Crystal structure of the halotolerant c-glutamyltranspeptidase from Bacillus subtilis in complex with glutamate reveals a unique architecture of the solvent-exposed catalytic pocket Kei Wada 1 , Machiko Irie 1 , Hideyuki Suzuki 2 and Keiichi Fukuyama 1 1 Department of Biological Sciences, Graduate School of Science, Osaka University, Japan 2 Division of Applied Biology, Graduate School of Science and Technology, Kyoto Institute of Technology, Japan Introduction c-Glutamyltranspeptidase (GGT; EC 2.3.2.2), an enzyme found in bacteria, yeast, plants, and mammals, is involved in the degradation of c-glutamyl com- pounds such as glutathione (GSH; c-glutamyl-cyste- inyl-glycine) [1], a process that is critical to maint enance of the cellular redox state [1–3]. GGT catalyzes the initial step of the degradation of extracellular GSH into its constituent amino acids, which are then Keywords electrostatic surface potential; glutathione; Ntn-hydrolase family; salt-tolerant; c-glutamyltranspeptidase Correspondence K. Fukuyama, Department of Biological Sciences, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan Fax: +81 6 6850 5425 Tel: +81 6 6850 5422 E-mail: fukuyama@bio.sci.osaka-u.ac.jp (Received 3 October 2009, revised 5 November 2009, accepted 8 December 2009) doi:10.1111/j.1742-4658.2009.07543.x c-Glutamyltranspeptidase (GGT; EC 2.3.2.2), an enzyme found in organ- isms from bacteria to mammals and plants, plays a central role in glutathi- one metabolism. Structural studies of GGTs from Escherichia coli and Helicobacter pylori have revealed detailed molecular mechanisms of catalysis and maturation. In these two GGTs, highly conserved residues form the catalytic pockets, conferring the ability of the loop segment to shield the bound c-glutamyl moiety from the solvent. Here, we have exam- ined the Bacillus subtilis GGT, which apparently lacks the amino acids corresponding to the lid-loop that are present in mammalian and plant GGTs as well as in most bacterial GGTs. Another remarkable feature of B. subtilis GGT is its salt tolerance; it retains 86% of its activity even in 3 m NaCl. To better understand these characteristics of B. subtilis GGT, we determined its crystal structure in complex with glutamate, a product of the enzymatic reaction, at 1.95 A ˚ resolution. This structure revealed that, unlike the E. coli and H. pylori GGTs, the catalytic pocket of B. subtilis GGT has no segment that covers the bound glutamate; consequently, the glutamate is exposed to solvent. Furthermore, calculation of the electro- static potential showed that strong acidic patches were distributed on the surface of the B. subtilis GGT, even under high-salt conditions, and this may allow the protein to remain in the hydrated state and avoid self-aggre- gation. The structural findings presented here have implications for the molecular mechanism of GGT. Structured digital abstract l MINT-7383558: GGT (uniprotkb:P54422) and GGT (uniprotkb:P54422) bind (MI:0407)by X-ray crystallography ( MI:0114) Abbreviations GGT, c-glutamyltranspeptidase; GSH, glutathione; L-subunit, large subunit; S-subunit, small subunit. 1000 FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS transported into the cell and reused as a cysteine source [1,3–5]. The localization of GGT differs by organism: in bacteria, GGT is expressed in the periplasmic space or secreted into the extracellular environment [1]; in mammalian cells, it is bound to the external surface of the plasma membrane [1,5]; and in plants, it is localized to the apoplast and the vacuole [6]. The mature GGT is a heterodimer comprising one large subunit (L-subunit;  40 kDa) and one small subunit (S-subunit;  20 kDa), generated by post- translational autocatalytic cleavage of the inactive pre- cursor protein ( 60 kDa) [7]. Crystallographic studies of GGTs from Escherichia coli and Helicobacter pylori have revealed the detailed molecular mechanisms of catalysis and maturation [8–10]. The side chain of Thr391 of the E. coli GGT precursor protein acts as the nucleophile for the cleavage, by which it becomes the new N-terminal residue of the S-subunit, and in turn acts as the nucleophile for the enzymatic reaction. Following cleavage, the C-terminal segment of the L-subunit (I378–Q390) moves away from the threo- nine, thereby forming the c-glutamyl moiety-binding pocket and concomitantly allowing the flexible loop (residues 438–449 of the E. coli GGT) to cover the pocket. Interestingly, the loop has been shown to shield the catalytic pocket from the solvent when the pocket is occupied by a substrate or inhibitor, whereas the loop is disordered when the pocket is empty [10– 12]. Hence, we assumed that the loop, which is called the ‘lid-loop’, is involved in recruiting the substrate by changing its conformation according to the conditions of the catalytic pocket, and is therefore necessary for the GGT reaction. Indeed, mutational analysis of H. pylori GGT demonstrated that substitution of a res- idue on the loop (Y433A) significantly diminished its catalytic activity [13]. We were interested in Bacillus subtilis GGT because it lacks the sequences corresponding to the lid-loop, and instead has extra residues at the C-terminus of the L-subunit, whereas the catalytic threonine and the resi- dues involved in substrate binding are mostly con- served (Fig. 1). A phylogenetic sequence alignment of confirmed and putative GGTs using the Microbial Genome Database [14] showed that, among the enzymes from 305 species, several bacterial GGTs, including those of Bacillus, Oceanobacillus and Staphy- lococcus, apparently lack the sequences corresponding to the lid-loop. Because the structural rearrangements occur at the active site pocket upon E. coli GGT mat- uration [10], the inserted residues at the L-subunit C-terminus in B. subtilis GGT could shield the active site pocket by occupying the position corresponding to the lid-loop. To better understand the significance of the absence of the lid-loop sequence and the presence of the extra residues at the L-subunit C-terminus, we embarked on a crystallographic analysis of B. subtilis GGT in complex with glutamate. Another goal of this analysis was to determine the structural factors underlying the salt tolerance of B. subtilis GGT [15,16]. In the presence of 3 m NaCl, B. subtilis GGT retained 86% of its hydrolytic activity, whereas E. coli GGT lost 90% of its activity, in com- parison with NaCl-free conditions. B. subtilis GGT has therefore attracted attention because of its possible application to the fermentation of food in high salt concentrations, e.g. with soy sauce and miso (fer- mented soybeans), the traditional Japanese seasonings, in which the desired taste depends mainly on the amount of glutamic acid present. Elucidation of the structural factors related to the salt tolerance of B. subtilis GGT may help in the development of bet- ter-engineered GGT. Here, we report the crystal structure of B. subtilis GGT in complex with glutamate, a product of the enzymatic reaction, at 1.95 A ˚ resolution, revealing the unique structure of the catalytic pocket. We also pres- ent a structural comparison between the salt-tolerant B. subtilis GGT and the nonhalophilic E. coli GGT. Results and Discussion Overall structure of B. subtilis GGT N-terminal His-tagged GGT from B. subtilis lacking the signal peptide composed of the first 35 residues was produced in E. coli. The recombinant protein was subjected to autocatalytic processing to yield a stable heterodimeric enzyme comprising an L-subunit (resi- dues 36–402) and an S-subunit (residues 403–587). The structure of this GGT in complex with glutamate, a product of the enzymatic reaction, was refined at 1.95 A ˚ resolution to R and R free values of 0.208 and 0.262, respectively. The asymmetric unit contains two GGT molecules, which bind one glutamate each at identical sites. Although the electron density for GGT was mostly of high quality and continuous, the densi- ties for the N-terminal His-tag segment and resi- dues 396–402, corresponding to the C-terminus of the L-subunit, were poorly defined; therefore, these resi- dues were not included in the model. B. subtilis GGT has a four-layer sandwich (a ⁄ b ⁄ b⁄ a) core structure, comprising two central b-sheets and surrounding a-helices, similar to the structure of the Ntn hydrolase superfamily [17,18]. Recently, the crystal structure of ligand free-GGT from B. subtilis was determined by Sharath et al. K. Wada et al. Structure of B. subtilis c-glutamyltranspeptidase FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS 1001 (Protein Data Bank ID: 2V36). To determine the struc- tural change caused by the binding of glutamate, the structures of ligand-free GGT and glutamate-bound GGT were superimposed. No significant structural change was observed; the rmsd for the Ca atom was 0.40 A ˚ . In the glutamate-bound GGT, one additional Fig. 1. Multiple sequence alignment of GGTs from several representative organisms. The sequence numbering is shown for B. subtilis GGT. Identical residues are highlighted in light green, and similar residues are boxed in blue. The residues of the catalytic nucleophile are high- lighted in orange. The residues that participate in hydrogen bonding with glutamate are highlighted in yellow. The secondary structural ele- ments of B. subtilis GGT are shown above the alignment, and 3 10 helices are labeled g. The figure was prepared with CLUSTALW [35] and ESPRIPT [36]. B. subtilis 168, NCBI accession no. NP_389723; E. coli K-12, NP_417904; H. pylori, NP_207909; Human, Homo sapiens, NM_005265; Pig, Sus scrofa, NM_214030; Rat, Rattus norvegicus, NM_053840. Structure of B. subtilis c-glutamyltranspeptidase K. Wada et al. 1002 FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS residue in the C-terminal region of the L-subunit was visible as compared with ligand-free GGT. One glutamate was bound to each of the deeply grooved catalytic pockets in the asymmetric unit, and the glutamate-binding modes are identical to each other (Fig. 2A). The a-carboxyl and a-amino groups of the bound glutamate are at the bottom of the pocket, and are held in this position by extensive hydrogen bonds and salt bridges in a similar manner as the c-glutamyl intermediate in the E. coli GGT molecule [8] and the glutamate complex in the H. pylori GGT molecule [13]. The carboxyl group is bonded with Arg113 Ng, Ser464 Oc, and Ser465 N, and the a-amino group with Glu442 Oe, Glu423 Oe, and Asp445 Od. The glutamate carbonyl oxygen at the e-position is hydrogen-bonded with the two main chain amino groups of Gly485 and Gly486, which are assumed to form the oxyanion hole. As compared with E. coli GGT (Fig. 2B) [8], all of the interactions with glutamate are identical, except that Glu423 and Glu442 in B. subtilis GGT are replaced by asparagine (Asn411) and glutamine (Gln430), respectively, in E. coli GGT. Unique structure of the catalytic pocket of B. subtilis GGT Comparison of the B. subtilis GGT structure with the previously reported E. coli and H. pylori GGT struc- tures [8,9] revealed a unique structural feature of B. sub- tilis GGT (Fig. 3). Unlike in the E. coli and H. pylori GGTs, in which the lid-loop covers the catalytic pocket when the pocket is occupied by the substrate or one of its analogs, in the B. subtilis GGT there was no ordered segment covering the bound glutamate in the catalytic pocket (Fig. 3A). Consequently, the bound glutamate is exposed to solvent, whereas the glutamates in both E. coli and H. pylori GGTs are buried as if in a cave (Fig. 3B,C). The segment of B. subtilis GGT that corre- sponds to the lid-loops of the E. coli and H. pylori GGTs (439–448 and 428–437 residues) is cut short. B. subtilis GGT has additional residues not present in most other GGTs at the C-terminal region of the L-subunit. In E. coli GGT, upon autocatalytic cleavage of the peptide bond on the N-terminal side of Thr391, the C-terminal segment of the newly produced L-sub- unit flips away, with the result that the C-terminus of the L-subunit and the N-terminus of the S-subunit become quite distant from one another (> 35 A ˚ ) [10]. The present crystallographic analysis has revealed that the L-subunit C-terminal segment in the mature form of B. subtilis GGT is located close to the catalytic pocket, although the seven C-terminal residues (396– 402) are invisible. The location of the L-subunit C-ter- minal segment in B. subtilis GGT is similar to that of the E. coli GGT precursor protein but distinct from that of the mature E. coli GGT. Hence, unlike other GGTs with known structures, it is assumed that B. subtilis GGT undergoes no significant structural change during maturation. In B. subtilis GGT, the seg- ment consisting of residues 391–395 undergoes no spe- cific interaction with neighboring residues. In summary, B. subtilis GGT does not have the lid- loop motif, and the C-terminal segment of the newly A B Fig. 2. Glutamate binding in the catalytic pocket of GGT. (A) Elec- tron density map for the bound glutamate in B. subtilis GGT. An omit F o – F c map for glutamate contoured at 2.0r (orange) is overlaid on the stick models of GGT and the bound glutamate. (B) The glutamate-binding mode in E. coli GGT. The bound glutamate and catalytic threonines are shown in orange and cyan, respec- tively. Dashed lines indicate hydrogen bonds. K. Wada et al. Structure of B. subtilis c-glutamyltranspeptidase FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS 1003 produced L-subunit appears to be changed little after autocatalytic processing. Moreover, additional residues at the L-subunit C-terminus are not involved in shield- ing the active site pocket from the solvent; therefore, the substrate ⁄ product is exposed to solvent when bound to the catalytic pocket. As described above, for E. coli GGT it has been well established that the cata- lytic reaction proceeds in the active site pocket, which is shielded from solvent by the lid-loop, as well as by release of the C-terminal segment from the active site pocket upon autocatalytic processing. The role of the lid-loop is made possible by its flexible nature, which allows it to adopt open or closed conformations. The structure of B. subtilis GGT shows that neither the lid-loop nor the alternative segment that covers the active site pocket is present, prompting questions about the role and significance of the lid-loop in GGT catalysis. During the preparation of this article, the crystal structures of Bacillus anthracis CapD, a GGT-related enzyme, in the absence and presence of a glutamate dipeptide were reported [19]. B. anthracis CapD cata- lyzes the cleavage of the c-glutamyl bond but, unlike GGTs, transfers the poly-c-d-glutamic acid to the pep- tidoglycan cell wall, and is therefore involved in link- ing the capsule of poly-c-d-glutamic acid to the bacterial envelope [20]. When the structure of CapD in complex with di-a-l-glutamate peptide, a nonhydrolyz- able analog of the substrate, is compared with that of the B. subtilis GGT–glutamate complex, it is apparent that CapD is a member of the Ntn hydrolase superfamily, like GGT. Differences in structural char- acteristics between the two enzymes can be observed at the active sites. CapD lacks the lid-loop, as seen in B. subtilis GGT, and therefore the ligand bound to CapD is exposed to solvent. A notable difference between the two enzymes can be seen in the manner of AB C Fig. 3. Comparison of the structures of GGTs from (A) B. subtilis, (B) E. coli and (C) H. pylori . The structure of each GGT is indicated in the upper panel. The catalytic pocket is shown in the square box, and the molecular surface corresponding to this region is shown in the lower panel. The bound glutamate molecule is depicted in each GGT with a space-filling model. The region corresponding to the unique C-terminal segment of B. subtilis GGT and the lid-loops of E. coli and H. pylori GGTs are shown as dark red and light green sticks, respectively. The characters N and C indicate the N-terminus and C-terminus, respectively. This figure was prepared with PYMOL [37]. Fig. 4. Comparison of the binding modes of the ligands. The struc- tures of B. subtilis GGT and B. anthracis CapD (Protein Data Bank ID: 3G9K) are shown in gray and brown, respectively. Stick models of the glutamate (orange) in GGT and the di-a-glutamate (purple) in CapD are shown. Hydrogen bonds in GGT and in CapD are shown in blue and orange, respectively. Structure of B. subtilis c-glutamyltranspeptidase K. Wada et al. 1004 FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS ligand binding to the enzyme (Fig. 4); there is little correspondence in the amino acids that are involved in the recognition of the functional groups of the ligands. Moreover, in GGT the a-carboxyl group of the gluta- mate protrudes into the groove, whereas in CapD the corresponding c-carboxyl group of the dipeptide ana- log is skewed towards the surface. This difference may reflect the different sizes of the physiological sub- strates; that is, the substrate of GGT is a small GSH, whereas that of CapD is a huge polymer. Structural basis for the salt tolerance B. subtilus GGT is so salt-tolerant that it retains most of its catalytic activity even in 3 m NaCl solution [16]. It has been reported that that the protein’s acidic sur- face enhances its stability by increasing solvation through increased water-binding capacity [21–23]; therefore, we analyzed its surface potential. The water- binding capacities of glutamate and aspartate have been reported to be 7.5 and 6.0 molecules per amino acid, respectively, whereas those for asparagine, serine and threonine have been estimated to be 2.0 molecules per amino acid [24]. B. subtilis and E. coli GGTs both have a negatively charged surface (Fig. 5A,B), whereas H. pylori GGT has positively charged patches globally distributed across its molecular surface (Fig. 5C), consistent with the theoretical pI value (9.12) calculated from the amino acid sequence. Contrary to our expectation that more negatively charged residues would be present on the molecular surface of B. subtilis GGT than on that of E. coli GGT, there was no significant difference in surface potential between the two GGTs. To better understand the factors underlying the high salt toler- ance, the effect of salt concentration on the enzyme sur- face properties was assessed by solving the Poisson– Boltzmann equation. At high salt concentrations, nota- bly different results were observed in the electrostatic surface potentials between the two GGTs; in B. subtilis GGT, apparent negatively charged areas were main- tained on the surface in the presence of 3 m monovalent ion, whereas in E. coli GGT, under the same condi- tions, the negatively charged patches on the surface that had been observed in the absence of salt had completely disappeared. The negatively charged areas of B. subtilis GGT under high-salt conditions even increased the solvation, owing to increased water-binding capacity. This may allow the protein to remain in a hydrated state, preventing the binding of inorganic cations in the high-salt solution, and also preventing self-aggregation. B. subtilis GGT may be applicable to the manufac- ture of fermented food, because this enzyme possesses A B C B. subtilis GGT (0 M NaCl) 90° E. coli GGT (0 M NaCl) B. subtilis GGT (3 M NaCl) E. coli GGT (3 M NaCl) H. py lori GGT ( 0 M NaCl ) Fig. 5. Surface electrostatic properties of GGTs. (A) Electrostatic potentials of B. subtilis (upper panels) and E. coli (lower panels) GGTs calculated using parameters without taking into account ion strength. (B) Electrostatic potentials of B. subtilis (upper panels) and E. coli (lower panels) GGTs calculated for 3 M concentration of monovalent ion. (C) The surface potential of H. pylori GGT calculated using para- meters without taking into account ion strength. The color scale ranges from )10 kT per electron (red) to +10 kT per electron (blue). The struc- tures are graphically depicted with the viewpoint looking down the cat- alytic pocket (right panels) or rotated 90° from this view (left panels). K. Wada et al. Structure of B. subtilis c-glutamyltranspeptidase FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS 1005 not only c-glutamyl compound-degrading activity (GGT activity) but also the steady glutaminase activity [16] that converts glutamine into glutamic acid and ammonia. Decreasing the level of glutamine in fer- mented foods improves the taste, as glutamine is spon- taneously converted to slightly sour pyroglutamic acid. Although glutaminases from fungi such as Aspergil- lus oryzae or Aspergillus sojae are applied to remove the glutamine during the process of fermentation, the activity of these glutaminases is seriously reduced by the high salt concentration. In contrast, halotolerant B. subtilis GGT could be used to increase the level of glutamic acid, a major component of the desired taste, at the same time decreasing the amount of glutamine, even under high-salt conditions, by both its GGT activity and its glutaminase activity. The structural basis for the salt tolerance presented here could be a guide for further improvements in the usefulness of B. subtilis GGT by protein engineering. Experimental procedures Cloning of the ggt gene into an expression vector The ggt gene from B. subtilis was amplified by PCR from the plasmid pCY167 (Suzuki H & Yamada C, Unpublished), using forward primer 5¢- CATATGGATGAGTACAAACA AGTAGATG-3¢ and reverse primer 5¢- GGATCCTCGAG CTCATTTACGTTTTAAATTAATGCCGAT-3¢ (underlin- ed sequences indicate NdeI and BamHI sites, respectively). The PCR product was initially subcloned into pTA2 (Toyobo, Osaka, Japan), and the sequence was confirmed. As the origi- nal ggt sequence has one NdeI site in the middle, we intro- duced a synonymous mutation into this site, using the Quikchange Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA, USA) with forward primer 5¢-GAAACGATGC ATTTGTCCTATGCCGACCGTGCGTC-3¢ and reverse primer 5¢-GACGCACGGTCGGCATAGGACAAATGCA TCGTTTC-3¢. The sequence of the second PCR product was also confirmed. Following the digestion of the second PCR product with NdeI and BamHI, the DNA fragment containing the ggt gene was ligated into the pCold-I vector (Takara Bio, Shiga, Japan), and pCold I–His 6 –ggt was subsequently generated. Overproduction and purification of B. subtilis GGT The pCold I–His 6 –ggt expression vector was transformed into E. coli C41(DE3) [25]. The transformant was grown at 37 °C in 3.6 L (900 mL · 4) of liquid TB containing ampi- cillin (50 lg ⁄ mL) to an attenuance of 0.6 at 600 nm. At this stage, expression of the N-terminal His-tagged GGT was induced by decreasing the temperature from 37 °Cto 15 °C, and then adding 1 mm isopropyl-b-d-thiogalactopyr- anoside. After induction, the transformant was cultured at 15 °C for 30 h. The cells were harvested, resuspended in 50 mm Tris ⁄ HCl (pH 7.8) containing 20 mm imidazole, and disrupted by sonication. The soluble fraction was mixed with His-select resin (Sigma, St Louis, MO, USA), and the N-terminal His-tagged GGT was purified by batch method according to the manufacturer’s protocol. Fractions contain- ing GGT were collected and concentrated with ammonium sulfate at 70% saturation. The precipitate was dissolved in 50 mm Tris ⁄ HCl (pH 7.8), and was then subjected to gel filtration using a HiPrep 16 ⁄ 60 Sephacryl S-200 HR column (GE Healthcare, Milwaukee, WI, USA). All purification steps were performed at 4 °C. Fractions containing GGT were monitored by absorption at 280 nm and GGT activity [26], and purity was confirmed by SDS ⁄ PAGE. Crystallization of B. subtilis GGT The purified GGT was concentrated to 10 mg ⁄ mL with a Vivaspin filter (GE Healthcare). All crystallization trials were performed at 4 °C, using the hanging-drop vapor-diffusion method. Crystallization drops containing 1 lL of protein solu- tion in 20 mm Hepes (pH 7.8) and 0.5 mml-glutamate, and 1 lL of precipitant solution, were equilibrated against 200 lL of precipitant solution. The initial trials were performed using the following commercially available sparse-matrix screening kits: Crystal Screen I, II and Lite, PEG ⁄ Ion screen (Hampton Research, Aliso Viejo, CA, USA), Wizard I–III (Emerald BioSystems, Bainbridge Island, WA, USA), and JB screen 1–6 (Jena Bioscience GmbH, Jena, Germany). The crystallization trials produced small crystals in sev- eral drops containing poly(ethylene glycol) as a precipitant (e.g. PEG ⁄ Ion screen, tube nos. 10, 11, 13 and 14; Wiz- ard III, tube no. 10; JB screen 3, C2). The crystals were improved using Additive Screen (Hampton Research), and the conditions were then manually optimized using home- made solutions. The best crystals of His 6 –GGT were grown in the drop containing a 1 : 1 mixture of protein solution [10 mg ⁄ mL in 20 mm Hepes (pH 7.8) containing 0.5 mm l-glutamate] and reservoir solution [poly(ethylene gly- col) 4000, 100 mm Mes (pH 7.0), 600 mm NaCl, and 5% (v ⁄ v) Jeffamine M-600]. Crystals grew to maximum dimensions of 0.05 · 0.1 · 0.4 mm in 1 week. Data collection for B. subtilis GGT The crystals were transferred to a reservoir solution con- taining 15% (v ⁄ v) glycerol as a cryoprotectant for a few seconds, and then flash-cooled in a cryostream at )180 °C. To measure the GGT crystals, we used a goniometer head with a large arc to alter the rotation axis of the mounted crystal; because the spaces between reciprocal points along the c*-axis were so small, it was necessary to choose a rota- tion axis nearly parallel to the c-axis to avoid overlapping Structure of B. subtilis c-glutamyltranspeptidase K. Wada et al. 1006 FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS of the diffraction spots on a frame derived from the adja- cent reciprocal lattice planes. Intensity datasets were collected at the BL38B1 station of the SPring-8 (Hyogo, Japan), using the oscillation method on an ADSC Q210 detector (Area Detector Systems Corporation, Poway, CA, USA) with synchrotron radiation (k = 1.000 A ˚ ). The crys- tal-to-detector distance was 200 mm; 1470 images were recorded at 0.1° intervals, with an exposure time of 15 s per image. The intensity data were processed and scaled with xds [27]. The results of the data collection are summarized in Table 1. Structure determination for GGT When we reached the stage of solving the phase problem, we learned that the coordinates for substrate-free B. subtilis GGT had been deposited by Sharath et al. at the RCSB Protein Data Bank (2V36). Because their crystal form was different from ours, we applied the molecular replacement method to solve our crystal structure, using their coordi- nates as the search probe. Rotational and translational searches of the diffraction data (15.0–4.0 A ˚ resolution), per- formed using molrep [28] from the ccp4 package, located two crystallographically independent molecules in an asym- metric unit. The structure was subjected to rigid-body refinement for 25–3.0 A ˚ resolution data, using cns [29]. The structure was further refined at 1.95 A ˚ resolution with the cns simulated annealing protocol, and this was fol- lowed by energy minimization and individual temperature- factor refinements; manual model building was performed with xfit [29a]. The electron density map at this stage was clear enough for exact assignment of the orientations of the two glutamates in the asymmetric unit, and the model was unambiguously fitted to the F o –F c map of the substrate- binding pocket of each molecule. The ordered water mole- cules were added to the model using the cns water-pick and water-delete functions. Finally, energy minimization and temperature-factor refinements were applied to the model. Although the crystallization drops contain 600 mm NaCl, the electron densities and the temperature factors of the picked atoms indicated that neither sodium ion nor chloride ion was bound to this GGT. Structure refinement statistics are summarized in Table 1. Atomic coordinates and structure factors have been deposited in the RCSB Pro- tein Data Bank (http://www.rcsb.org) under accession number 3A75. The software programs used were as follows: promotif [30] for secondary structure assignment, procheck [31] for the validity of the final model, and lsqman [32] for super- position and rmsd values of the structures. The electrostatic potentials of the molecular surface were calculated with pbeq-solver [33], which uses the Poisson–Boltzmann equa- tions module from the biomolecular simulation program charmm [34]. Acknowledgements We thank S. Baba, N. Mizuno and T. Hoshino for their assistance with data collection using the synchro- tron radiation at SPring-8 (Hyogo, Japan). The syn- chrotron radiation experiments were performed at BL38B1, SPring-8, with the approval of the Japan Synchrotron Radiation Research Institute (Proposal No. 2008B1079). We also thank M. Sugishima of Kurume University for valuable comments and sugges- tions, and N. Kaseda of Osaka University for technical assistance. This work was supported by a grant from the Japan Foundation for Applied Enzymology (to K. Fukuyama), Grants-in-Aid for Scientific Research 21380059 (to H. Suzuki), 20370037 (to K. Fukuyama) and 21770112 (to K. Wada) from the Ministry of Table 1. Crystallographic data and refinement statistics (values in parentheses are for the outermost shell). Crystallographic data Space group P 2 1 2 1 2 1 Cell parameters (A ˚ ) a = 49.4, b = 98.9, c = 227.9 Resolution range (A ˚ ) 25.0–1.95 (2.02–1.95) Observed reflections 408 102 Unique reflections 76 177 Mean I ⁄ r (I) 18.3 (4.6) Redundancy 5.4 (3.4) Completeness (%) 92.5 (86.2) R sym (%) a 6.2 (26.2) Refinement statistics R cryst (%) b 20.8 R free (%) c 26.2 Disordered regions d Molecule A L-subunit 36–37, 396–402 S-subunit 585–587 Molecule B L-subunit 36–38, 396–402 S-subunit – rmsd from ideal values Bond length (A ˚ ) 0.017 Bond angle (°) 1.9 Average B-factor (A ˚ 2 ) 23.3 Ramachandran plot Most favored (%) 91.0 Additionally allowed (%) 8.8 Generously allowed (%) 0.0 Disallowed (%) e 0.2 a R sym = P hkl P i |I i (hkl )–<I(hkl )>| ⁄ P hkl P i I i (hkl ), where <I(hkl )> is the average intensity over equivalent reflections. b R cryst = P ||F obs (hkl )| – |F calc (hkl )|| ⁄ P |F obs (hkl )|. c R free is the R-value calculated for 5% of the dataset not included in the refinement. d Numerals shown are invisible residue numbers. e Glu423, which corresponds to Asn411 in E. coli GGT, in the two crystallographically independent molecules. K. Wada et al. Structure of B. subtilis c-glutamyltranspeptidase FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS 1007 Education, Culture, Sports, Science and Technology of the Japanese Government, and a research grant from Towa Food Research Foundation (to H. Suzuki). References 1 Tate SS & Meister A (1981) c -Glutamyl transpeptidase: catalytic, structural and functional aspects. Mol Cell Biochem 39, 357–368. 2 Mehdi K & Penninckx MJ (1997) An important role for glutathione and c-glutamyltranspeptidase in the supply of growth requirements during nitrogen starva- tion of the yeast Saccharomyces cerevisiae. Microbiology 143, 1885–1889. 3 Martin MN, Saladores PH, Lambert E, Hudson AO & Leustek T (2007) Localization of members of the c-glut- amyl transpeptidase family identifies sites of glutathione and glutathione S-conjugate hydrolysis. Plant Physiol 144, 1715–1732. 4 Suzuki H, Kumagai H & Tochikura T (1986) c-Glutam- yltranspeptidase from Escherichia coli K-12: formation and localization. J Bacteriol 168, 1332–1335. 5 Taniguchi N & Ikeda Y (1998) c-Glutamyl transpepti- dase: catalytic mechanism and gene expression. Adv Enzymol Relat Areas Mol Biol 72, 239–278. 6 Ohkama-Ohtsu N, Fukuyama K & Oliver DJ (2009) Roles of c-glutamyl transpeptidase and c-glutamyl cyclotransferase in glutathione and glutathione- conjugate metabolism in plants. Adv Bot Res 52, 87–113. 7 Suzuki H & Kumagai H (2002) Autocatalytic process- ing of c-glutamyltranspeptidase. J Biol Chem 277, 43536–43543. 8 Okada T, Suzuki H, Wada K, Kumagai H & Fukuyama K (2006) Crystal structures of c-glutamyl- transpeptidase from Escherichia coli, a key enzyme in glutathione metabolism, and its reaction intermediate. Proc Natl Acad Sci USA 103, 6471–6476. 9 Boanca G, Sand A, Okada T, Suzuki H, Kumagai H, Fukuyama K & Barycki JJ (2007) Autoprocessing of Helicobacter pylori c-glutamyltranspeptidase leads to the formation of a threonine–threonine catalytic dyad. J Biol Chem 282, 534–541. 10 Okada T, Suzuki H, Wada K, Kumagai H & Fukuyama K (2007) Crystal structure of the c-glutamyl- transpeptidase precursor protein from Escherichia coli. Structural changes upon autocatalytic processing and implications for the maturation mechanism. J Biol Chem 282, 2433–2439. 11 Wada K, Hiratake J, Irie M, Okada T, Yamada C, Kumagai H, Suzuki H & Fukuyama K (2008) Crystal structures of Escherichia coli c-glutamyltranspeptidase in complex with azaserine and acivicin: novel mechanis- tic implication for inhibition by glutamine antagonists. J Mol Biol 380, 361–372. 12 Williams K, Cullati S, Sand A, Biterova EI & Barycki JJ (2009) Crystal structure of acivicin-inhibited c-glut- amyltranspeptidase reveals critical roles for its C-termi- nus in autoprocessing and catalysis. Biochemistry 48, 2459–2467. 13 Morrow AL, Williams K, Sand A, Boanca G & Barycki JJ (2007) Characterization of Helicobacter pylori c-glut- amyltranspeptidase reveals the molecular basis for sub- strate specificity and a critical role for the tyrosine 433-containing loop in catalysis. Biochemistry 46, 13407–13414. 14 Uchiyama I (2007) MBGD: a platform for microbial comparative genomics based on the automated con- struction of orthologous groups. Nucleic Acids Res 35, D343–D346. 15 Minami H, Suzuki H & Kumagai H (2003) A mutant Bacillus subtilis c-glutamyltranspeptidase specialized in hydrolysis activity. FEMS Microbiol Lett 224, 169–173. 16 Minami H, Suzuki H & Kumagai H (2003) Salt-tolerant c-glutamyltranspeptidase from Bacillus subtilis 168 with glutaminase activity. Enzyme Microb Technol 32, 431–438. 17 Duggleby HJ, Tolley SP, Hill CP, Dodson EJ, Dodson G & Moody PC (1995) Penicillin acylase has a single- amino-acid catalytic centre. Nature 373, 264–268. 18 Oinonen C, Tikkanen R, Rouvinen J & Peltonen L (1995) Three-dimensional structure of human lysosomal aspartylglucosaminidase. Nat Struct Biol 2, 1102–1108. 19 Wu R, Richter S, Zhang RG, Anderson VJ, Missiakas D & Joachimiak A (2009) Crystal structure of Bacillus anthracis transpeptidase enzyme CapD. J Biol Chem 284, 24406–24414. 20 Richter S, Anderson VJ, Garufi G, Lu L, Budzik JM, Joachimiak A, He C, Schneewind O & Missiakas D (2009) Capsule anchoring in Bacillus anthracis occurs by a transpeptidation reaction that is inhibited by capsidin. Mol Microbiol 71, 404–420. 21 Frolow F, Harel M, Sussman JL, Mevarech M & Shoham M (1996) Insights into protein adaptation to a saturated salt environment from the crystal structure of a halophilic 2Fe–2S ferredoxin. Nat Struct Biol 3, 452–458. 22 Premkumar L, Greenblatt HM, Bageshwar UK, Savchenko T, Gokhman I, Sussman JL & Zamir A (2005) Three-dimensional structure of a halotolerant algal carbonic anhydrase predicts halotolerance of a mammalian homolog. Proc Natl Acad Sci USA 102, 7493–7498. 23 Grochowski P & Trylska J (2008) Continuum molecular electrostatics, salt effects, and counterion binding – a review of the Poisson–Boltzmann theory and its modi- fications. Biopolymers 89, 93–113. 24 Kuntz ID (1971) Hydration of macromolecules. IV. Polypeptide conformation in frozen solutions. JAm Chem Soc 93, 516–518. Structure of B. subtilis c-glutamyltranspeptidase K. Wada et al. 1008 FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS 25 Miroux B & Walker JE (1996) Over-production of pro- teins in Escherichia coli: mutant hosts that allow synthe- sis of some membrane proteins and globular proteins at high levels. J Mol Biol 260, 289–298. 26 Suzuki H, Kumagai H & Tochikura T (1986) c-Glutam- yltranspeptidase from Escherichia coli K-12: purification and properties. J Bacteriol 168, 1325–1331. 27 Kabsch W (1993) Automatic processing of rotation diffraction data from crystals of initially unknown sym- metry and cell constants. J Appl Crystallogr 26, 795–800. 28 Vagin A & Teplyakov A (2000) An approach to multi- copy search in molecular replacement. Acta Crystallogr D Biol Crystallogr 56, 1622–1624. 29 Bru ¨ nger AT, Adams PD, Clore GM, DeLano WL, Gros P, Grosse-Kunstleve RW, Jiang JS, Kuszewski J, Nilges M, Pannu NS et al. (1998) Crystallography & NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr D Biol Crystallogr 54, 905–921. 29a McRee DE (1999) XtalView ⁄ Xfit – a versatile program for manipulating atomic coordinates and electron den- sity. J Struct Biol 125 , 156–165. 30 Hutchinson EG & Thornton JM (1996) PROMOTIF: a program to identify and analyze structural motifs in proteins. Protein Sci 5, 212–220. 31 Laskowski RA, MacArthur MW, Moss DS & Thornton JM (1993) PROCHECK: a program to check the ste- reochemical quality of protein structures. J Appl Crys- tallogr 26, 283–291. 32 Kleywegt GJ & Jones TA (1994) A super position. ESF ⁄ CCP4 Newslett 31, 9–14. 33 Jo S, Kim T, Iyer VG & Im W (2008) CHARMM- GUI: a web-based graphical user interface for CHARMM. J Comput Chem 29, 1859–1865. 34 Brooks BR, Brooks CL 3rd, Mackerell AD Jr, Nilsson L, Petrella RJ, Roux B, Won Y, Archontis G, Bartels C, Boresch S et al. (2009) CHARMM: the biomolecular simulation program. J Comput Chem 30, 1545– 1614. 35 Thompson JD, Higgins DG & Gibson TJ (1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weight- ing, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22, 4673–4680. 36 Gouet P, Courcelle E, Stuart DI & Metoz F (1999) ESPript: analysis of multiple sequence alignments in PostScript. Bioinformatics 15, 305–308. 37 DeLano WL (2002) PyMOL. DeLano Scientific, San Carlos, CA. K. Wada et al. Structure of B. subtilis c-glutamyltranspeptidase FEBS Journal 277 (2010) 1000–1009 ª 2010 The Authors Journal compilation ª 2010 FEBS 1009 . Crystal structure of the halotolerant c-glutamyltranspeptidase from Bacillus subtilis in complex with glutamate reveals a unique architecture of the solvent-exposed. pockets in the asymmetric unit, and the glutamate- binding modes are identical to each other (Fig. 2A) . The a- carboxyl and a- amino groups of the bound glutamate