 4 and in both alpha- and beta-modes below pH 4, and (3) alpha-GlcNAc binds to turkey lysozyme predominantly in the beta-mode over the whole pH range studied.

The interactions of deoxy derivatives of GlcNAc, 6-deoxy-GlcNAc, and 3-deoxy-GlcNAc with hen egg-white lysozyme [EC 3.2.1.17] were studied at various pH's by measuring the changes in the circular dichroic (CD) band at 295 nm. It was shown that 6-deoxy-GlcNAc and 3-deoxy-GlcNAc bind at subsite C of lysozyme and compete with GlcNAc. The pH dependence of the binding constant of 6-deoxy-GlcNAc was the same as that of GlcNAc. On the other hand, the binding constants of 3-deoxy-GlcNAc were 3--10 times smaller than those of GlcNAc in the pH range from 3 to 9. X-ray crystallographic studies show that O(6) and O(3) of GlcNAc at subsite C are hydrogen-bonded to the indole NH's of Trp 62 and Trp 63, respectively, but the above results indicate that Trp 63, not Trp 62, is important for the interaction of GlcNAc with lysozyme.

The interaction of N-acetyl-chitotriose ((GlcNAc)3) with human lysozyme [EC 3.2.1.17] was studied at various pH values by measuring changes in the circular dichroic (CD) band at 294 or 255 nm and the data were compared with the results for hen and turkey lysozymes reported previously (Kuramitsu et al. (1974) J. Biochem.76, 671-683; Kuramitsu et al. (1975) J. Biochem. 77, 291-301). The pH dependence of the binding constant of (GlcNAc)3 to human lysozyme was different from those for hen and turkey lysozymes. The catalytic carboxyls of human lysozyme, Asp 52 and Glu 35, were not perturbed on binding of (GlcNAc)3. This is consistent with the previous findings that the macroscopic pK values of Asp 52 and Glu 35 of human lysozyme are 3.4 and 6.8 at 0.1 ionic strength and 25 degrees and were unchanged on complexing with (GlcNAc)3. An ionizable group with pK 4.5, which participates in the binding of (GlcNAc)3 to hen lysozyme and was assigned as Asp 101, did not participate in the binding of the saccharide to human lysozyme. Between pH 9 and 11, the binding constants of (GlcNAc)3 to hen lysozyme remained unchanged, whereas perturbation of an ionizable group with pK 10.5 to 10.0 was observed for human lysozyme. This group may be Tyr 62 in the active-site cleft. The binding constants of (GlcNAc)3 to human lysozyme molecules having different microscopic protonation forms, with respect to the catalytic carboxyls, were estimated using the binding constants obtained in the present experiments and the microscopic ionization constants of the catalytic carboxyls obtained previously. All four species of human lysozyme had similar binding constants to (GlcNAc)3. This result is different from those for hen and turkey lysozymes.

Purified bovine heart two-band cytochrome c1 subcomplex was dissociated by treatment with p-chloromercuribenzoic acid (pCMB) into its heme subunit and a colorless subunit called hinge protein, which is essential for the formation of cytochrome c1-c complex. The subcomplex was found by titration to react with 4 mol of pCMB per mol of cytochrome c1. The contents of mercury of the dissociated heme subunit and the hinge protein were 3 and 1 mol per mol of polypeptide, respectively. These results, together with the sequence analysis, indicated that the three cysteine residues in cytochrome c1 heme subunit not involved in heme-binding existed in free thiol form. One of the five cysteine residues in the hinge protein was in free form and four in two disulfide bonds. The dissociated hinge protein was digested with staphylococcal protease and the cysteine-containing peptides were separated by reversed-phase high-performance liquid chromatography (HPLC). The content of mercury and the result of performic acid oxidation of cystine peptides revealed that Cys-30 existed in free thiol form and two disulfide bridges were formed between Cys-24 and Cys-68 and between Cys-40 and Cys-54. The conformation of the hinge protein was predicted to be composed largely of either two-alpha-helical or four-alpha-helical conformation with the amino (N)-terminal 20 residues being in a random structure.

The difference spectra of hen and turkey egg-white lysozymes [EC 3.2.1.17] produced by acidification were measured. The difference spectra of both lysozymes had peaks at 295 and 301 nm which are characteristic of tryptophyl residues. The pH dependence curves of the extinction differences (delta eplision) at 301 nm and 295 nm for hen lysozyme were identical with the corresponding curves for turkey lysozyme. The pH dependence of delta eplision at 301 nm was analyzed assuming that the extinction at 301 nm is due to Trp 108 only, which interacts with the catalytic carboxyls, Glu 35 and Asp 52. The macroscopic pK values of Glu 35 and Asp 52 in both lysozymes thus determined were 6.0 and 3.3, respectively. These values were in excellent agreement with those determined by measuring the pH dependence of the circular dichroic band at 305 nm (Kuramitsu et al. (1974) J. Biochem, 76, 671-683; (1975) ibid. 77, 291-301). The pH dependence of delta eplision at 295 nm could not be completely explained in terms of the electrostatic effects of the catalytic groups on Trp 108.