Crystal Structures of 4-�-Glucanotransferase fromThermococcus litoralis and Its Complex with an Inhibitor*

Received for publication, December 23, 2002, and in revised form, February 24, 2003Published, JBC Papers in Press, March 4, 2003, DOI 10.1074/jbc.M213134200

Hiromi Imamura‡§, Shinya Fushinobu‡¶, Masaki Yamamoto¶, Takashi Kumasaka¶�,Beong-Sam Jeon‡**, Takayoshi Wakagi‡, and Hiroshi Matsuzawa‡‡§§

From the ‡Department of Biotechnology, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan,the ¶Harima Institute, The Institute of Physical and Chemical Research (RIKEN), 1-1-1 Kouto, Mikazuki-cho, Sayo-gun,Hyogo 678-5148, Japan, and the ‡‡Department of Bioscience and Biotechnology, Aomori University, 2-3-1 Kohbata,Aomori 030-0943, Japan

Thermococcus litoralis 4-�-glucanotransferase (TLGT)belongs to glucoside hydrolase family 57 and catalyzesthe disproportionation of amylose and the formation oflarge cyclic �-1,4-glucan (cycloamylose) from linear am-ylose. We determined the crystal structure of TLGT withand without an inhibitor, acarbose. TLGT is composedof two domains: an N-terminal domain (domain I), whichcontains a (�/�)7 barrel fold, and a C-terminal domain(domain II), which has a twisted �-sandwich fold. In thestructure of TLGT complexed with acarbose, the inhib-itor was bound at the cleft within domain I, indicatingthat domain I is a catalytic domain of TLGT. The acar-bose-bound structure also clarified that Glu123 andAsp214 were the catalytic nucleophile and acid/base cat-alyst, respectively, and revealed the residues involvedin substrate binding. It seemed that TLGT produceslarge cyclic glucans by preventing the production ofsmall cyclic glucans by steric hindrance, which isachieved by three lids protruding into the active sitecleft, as well as an extended active site cleft. Interest-ingly, domain I of TLGT shares some structural featureswith the catalytic domain of Golgi �-mannosidase fromDrosophila melanogaster, which belongs to glucosidehydrolase family 38. Furthermore, the catalytic residueof the two enzymes is located in the same position. Theseobservations suggest that families 57 and 38 evolvedfrom a common ancestor.

In the maltose metabolism of the hyperthermophilicarchaeon, Thermococcus litoralis, 4-�-glucanotransferase

(TLGT)1 (EC 2.4.1.25) plays a key role, producing glucose anda series of maltodextrins through intermolecular transglycosy-lation of maltose that has been transported into the cells (1). Inaddition to intermolecular transglycosylation, TLGT also cata-lyzes intramolecular transglycosylation in vitro, where it cy-clizes amylose to produce cyclic �-1,4-glucans (cycloamyloses,CAs) (2) with 16 to several hundred glucose units.2 The degreeof polymerization of known CAs varies, from six to severalhundreds of glucose units. �-, �-, and �-cyclodextrins, which arethe smallest CA species, consisting of 6, 7, and 8 glucose units,respectively, are well known doughnut-shaped rigid moleculesand are able to accommodate guest molecules in their centralcavity, yielding inclusion complexes. Although larger species ofCA, which were recently found to be products of potato 4-�-glucanotransferase (3), also form inclusion complexes, theyhave flexible single-helical conformations in an aqueous solu-tion unlike cyclodextrins (4), and they lose their flexibility andfold into a compact structure during complex formation (5).Because of this structural difference from cyclodextrins, largeCAs show the advantageous features of the formation of inclu-sion complexes and higher solubility in water and thus areexpected to be valuable for future industrial use (6). Interest-ingly, an artificial chaperone activity of large CAs has also beenreported (7).

According to Henrissat’s classification (8, 9), TLGT belongsto family 57 of the glycoside hydrolases. Most family 57 en-zymes catalyze reactions similar to those of some �-amylasefamily members (families 13, 70, and 77). However, no se-quence similarity has been detected between family 57 and�-amylase family enzymes. The three-dimensional structuresof many �-amylase family enzymes, including that ofThermus aquaticus amylomaltase (10), which produces largeCAs, have been determined (11–13), and the amino acid resi-dues involved in the catalysis have also been studied exten-sively (for a review, see Ref. 14). In contrast to �-amylasefamily enzymes, family 57 enzymes have received less investi-gation. Although the catalytic nucleophile of TLGT was re-cently determined (15), its three-dimensional structure, acid/base catalyst, and mechanism for large CA production remainunknown.

Previously, we made a preliminary report of the crystalstructure of TLGT (16). In this paper, we describe the detailedstructures of TLGT with and without a tetrasaccharide inhib-itor, acarbose. The structures revealed the residues involved in

* This work was supported in part by Grants-in-Aid for ScientificResearch 10460035 and 12460047 (to H. M.) from the Japan Society forthe Promotion of Science. The costs of publication of this article weredefrayed in part by the payment of page charges. This article musttherefore be hereby marked “advertisement” in accordance with 18U.S.C. Section 1734 solely to indicate this fact.

The atomic coordinates and structure factors (codes 1K1W, 1K1X, and1K1Y) have been deposited in the Protein Data Bank, Research Collabo-ratory for Structural Bioinformatics, Rutgers University, New Bruns-wick, NJ (http://www.rcsb.org/).

§ Supported by a fellowship for young researcher of the Japan Societyfor the Promotion of Science. To whom correspondence may be ad-dressed. Present address: ATP System Project, ERATO, JST, 5800-3Nagatsuta, Midori-ku, Yokohama 226-0026, Japan. Fax: 81-45-922-5239; E-mail: himamura-ra@res.titech.ac.jp.

� Present address: Dept. of Life Science, Tokyo Institute of Technol-ogy, 4259 Nagatsuta, Midori-ku, Yokohama 226-8501, Japan.

** Presentaddress:Dept. ofMicrobiology,CollegeofMedicine,Gyeong-sang National University, 90 Chilam-Dong, Jinju, Gyeong-Nam 660-750, Korea.

§§ To whom correspondence may be addressed. Fax: 81-17-738-2030;E-mail: hmatsuza@aomori-u.ac.jp.

1 The abbreviations used are: TLGT, 4-�-glucanotransferase fromT. litoralis; CA(s), cycloamylose(s); CGTase, cyclodextrin glucano-transferase.

2 B. S. Jeon and H. Matsuzawa, unpublished results.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 21, Issue of May 23, pp. 19378–19386, 2003© 2003 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www.jbc.org19378

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

catalysis and substrate binding of the enzyme and providedinsight to investigate the mechanism for the production of largeCAs.

EXPERIMENTAL PROCEDURES

Purification and Crystallization—The nonlabeled enzyme was ex-pressed in Escherichia coli strain BL21(DE3) and purified as de-scribed previously (15) with some modifications. In the last purifica-tion step, the buffer was changed to 5 mM Tris-HCl, pH 7.5, and theenzyme was concentrated to 10–20 mg/ml by centrifugal filtrationusing a Centriplus 30 (Millipore). The seleno-methionine-labeled en-zyme was coexpressed in the methionine auxotrophic E. coli strainB834(DE3) (Novagen) with GroELS and tRNA cognates for AGA andAGG in a medium containing seleno-methionine. The purificationprocedures for the labeled enzyme were the same as those for thenonlabeled enzyme, except that the buffer used in the last step wasreplaced by 5 mM Tris-HCl, pH 7.5, containing 10 mM dithiothreitol.

Form I crystals were obtained at 25 °C by the hanging drop vapordiffusion method by mixing 5 �l of protein solution with 5 �l of areservoir solution comprising 1.9 M ammonium sulfate, 2% (w/v)polyethylene glycol 400, and 0.1 M HEPES-NaOH, pH 7.5. The crys-tals were harvested in a solution comprising 2.2 M ammonium sulfate,2% (w/v) polyethylene glycol 400 and 0.1 M HEPES-NaOH, pH 7.5.The concentration of trehalose (a gift from Hayashibara (Okayama,Japan), recrystallized before use) in the harvesting solution wasincreased stepwise to 25% (w/v), and then the crystals were subjectedto flash freezing before data collection. Form II crystals were grown at25 °C by the sitting drop vapor diffusion method. The drops weremade by mixing 5 �l of protein solution with 5 �l of a reservoircomprising 35% (w/v) 2,4-dimethylpentanediol (Hampton Research),20 mM calcium chloride, and 0.1 M Tris-HCl, pH 8.0. The crystals weretransferred to a solution comprising 40% (w/v) 2,4-dimethylpen-tanediol, 20 mM calcium chloride, and 0.1 M Tris-HCl, pH 8.0, followedby flash freezing. Acarbose-bound crystals were obtained by soaking

FIG. 1. Schematic representation of TLGT. a, two TLGT monomers in an asymmetric unit of form II-free. Calcium ions are presented asyellow space-filled models. Two catalytic residues and Tris are presented as stick models. b, stereoview of a C� trace of a TLGT monomer. ChainA of form II-free is shown. Every 40th C� is represented as a sphere. The molecule is viewed from the same perspective as in a. Figs. 1, 3, 4c, 6, and8 were prepared with Molscript (44) and Raster3d (45).

TABLE IData collection and refinement statistics

Form I-Se Form II-free Form II-complex

Beamline SPring-8 BL45PX SPring-8 BL45PX Photon factory BL18BCell constant (Å) a � b � 124.9, c � 246.9 a � 137.7,

b � 161.5, c � 70.4a � 137.6,

b � 161.0, c � 70.2Wavelength (Å) 1.0400 0.9793 0.9795 1.0200 1.0000Resolution range (Å)a 20–2.80 20–2.80 20–2.80 50–2.40 40–2.4

(2.90–2.80) (2.90–2.80) (2.90–2.80) (2.49–2.40) (2.53–2.40)Observed reflections 107608 104954 106848 394421 274277Unique reflectionsa 27918 (2760) 27937 (2724) 27795 (2731) 61900 (6064) 61753 (8891)Completeness (%)a 98.4 (98.9) 98.5 (98.2) 98.3 (98.2) 99.7 (99.3) 99.9 (99.9)�I�/��� (%)a 11.9 (7.2) 9.0 (4.5) 11.7 (7.0) 10.6 (9.8) 6.3 (2.2)Rmerge (%)a 4.2 (18.6) 6.7 (24.6) 4.4 (17.7) 8.3 (15.7) 9.0 (32.1)

Figure of meritb 0.579 (0.710)R-factor/Rfree (%)a 22.1/28.9 (29.0/37.4) 19.5/23.6 (24.0/26.4) 19.8/25.6 (25.5/33.4)No. of protein atoms 5272 10695 10695No. of solvent atoms 68 499 572No. of heteroatoms 18 11 79Average B (Å2)

Protein 56.5 35.6 40.7Water 40.6 30.9 35.0Substrates/Analogs 64.1 40.9

Root mean square deviationsBond lengths (Å) 0.0071 0.0062 0.0065Angles (°) 1.41 1.26 1.27

Protein Data Bank accession codes 1K1W 1K1X 1K1Ya The values in parentheses are for the outer resolution shell.b The value in parenthesis is after density modification. The values are for the resolution range of 9.4–2.8 Å.

Crystal Structure of Glycoside Hydrolase Family 57 19379

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

crystals overnight in a solution comprising 40% (w/v) 2,4-dimeth-ylpentanediol, 20 mM calcium chloride, 10 mM acarbose (a gift fromBayer), and 0.1 M Tris-HCl, pH 8.0, followed by flash freezing.

Data Collection—A multiple wavelength anomalous dispersion dataset for a seleno-methionine-labeled form I crystal was collected atwavelengths of 1.0400 Å (remote), 0.9793 Å (peak), and 0.9795 Å (edge)at 100 K on beamline BL45PX of SPring-8 (Hyogo, Japan). A nativedata set for a form II crystal was collected at a wavelength of 1.0200 Åat 100 K on beamline BL45PX of SPring-8. An acarbose-bound data setfor a form II crystal was collected at a wavelength of 1.0000 Å at 100 Kon beamline BL18B of the Photon Factory (Tsukuba, Japan). The firsttwo data sets were processed with DENZO and Scalepack (17), and thelast one was processed with Mosflm (18).

Phase Calculation and Refinement—For phase calculation of themultiple wavelength anomalous dispersion data set, the programSOLVE (19) was used. The multiple wavelength anomalous dispersionmap was improved by density modification using the program DM (20).The model for “form I-Se” was built into the resultant electron densitymap using the program O (21) and refined to 2.8 Å resolution, includingsimulated annealing, bulk solvent correction, and grouped B-factorrefinement with the program CNS (22). Water molecules were pickedautomatically from Fo � Fc electron density maps using the programCNS and checked manually at a graphic station. The phase for formII-free crystal was calculated by molecular replacement using form I-Seas the initial model using the program CNS. Noncrystallographic sym-

metry was found on the self-rotation function, but it was not used forcross-rotation function nor translation function. Density modificationwas not applied to the molecular replacement solution. The calculatedmodel was refined to 2.4 Å resolution, including rigid body minimiza-tion, simulated annealing, and grouped B-factor refinement with theprogram CNS. The two TLGT monomers in the asymmetric unit werenot restrained during refinement, because there are some variations inthe conformation of the two molecules as described later. The finalmodel was found to exhibit good geometry, as determined using theprogram Procheck (23); 88.7% of the residues have �/� angles in the“most favored region” of a Ramachandran plot. The model for form IIcomplex was also refined using the program CNS. During refinement ofform II complex, the ligands were added based on 2Fo � Fc and Fo �Fc electron density maps. The refinement statistics are presented inTable I. A structural data base search was performed using the DALIserver (24). Least mean square fitting of the structures was carried outwith the program LSQMAN (25). The oligosaccharide model shown inFig. 7 was constructed using the program XtalView (26).

Site-directed Mutagenesis and Enzyme Assay—The D214N mutantwas generated with a QuikChange site-directed mutagenesis kit (Strat-agene). An oligonucleotide with the sequence 5�-GTGTTCCATGA-CAATGGTGAAAAGTTCGG-3� and a complementary oligonucleotide,which replaced the codon for Asp214 (GAC) with AAT and introduced aHphI site for rapid screening of the mutation, were used. The mutationwas reconfirmed by sequencing with an ABI PRISM 310 DNA se-

FIG. 2. Sequence alignment of TLGT and two family 57 enzymes. PFAM, �-amylase from Pyrococcus furiosus; DTAM, �-amylase A fromDictyoglomus thermophilum. The residue numbers refer to the amino acid sequence of TLGT. Alignment was performed with ClustalW, version1.81 (46). The secondary structure was assigned with DSSP (47). Identical residues are shown in white on red, and homologous residues are in redletters. Arrowheads indicate the two catalytic residues. The closed blue circles represent residues whose side chains are involved in substratebinding. The residues indicated by closed yellow circles are involved in calcium binding. Open purple and green circles indicate the residues thatcontribute to the intermonomer interactions via hydrophobic interactions and hydrogen bond networks, respectively. Functions of the residueswere determined for TLGT as described in this study. This figure was generated using ESPript (48).

Crystal Structure of Glycoside Hydrolase Family 5719380

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

quencer (Applied Biosystems). Activity toward maltotriose was meas-ured as described previously (27). One unit of activity was defined asthe amount of enzyme that liberated 1 �mol of glucose from maltotrioseper min at 80 °C.

RESULTS

Structure Determination—TLGT was crystallized in twoforms (forms I and II). Form I crystals belonged to hexagonalspace group P6422 and form II crystals to orthorhombic spacegroup P21212. First, we determined the structure at 2.8 Åresolution by multiple wavelength anomalous dispersionmethod using a form I crystal labeled with seleno-methionine.The model was refined to an R-factor of 22.1% (Table I). Wecalled this structural model form I-Se. The form I crystal con-tained one TLGT monomer/asymmetric unit. Next, we deter-mined the structures of a form II crystal through molecularreplacement in the free form, designated as form II-free, and ina complex form with acarbose, designated as the form II-com-plex. Acarbose, an inhibitor of �-amylase family enzymes,�-amylases, and glucoamylases, is a maltotetraose analog com-posed of a nonreducing end inhibitory group, acarviosine, and areducing end maltose. Acarbose also inhibits the activity ofTLGT.3 The form II crystal contained two TLGT monomers(chains A and B)/asymmetric unit (Fig. 1a). The form II-freestructure was refined to an R-factor of 19.5% at 2.4 Å resolution(Table I).

Overall Structure—The TLGT monomer has approximatedimensions of 50 � 55 � 90 Å. TLGT is composed of two distinctdomains; an N-terminal domain (domain I, residues 1–381) anda C-terminal domain (domain II, residues 389–659) (Fig. 1),comprising 20 �-helices, 24 �-strands, and eight 310-helices(Fig. 2). There is no cross-over between the two domains.

Domain I—The core of domain I is a (�/�)7 barrel (residues1–258), in which seven central parallel �-strands (�1, �2, �3,�4, �5, �7, and �8) are flanked by helices (Fig. 3a). The (�/�)7barrel is followed by a helical region that consists of 10 helicesincluding the three-helix bundle made of �16, �17, and �18(Figs. 1a and 3b) and covers the C termini of the (�/�)7 barrel,forming a cleft between them. The observation that acarbosewas bound in the cleft in the form II complex (see below) andthat the catalytic nucleophile Glu123 (15) was located in themiddle of the cleft indicates that the cleft is an active site anddomain I is a catalytic domain.

Domain II—Domain II of TLGT is composed entirely of�-strands, with the exception of two short �-helices (�19 and�20). This domain is characterized by a �-sandwich fold, inwhich two layers of anti-parallel �-sheets are arranged in anearly parallel manner. The layer adjacent to domain I consistsof 10 �-strands (�10, �11, �12, �13, �14, �15, �20, �21, �22,and �23), and the other layer consists of seven �-strands (�16,�17, �18, �19, �21, �22, and �24). �21 and �22 bend sharplyand lie across the two layers. The same fold is found in chon-droitin AC lyase from Flavobacterium heparinum (ProteinData Bank code 1CB8), �-galactosidase from E. coli (1BGL),hyaluronate lyase from Streptococcus pneumoniae (1EGU),methylamine oxidase from Hansenula polymorpha (1A2V), andcopper amine oxidases from pea seedlings (1KSI) and E. coli(1OAC). Although the function of domain II is unclear at pres-ent, this domain may play a role in transglycosylation reactionof TLGT, because it has been suggested that �-sandwich do-main of E. coli �-galactosidase is involved in transglycosylationreaction (28).

Calcium-binding Site—A calcium ion was bound at the loopbetween �10 and �11 in domain II (Figs. 1a and 2). The calciumion was coordinated with O�1 of Asp392, O�2 of Asp394, O�1 ofAsp396, the main chain carbonyl oxygen of Arg398, and O1 andO2 of Glu400.

In form II-free, one additional calcium ion was identifiedbetween Glu60 of chain B and Asn248 of chain A of an adjacentasymmetric unit (data not shown). Because no form II crystalswere observed when calcium chloride was omitted from thecrystallization buffer, the latter calcium ion may promote crys-tallization by tightening the interaction between the two TLGTmolecules.

Catalytic Residues—An acarbose-TLGT complex was ob-tained by soaking a form II crystal in a buffer containingacarbose. The structure of the complex was determined at 2.4 Åresolution and refined to an R-factor of 19.8% (Table I). Elec-tron density corresponding to an intact acarbose molecule wasclearly observed for the active site of chain A (Fig. 4a). Asexpected, the acarbose molecule bound to subsites �1 to �3,where the acarviosine moiety, the inhibitory disaccharidegroup of acarbose, occupied subsites �1 and �1 and the malt-ose moiety occupied subsites �2 and �3 (Fig. 5a). The nomen-clature for the subsites is according to Davies et al. (29). This isthe inhibitory binding mode observed for most of the structuresof �-amylase family enzymes in complex with acarbose (30).3 H. Imamura and H. Matsuzawa, unpublished results.

FIG. 3. Schematic representations of the catalytic domain of families 57 and 38. a, (�/�)7 barrel core of domain I of TLGT. b, domain Iof TLGT. The two catalytic residues, Glu123 and Asp214, are represented as ball-and-stick models. Residues 259–321 of TLGT are showntransparently for clear understanding. �-Strands are numbered sequentially from the N termini. c, catalytic domain of D. melanogaster�-mannosidase II (Protein Data Bank code 1HTY). The catalytic nucleophile, Asp204, is represented as ball-and-stick model. �-Strands in thisdomain are numbered sequentially from the N termini.

Crystal Structure of Glycoside Hydrolase Family 57 19381

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Enzymatic hydrolysis of glycosidic linkages can be classifiedinto two major types according to the anomeric configuration ofthe product, retaining and inverting, and in both cases the

catalytic residues are typically two carboxylates (31). In retain-ing glycoside hydrolases, such as TLGT (2, 15), one residue actsas a nucleophile and the other as an acid/base catalyst. Glu123

FIG. 4. Bound ligands in the form II complex of TLGT. Stereoviews of the refined models of acarbose in chain A (a) and maltose in chainB (b) are presented, together with Fo � Fc electron density maps contoured at 4 �. Each ligand was omitted from the phase calculation. c, acarboseand the two catalytic residues. a and b were generated using XtalView (26) and Raster3d (45).

Crystal Structure of Glycoside Hydrolase Family 5719382

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

is close (3.15 Å) to the C1 atom of the valienamine moiety atsubsite �1, allowing nucleophilic attack (Fig. 4c), which isconsistent with the results of a cross-linking study that dem-onstrated Glu123 to be a catalytic nucleophile (15). It is consid-ered most likely that Asp214 is the acid/base catalyst of TLGT,because O�2 of Asp214 is only 2.96 Å away from the amidegroup of the valienamine moiety (this amide group is replacedby a glucosidic oxygen in a native substrate) (Fig. 4c), and thereis no other acidic residue nearby. The average distances be-tween all four pairs of O atoms of Glu123 and Asp214 (6.72 and6.97 Å in the acarbose-free and acarbose complex structures,respectively) are in appropriate range for retaining enzymes(32). We generated the D214N mutant, in which Asp214 wasreplaced by Asn. The specific activity of the D214N mutant(0.0016 units/mg) was decreased about 10,000-fold as comparedwith that of the wild-type enzyme (17.7 units/mg). These re-sults indicate that Asp214 is the acid/base catalyst.

Subsite Structure—Fig. 5a shows interactions between acar-bose and the protein. The valienamine moiety at subsite �1 is

within hydrogen bonding distance of His11 and Asp354 andinteracts with His13, Glu216, and Trp357 via water molecules.The 6-deoxyglucose moiety is fixed at subsite �1 with Arg124,Asp213, and Asp214 through hydrogen bonds, with Tyr272

through aromatic stacking, and with Trp221 through a hydro-phobic interaction. The glucose moiety at subsite �2 is boundto Arg182 and Asp213 though hydrogen bonds, to Tyr183 throughan aromatic stacking interaction, and to Trp221 through a hy-drophobic interaction. The glucose moiety at subsite �3 seemsto be more flexible, because it only exhibits a stacking interac-tion with Phe187 and a hydrophobic interaction with Trp221, i.e.no hydrogen bonding interaction with the protein.

In contrast to chain A, acarbose did not bind at the active siteof chain B of the form II complex, and only a Tris molecule wasidentified there (data not shown). Tris was also found at theactive site of chain B of the form II-free (Fig. 1a). Althoughacarbose did not bind to the active site, we found electrondensity corresponding to disaccharide, which seems to be malt-ose, at the edge of the active site cleft of chain B (Fig. 4b). Thereducing end of maltose is �14.5 Å apart from subsite �1,suggesting that this binding site corresponds to subsites �5and �6. The glucose moiety at subsite �5 is bound to His368

through a hydrogen bond and to Phe19 through a hydrophobicinteraction (Fig. 5b). The glucose moiety at subsite �6 is withinhydrogen bonding distance of Arg371 and exhibits hydrophobicinteractions with Phe19 and Tyr601 (Fig. 5b). In chain A, malt-ose did not bind to this site, because this region was involved inthe crystal contact. From the complex structure with acarbose,TLGT was revealed to possess at least nine subsites, �6 to �3,although the residues forming subsites �4 to �2 could not bedetermined in this study.

Unexpectedly, a glucose molecule was found at subsite �1 ofform I-Se (data not shown). Glucose interacted with Arg124,Asp214, Asp213, and Tyr272. Interactions between glucose andthese residues in form I-Se are the same as those observedbetween the �1 glucose moiety of acarbose and the correspond-ing residues in the form II complex. Glucose was probablyderived from trehalose reagent, which was used as a cryopro-tectant and contained a trace amount of glucose.

The active site cleft of TLGT is tunnel-like in shape, asevidenced by the three lids that cover the cleft (Figs. 6 and 7).The first lid (lid 1, residues 220–224) protrudes from the (�/�)7barrel (Fig. 2). The second (lid 2, residues 358–363) and third(lid 3, 627–630) lids protrude from the three-helix bundle anddomain II, respectively (Fig. 2). Upon binding of acarbose, theconformations of lids 2 and 3 change significantly (Fig. 6). Inthe absence of acarbose, the side chains of Val360 and Phe361

are directed toward the active site cleft. The bound acarbosecollides with the side chains of Val360 and Phe361, leading tomovement of lid 2. This movement induces a large movement oflid 3 (Fig. 6). For example, the C� of Ser627 and Glu628 move 4.2and 6.2 Å, respectively. In addition to the movements of thesetwo regions, the 1 axis of Tyr183 and the 2 axis of Phe187

rotated to interact with the pyranose rings of the maltosemoiety via hydrophobic stacking at subsites �2 and �3, respec-tively (Fig. 6).

Subunit Interface—Although the TLGT gene encodes a 78-kDa polypeptide, purified TLGT is eluted as a 168-kDa proteinin gel filtration chromatography using Superdex 200 column(Amersham Biosciences) (data not shown). This indicates thatTLGT is a homodimer, as previously suggested by Xavier et al.(1). In a form II crystal, two TLGT monomers, which are in anasymmetric unit, interact via the same surface of domain I,because there is a pseudo-2-fold axis between them (Fig. 1a). Ina form I crystal, two TLGT monomers, which are in adjacentasymmetric units associated through a 2-fold axis, interact in

FIG. 5. Schematic drawing of subsite structures of TLGT. a,interactions between acarbose and residues at subsites �1 to �3. b,interactions between maltose and residues at subsites �6 and �5.

Crystal Structure of Glycoside Hydrolase Family 57 19383

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

the same manner (data not shown). This observation indicatesthat the two TLGT monomers interact in the same manner asobserved in these crystals. Two hydrogen bond networks and ahydrophobic patch form the primary contribution to the inter-actions between the two subunits (Fig. 8a). One hydrogen bondnetwork is constructed from two water molecules and six res-idues: Glu166 and Tyr266 from one monomer, and Ala91, Lys314,Asn324, and Lys328 from the other (Fig. 8a). The hydrophobicpatch is formed by Leu287, Phe288, Phe291, Leu295, Tyr304,Phe307, and Val308 (Fig. 8b). In particular, Leu295 and Val308

are conserved (Fig. 2) and in contact with their counterparts inthe other monomer (Fig. 8b), which suggests that they play acentral role in the hydrophobic contact in this region. Theproportion of the buried surface area (�1500 Å2) is largeenough for dimerization, compared with the total molecularsurface of a monomer (�22300 Å2). Oligomerization is knownto be one of the strategies by which proteins acquire thermo-stability (33). Because the dimer interface is located on the

opposite side of the active site, dimerization seems to contrib-ute to the thermostability rather than the activity includingamylose cyclization.

DISCUSSION

One of the interesting features of TLGT is the (�/�)7 barrelfold, which forms the core of domain I (Fig. 3a). Proteins havingthe (�/�)7 barrel fold are very rare (34), although there is alarge number of �-barrel proteins. Cellobiohydrolase (35) andendoglucanase (36), which both belong to glycoside hydrolasefamily 6, are two examples of (�/�)7 barrel proteins. The central�-sheet of the barrel of family 6 proteins is not completelyclosed, because there is only one hydrogen bond between themain chains of the first (�1) and last (�7) strands. The barrel ofTLGT has a more open conformation, because there are twosuch “nonclosures” in the barrel of TLGT: between �3 and �4and between �7 and �8.

Previously, we found that the nucleophilic residues of TLGTand class II �-mannosidases (glycoside hydrolase family 38) arelocated at the same position in the amino acid sequences de-spite their low sequence similarity (15). When the two familieswere compared at the three-dimensional level, striking struc-tural similarities were observed in the catalytic domains ofTLGT (Fig. 3b) and Drosophila melanogaster Golgi �-mannosi-dase II (37) (Fig. 3c): a �-sheet core surrounded by �-helices, athree-helix bundle, and a catalytic nucleophile at the fourth �-�loop. These findings strongly suggest that the two families have

FIG. 6. Stereoview of the movements at the active site upon binding of acarbose on TLGT. The native (chain A of form II-free) andcomplex (chain A of form II-complex) are colored orange and green, respectively.

FIG. 7. Surface representation around the active site of TLGT.The molecular surface of chain A of the form II-complex, a boundacarbose, and a modeled maltotetradecaose are presented. The C atomsof acarbose and maltotetradecaose are colored white and yellow, respec-tively. The oxygen and nitrogen atoms are colored red and blue, respec-tively. Lid 2 is concealed behind lid 3. The figures were prepared withthe programs SPOCK (49), Molscript (44), and Raster3d (45).

FIG. 8. Dimer interfaces of TLGT. A hydrogen bond network (a)and a hydrophobic patch (b) in form II-free. Chain A is presented inblue, and chain B is in green. The broken orange lines indicate hydrogenbonds. The calcium ion is also represented as an yellow space-filledmodel.

Crystal Structure of Glycoside Hydrolase Family 5719384

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

evolved from a common ancestor. Despite the similarities, how-ever, there are also significant structural differences betweenthe two enzymes. The most important one is that the catalyticdomain of �-mannosidase II does not have a barrel structure,because there is no hydrogen bond that connects the two�-strands (7 and 8 in Fig. 3c) corresponding to �7 and �8 ofTLGT. The �� fold in �-mannosidase II may be a result ofprotein evolution from the (�/�)7 barrel fold found in TLGT andvice versa. This difference is interesting from the standpoint ofthe evolution of protein folding.

In the crystal structures of amylase-acarbose complexes,acarbose was often found in a modified form, because of thetransglycosylation activity of amylase (see Ref. 30). However,it is clear that acarbose bound to the active site of TLGT(chain A in the form II-complex) is an intact one, because theelectron density indicates that the sugar ring at the subsite�1 is somewhat flattened and that O-6 is not present at thesubsite �1. Why maltose bound to the subsite �5 and �6 ofchain B in the form II-complex is uncertain. Because maltosewas not detected in the acarbose reagent when analyzed onthin layer chromatography (data not shown), acarviosinemoiety of acarbose may have flexible conformation, and onlymaltose moiety may be observed as a clear electron density.Additional electron density at the nonreducing end seen inFig. 4b can be explained by this idea. Such a case has beenalso reported in the crystal structure of Bacillus circulansxylanase complexed with xylotetraose, in which electron den-sity for only two xylose residues was observed (38). It is alsounclear why acarbose did not bind to the active site of chainB in the form II crystal. Although TLGT has a larger Ki value(�0.6 � 10�3 M) for acarbose than �-amylase family enzymes(0.6 � 10�7 to 0.8 � 10�4 M) (39), the acarbose concentration(10 mM) in the soaking solution is sufficient. One possibilityis that conformational changes in chain B prevent the bind-ing of acarbose. When compared with form I and chain A ofform II, the slight movements of �8, �9, �10, �11, �13, �14,3105, �22, and �23, and the destruction of some ion pairs inchain B were observed (data not shown), probably because ofcrystal packing. These movements seem to change chain Binto an inactive form that cannot bind the substrate.

There are several kinds of 4-�-glucanotransferases, includ-ing cyclodextrin glucanotransferase (CGTase), many of whichproduce CAs (6). The minimum ring size of CAs varies with theenzyme. The smallest products of CGTase are CAs with 6–8glucose units. Potato 4-�-glucanotransferase (3) and T. aquati-cus amylomaltase (40) produce CAs with 17 or more and 22 ormore glucose units, respectively. Sixteen is the minimum ringsize of CAs produced by TLGT2 and Thermococcus kodakaraen-sis 4-�-glucanotransferase (also a family 57 enzyme) (41). Whatdetermines the minimum ring size of CAs produced by theseenzymes? It has been proposed that in the case of large CAformation by T. aquaticus amylomaltase, the product wrapsaround the bulky loop, which partially covers the active site(30). In addition to the loop structure, it was thought that theexistence of a second ligand-binding site apart from the cata-lytic site also contributed to the large CA formation (30). Incontrast, there is no such large steric hindrance in CGTase.Although T. aquaticus amylomaltase and TLGT have differentstructures, they seem to have adapted a similar strategy toproduce large CAs. In the form II-complex of TLGT, acarbosepartially wraps around lid 1 (Fig. 7). Lid 1 seems to prevent theformation of small CAs because of its considerable steric hin-drance. Because amylose forms a helical structure in whichglucose units are connected through O-3–O-2� hydrogen bonds,�-1,4-linked glucose polymers tend to become circularized in anaqueous solution. In fact, the reorganization of O-3–O-2� hy-

drogen bonds is proposed to be one of the forces in the circu-larization step responsible for the formation of cyclodextrins byCGTase (42). If there is no steric hindrance by the loop ob-served in TLGT and amylomaltase, smaller CAs might be pro-duced. When we included a polysaccharide chain in the struc-ture of chain A of the form II complex, a chain of 14 glucoseresidues was built into the structure, in addition to the tet-rasaccharide inhibitor at the catalytic site (Fig. 7). This poly-saccharide model and acarbose correspond to a CA with 18glucose units. Because TLGT predominantly produces a CAwith 18–20 glucose units,2 this model seems to represent themost common mode of CA binding in TLGT. When the mini-mum ring size of CA (16 glucose residues) is produced, someassumption need to be made. From this model, in addition to lid1, the large distance (�27.5 Å) between subsites �6 and �3,which is caused by the relatively extended structure of thecleft, coupled with the steric hindrance of the side chains ofPhe19 and Trp21 also seem to prevent the formation of smallCAs in TLGT (Fig. 7). When amylose circularizes, the nonre-ducing end of a polysaccharide chain must travel a considerabledistance to the acceptor site in the active site. It has beenproposed that the movement of some hydrophobic residuesassists in the circularization of a substrate in CGTase (43). Thestructural movements observed in TLGT (Fig. 6) may alsofacilitate the circularization of amylose.

Acknowledgments—We are indebted to the staff of SPring-8 and thePhoton Factory.

REFERENCES

1. Xavier, K. B., Peist, R., Kossmann, M., Boos, W., and Santos, H. (1999) J.Bacteriol. 181, 3358–3367

2. Jeon, B. S., Taguchi, H., Sakai, H., Ohshima, T., Wakagi, T., and Matsuzawa,H. (1997) Eur. J. Biochem. 248, 171–178

3. Takaha, T., Yanase, M., Takata, H., Okada, S., and Smith, S. M. (1996) J. Biol.Chem. 271, 2902–2908

4. Kitamura, S., Isuda, H., Shimada, J., Takada, T., Takaha, T., Okada, S.,Mimura, M., and Kajiwara, K. (1997) Carbohydr. Res. 304, 303–314

5. Kitamura, S., Nakatani, K., Takaha, T., and Okada, S. (1999) Macromol.Rapid Commun. 20, 612–615

6. Takaha, T., and Smith, S. M. (1999) Biotechnol. Genet. Eng. Rev. 16, 257–2807. Machida, S., Ogawa, S., Xiaohua, S., Takaha, T., Fujii, K., and Hayashi, K.

(2000) FEBS Lett. 486, 131–1358. Henrissat, B., and Bairoch, A. (1996) Biochem. J. 316, 695–6969. Coutinho, P. M., and Henrissat, B. (1999) Carbohydrate-Active Enzymes

server, afmb.cnrs-mrs.fr/�cazy/CAZY/index.html10. Przylas, I., Tomoo, K., Terada, Y., Takaha, T., Fujii, K., Saenger, W., and

Strater, N. (2000) J. Mol. Biol. 296, 873–88611. Matsuura, Y., Kusunoki, M., Harada, W., and Kakudo, M. (1984) J. Biochem.

(Tokyo) 95, 697–70212. Klein, C., and Schulz, G. E. (1991) J. Mol. Biol. 217, 737–75013. Katsuya, Y., Mezaki, Y., Kubota, M., and Matsuura, Y. (1998) J. Mol. Biol. 281,

885–89714. Svensson, B. (1994) Plant Mol. Biol. 25, 141–15715. Imamura, H., Fushinobu, S., Jeon, B. S., Wakagi, T., and Matsuzawa, H.

(2001) Biochemistry 40, 12400–1240616. Imamura, H., Fushinobu, S., Yamamoto, M., Kumasaka, T., Wakagi, T., and

Matsuzawa, H. (2001) J. Appl. Glycosci. 48, 171–17517. Otwinowski, Z., and Minor, W. (1997) Methods Enzymol. 276, 307–32618. Leslie, A. G. W. (1992) Joint CCP4 and ESF-EAMCB Newsletter on Protein

Crystallography 2619. Terwilliger, T. C., and Berendzen, J. (1999) Acta Crystallogr. Sect. D Biol.

Crystallogr. 55, 849–86120. Cowtan, K. (1994) Joint CCP4 and ESF-EAMCB Newsletter on Protein Crys-

tallography 31, 34–3821. Jones, T. A., Zou, J. Y., Cowan, S. W., and Kjeldgaard, M. (1991) Acta Crys-

tallogr. Sect. A 47, 110–11922. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-

Kunstleve, R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S., Read,R. J., Rice, L. M., Simonson, T., and Warren, G. L. (1998) Acta Crystallogr.Sect. D Biol. Crystallogr. 54, 905–921

23. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993)J. Appl. Crystallogr. 26, 283–291

24. Holm, L., and Sander, C. (1993) J. Mol. Biol. 233, 123–13825. Kleywegt, G. J. (1999) Acta Crystallogr. Sect. D Biol. Crystallogr. 55,

1878–188426. McRee, D. E. (1999) J. Struct. Biol. 125, 156–16527. Imamura, H., Jeon, B. S., Wakagi, T., and Matsuzawa, H. (1999) FEBS Lett.

457, 393–39628. Juers, D. H., Huber, R. E., and Matthews, B. W. (1999) Protein Sci. 8, 122–13629. Davies, G. J., Wilson, K. S., and Henrissat, B. (1997) Biochem. J. 321, 557–55930. Przylas, I., Terada, Y., Fujii, K., Takaha, T., Saenger, W., and Strater, N.

(2000) Eur. J. Biochem. 267, 6903–6913

Crystal Structure of Glycoside Hydrolase Family 57 19385

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

31. McCarter, J. D., and Withers, S. G. (1994) Curr. Opin. Struct. Biol. 4, 885–89232. Davies, G., and Henrissat, B. (1995) Structure 3, 853–85933. Vieille, C., and Zeikus, G. J. (2001) Microbiol. Mol. Biol. Rev. 65, 1–4334. Nagano, N., Hutchinson, E. G., and Thornton, J. M. (1999) Protein Sci. 8,

2072–208435. Rouvinen, J., Bergfors, T., Teeri, T., Knowles, J. K., and Jones, T. A. (1990)

Science 249, 380–38636. Spezio, M., Wilson, D. B., and Karplus, P. A. (1993) Biochemistry 32,

9906–991637. van den Elsen, J. M., Kuntz, D. A., and Rose, D. R. (2001) EMBO J. 20,

3008–301738. Wakarchuk, W. W., Campbell, R. L., Sung, W. L., Davoodi, J., and Yaguchi, M.

(1994) Protein Sci. 3, 467–47539. Kim, M. J., Lee, S. B., Lee, H. S., Lee, S. Y., Baek, J. S., Kim, D., T. W., M.,

Robyt, J. F., and Park, K. H. (1999) Arch. Biochem. Biophys. 371, 277–28340. Terada, Y., Fujii, K., Takaha, T., and Okada, S. (1999) Appl. Environ. Micro-

biol. 65, 910–91541. Tachibana, Y., Takaha, T., Fujiwara, S., Takagi, M., and Imanaka, T. (2000)

J. Biosci. Bioeng. 90, 406–40942. Uitdehaag, J. C., van der Veen, B. A., Dijkhuizen, L., Elber, R., and Dijkstra,

B. W. (2001) Proteins 43, 327–33543. Uitdehaag, J. C., Kalk, K. H., van Der Veen, B. A., Dijkhuizen, L., and

Dijkstra, B. W. (1999) J. Biol. Chem. 274, 34868–3487644. Kraulis, P. J. (1991) J. Appl. Crystallogr. 24, 946–95045. Merrit, E. A., and Murphy, M. E. P. (1994) Acta Crystallogr. Sect. D Biol.

Crystallogr. 50, 869–87346. Thompson, J. D., Higgins, D. G., and Gibson, T. J. (1994) Nucleic Acids Res. 22,

4673–468047. Kabsch, W., and Sander, C. (1983) Biopolymers 22, 2577–263748. Gouet, P., Courcelle, E., Stuart, D. I., and Metoz, F. (1999) Bioinformatics 15,

305–30849. Christopher, J. A., and Baldwin, T. O. (1998) J. Mol. Graph. Model. 16, 285

Crystal Structure of Glycoside Hydrolase Family 5719386

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from

Jeon, Takayoshi Wakagi and Hiroshi MatsuzawaHiromi Imamura, Shinya Fushinobu, Masaki Yamamoto, Takashi Kumasaka, Beong-Sam

Complex with an Inhibitor and ItsThermococcus litoralis-Glucanotransferase from αCrystal Structures of 4-

doi: 10.1074/jbc.M213134200 originally published online March 4, 20032003, 278:19378-19386.J. Biol. Chem. 

  10.1074/jbc.M213134200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/278/21/19378.full.html#ref-list-1

This article cites 47 references, 8 of which can be accessed free at

by guest on March 17, 2018

http://ww

w.jbc.org/

Dow

nloaded from