View metadata, citation and similar papers at core.ac.uk brought to you by CORE Researchprovided Articleby Elsevier -179 Publisher Connector

Three-dimensional structure of a barley β-D-glucan exohydrolase, a family 3 glycosyl Joseph N Varghese1*, Maria Hrmova2 and Geoffrey B Fincher2*

Background: Cell walls of the starchy endosperm and young vegetative tissues Addresses: 1Biomolecular Research Institute, 343 of barley (Hordeum vulgare) contain high levels of (1→3,1→4)-β-D-glucans. Royal Parade, Parkville, Victoria 3052, Australia and 2Department of Plant Science, University of The (1→3,1→4)-β-D-glucans are hydrolysed during wall degradation in Adelaide, Waite Campus, South Australia 5064, germinated grain and during wall loosening in elongating coleoptiles. These key Australia. processes of plant development are mediated by several polysaccharide endohydrolases and exohydrolases. *Corresponding authors. E-mail: [email protected] [email protected] Results: The three-dimensional structure of barley β-D-glucan exohydrolase isoenzyme ExoI has been determined by X-ray crystallography. This is the first Key words: barley β-D-glucan exohydrolase reported structure of a family 3 glycosyl hydrolase. The is a two-domain, isoenzyme ExoI, β-D-glucan, family 3 glycosyl globular of 605 amino acid residues and is N-glycosylated at three sites. hydrolase, X-ray crystallography α β The first 357 residues constitute an ( / )8 TIM-barrel domain. The second Received: 11 September 1998 domain consists of residues 374–559 arranged in a six-stranded β sandwich, Revisions requested: 13 October 1998 which contains a β sheet of five parallel β strands and one antiparallel β strand, Revisions received: 19 November 1998 with three α helices on either side of the sheet. A glucose moiety is observed in Accepted: 15 December 1998 a pocket at the interface of the two domains, where Asp285 and Glu491 are Published: 29 January 1999 believed to be involved in catalysis. Structure February 1999, 7:179–190 Conclusions: The pocket at the interface of the two domains is probably the http://biomednet.com/elecref/0969212600700179 of the enzyme. Because amino acid residues that line this active-site © Elsevier Science Ltd ISSN 0969-2126 pocket arise from both domains, activity could be regulated through the spatial disposition of the domains. Furthermore, there are sites on the second domain that may bind carbohydrate, as suggested by previously published kinetic data indicating that, in addition to the catalytic site, the enzyme has a second specific for (1→3,1→4)-β-D-glucans.

Introduction the second site does not appear to bind (1→3)-β-glucans Two basic β-glucan exohydrolases with apparent molecu- or other, smaller, substrates [6]. lar masses of 69 kDa and 71 kDa have been purified from extracts of eight day germinated barley seedlings and are Glycosyl have been classified into a number of designated isoenzymes ExoI and ExoII, respectively [1]. families on the basis of amino acid sequence alignments β-Glucan exohydrolases have also been detected in elon- and hydrophobic cluster analyses [7]. It is becoming gating coleoptiles of barley [2,3] and maize [4], and in apparent that members of each family have similar three- suspension-cultured soybean cells [5]. The barley dimensional (3D) conformations. For example, in previous preferentially hydrolyse the (1→3)-β-glucan, work from our laboratories, it was shown that barley laminarin, but also hydrolyse (1→3,1→4)-glucans, (1→3)-β-glucan endohydrolases and (1→3,1→4)-β-glucan (1→3,1→6)-β-glucans and a range of glucooligosaccha- endohydrolases, both of which are classified in family 17, β α β rides and aryl- -glucosides [6]. Both isoforms release have ( / )8 barrel folds that are very similar; the root mean single glucosyl residues from the non-reducing ends of square deviation (rmsd) in Cα positions is 0.65 Å over 278 substrates, and proton nuclear magnetic resonance amino acid residues [8]. Sequence alignments and (NMR) shows that the anomeric configuration of released hydrophobic-cluster analyses show that the barley glucose is retained after hydrolysis [1]. Detailed kinetic β-glucan exohydrolases are members of the family 3 glyco- analyses showed that resultant Hill plots had gradients of syl hydrolases [6,7], but no 3D structures are currently 1.0 both for laminarin and for 4-nitrophenyl β-glucoside, available for this group of enzymes. However, we recently but 2.0 for the cell wall (1→3,1→4)-β-glucan from barley crystallized β-glucan exohydrolase isoenzyme ExoI from [6]. The gradient of 2.0 indicated positive of barley using vapour diffusion in the presence of ammo- binding of (1→3,1→4)-β-glucan to the enzyme and led to nium sulphate and polyethylene glycol [9]. Crystal dimen- the conclusion that barley β-glucan exohydrolases have sions of up to 0.8 mm × 0.4 mm × 0.6 mm were observed another binding site for cell wall (1→3,1→4)-β-glucans; and the crystals diffracted to at least 2.2 Å resolution [9]. 180 Structure 1999, Vol 7 No 2

In the present work, the 3D structure of the barley Table 1 β-glucan exohydrolase isoenzyme ExoI has been deter- Data collection and MIR phasing statistics. mined by X-ray crystallography and this therefore repre- sents the first structure reported for a family 3 glycosyl Native PIP MHG hydrolase and for a β-glucan glucohydrolase. Data collection Resolution (Å) 2.2 2.2 2.6 Results and discussion Completeness* 78 (32) 75 (33) 70 (38) The 3D structure of barley β-glucan exohydrolase isoen- Multiplicity 3.7 2.5 3.5 zyme ExoI, a family 3 glycosyl hydrolase, has now been 〈I/σ(I)〉* 20.1 (2.8) 19.8 (2.4) 11.1 (2.1) solved at 2.2 Å resolution by X-ray crystallography, using Rmerge* 5.4 (14.2) 8.2 (20.5) 9.3 (17.2) Wilson B 26.7 41.8 39.7 multiple isomorphous replacement (MIR) techniques MIR phasing (Table 1). The heavy atom derived phases were accurate Resolution (Å) 90–2.3 90–2.6 enough to build the entire molecule unambiguously and Sites per molecule 4† 2 simulated annealing was not necessary as a means of refin- Total occupation 1.84 1.7 Phasing power‡§ 0.9 (1.2) [0.7] 1.0 (1.1) [1.1] ing the structure. ¶§ Rcullis 0.77 (0.79) [0.76] 0.87 (0.89) [0.61] The complete amino acid sequence of the enzyme has been *Numbers in parentheses are the values in the high resolution shell, Σ 〈 〉 Σ deduced from the nucleotide sequence of a near full-length where Rmerge = (Ii– I ) Ii summed over all independent reflections. † § cDNA (AJ Harvey and GBF, unpublished results) and from Each site was split into pairs. The three numbers represent summation over: centric (acentric) [acentric anomalous] reflections. ‡Root mean amino acid sequences of tryptic peptides generated from the √ Σ 2 square (rms) fh/residual = ( ( Fderi–FPH) ), where fh and FPH are the purified enzyme [1]. The mature enzyme contains 605 calculated heavy-atom and derivative structure factors, respectively. ¶ Σ Σ amino acid residues, and its sequence has been lodged in Rcullis = ||fh|–(|Fderi|–|FNati|)|/ ||Fderi|–|FNati||, where Fderi and FNati are the the GenBank and EMBL databases under accession observed derivative and native structure factors, respectively. number AF102868. The X-ray data were collected from two crystals [9] and electron density was continuous for the entire Cα polypeptide chain, except that the last three amino However, the (1→3)-β-glucan exohydrolase from the acid residues at the C terminus could not be observed in the human pathogenic fungus Candida albicans, which is a electron-density map. The electron densities of most amino member of the family 5 group of glycosyl hydrolases [12,13], acid sidechains were well-defined and identifiable residues will probably exhibit significant structural differences. corresponded in all cases to those deduced from the cDNA sequence. In several instances, the electron density of Domain structure sidechains was used to confirm that the nucleotide sequence Barley β-glucan exohydrolase isoenzyme ExoI is a glyco- was translated in the correct reading frame; the cDNA is par- sylated protein consisting of two distinct domains, which ticularly GC-rich and sequencing compressions were are connected by a helix-like strand of 16 residues common (AJ Harvey and GBF, unpublished results). (Figures 1 and 2). The first domain (residues 1–357) forms α β an ( / )8 barrel and the second domain (residues 374–559) β α β β Two isoforms of the barley -glucan exohydrolase family forms an ( / )6 sheet, consisting of five parallel strands have been purified and characterized [1,2], and Southern and one antiparallel strand, with three helices on either hybridization analyses show that there are approximately side of it. The C terminus forms a long antiparallel loop eight members of the barley β-glucan exohydrolase gene (Figure 1). Overall, the molecule is an elongated structure family (AJ Harvey and GBF, unpublished results). Isoen- of about 82 Å in length. It is 55 Å in width and 35 Å thick α β zymes ExoI and ExoII exhibit 72% sequence identity at the at the ( / )8-barrel domain, but it tapers to 30 Å in width α β amino acid level (AJ Harvey and GBF, unpublished and 26 Å in thickness at the ( / )6-sheet domain. results). We predict that the 3D structures of isoenzyme ExoII and other barley isoforms (N Sakurai, personal com- Domain 1 is a short cylindrical structure of approximately munication), together with related β-glucan exohydrolases 55 Å in diameter and 35 Å in thickness and, when viewed from maize [4] and higher plants generally [5], will be very down the cylinder axis of the disk with domain 2 below β α β similar to that described here for barley -glucan exohydro- (Figure 1), the pseudo-eightfold ( / )8 barrel axis points lase isoenzyme ExoI. Furthermore, other family 3 glycosyl up and out at an angle of about 45° to the right of the hydrolases, such as the β-glucosidase from nasturtium cylinder axis. Domain 2 is wedge-shaped and is about 45 Å (Tropaeolum majus) [10] and the β-glucan glucohydrolase long, 26 Å wide and 26 Å thick. Domain 2 lies in close from Nicotiana tabacum (N Koizumi, unpublished results, contact with the lower cylinder surface of domain 1 and GenBank accession number G3582436) and Pseudomonas protrudes downwards at an angle of 30°, with the upper fluorescens [11], which have sequence identities relative to loops of domain 2 bordering the barrel axis of domain 1. the barley β-glucan exohydrolase of 71%, 72% and 44%, The upper surface of the molecule is quite flat, apart from respectively, are likely to adopt similar protein folds. the protrusion of carbohydrate attached to Asn498. Research Article Barley β-glucan exohydrolase Varghese, Hrmova and Fincher 181

Figure 1

A MOLSCRIPT [50] stereo representation of the overall structure of barley β-glucan α β exohydrolase isoenzyme ExoI. (a) The ( / )8 barrel that comprises domain 1 is represented by red α helices and green β strands, with α β purple coils connecting them. The ( / )6 sheet that constitutes domain 2 is represented by orange α helices and light blue β strands, with yellow coils connecting them. The C terminus is represented by light green antiparallel sheets. The yellow and green balls represent the N terminus and C terminus, respectively, and the two light blue balls and the three dark blue balls represent the two putative catalytic amino acid residues and the three N-glycosylation sites, respectively. Helices are labelled with the uppercase letter of the β strands (lower case) they follow. Positive and negative subscripts on helices represent successive helices following or preceding the corresponding strands, respectively. The bound glucose moiety is shown in the active- site region at the interface of the two domains. (b) A Cα trace of the protein in the same orientation, with every tenth residue numbered. The colour scheme is as in (a). Transparent GRASP [51] molecular surfaces of the two domains are overlayed on the α β figure, with the ( / )8 barrel in pink, and the α β ( / )6 sheet in green. The asparagine-linked carbohydrate sites are represented by orange spheres, and blue, red and purple spheres represent N-acetyl-D-glucosaminyl, mannosyl and fucosyl residues, respectively.

The two domains are very closely associated and a shallow The four remaining cysteine residues in the enzyme groove runs along the upper edge of the interface of the (Cys81, Cys 146, Cys204, Cys427) are unpaired and two of domains. In the central region of this groove is a small these (Cys146 and Cys427) are bound to mercury ions in pocket, and there is a cleft between the barrel domain and the crystal. Six cis proline residues were observed at posi- the C-terminal loop, toward the right-hand side (Figure 1). tions 146, 295, 318, 405, 504 and 579 in the amino acid A glucose moiety is bound noncovalently, but stably, in sequence, and cis peptide bonds were found before the pocket. The pocket sits above the pseudo-eightfold residues His207 and Val209. α β axis of the ( / )8 barrel of domain 1. α β The ( / )8-barrel domain There are disulphide bridges connecting Cys151 and The N terminus begins at the top of the molecule when Cys159 in domain 1, and Cys513 and Cys518 in domain 2. viewed in the orientation shown in Figure 1 (see Figure 2 182 Structure 1999, Vol 7 No 2

Figure 2

11020304050607080 DYVLYKDATKP VEDRVADLLGRMTLAEKI GQMTQI ERLVATP DVLRDNF I GS LLS GGGS VP RKGATAKEWQDMVDGFQKA A-2 A-1 a A b B

90 100 110 120 130 140 150 160 CMSTRLGI P MI YGI DAVHGQNNVYGATI F P HNVGLGATRDP YLVKRI GEATALEVRATGI QYAF AP CI AVCRDP RWGRCY cC1C2d cis 170 180 190 200 210 220 230 240 ESYSEDRRI VQSMTELIPGLQGDVP KDF TS GMPF VAGKNKVAACAKHF VGDGGTVDGI NENNTI I NREGLMNI HMPAYKN DeE1 E2 cis cis CHO 250 260 270 280 290 300 310 320 AMDKGVSTVMI S YS S WNGVKMHANQDLVTGYLKDTLKF KGF VI S DWEGI DRI TTP AGS DYS YS VKAS I LAGLDMI MVPNK fFgG1 G2 h Catalytic cis cis 330 340 350 360 370 380 390 400 YQQF I S I LTGHVNGGVI P MSRI DDAVTRI LRVKF TMGLF ENP YADP AMAEQLGKQEHRDLAREAARKS LVLLKNGKTS TD H1 H2 H3 i

410 420 430 440 450 460 470 480 AP LLP LP KKAP KI LVAGS HADNLGYQCGGWTI EWQGDTGRTTVGTTI LEAVKAAVDP S TVVVF AENP DAEF VKS GGF S YA jJ1 J2k K cis 490 500 510 520 530 540 550 560 I VAVGEHP YTETKGDNLNLTI P EP GLS TVQAVCGGVRCATVLI S GRP VVVQP LLAAS DALVAAWLP GS EGQGVTDALF GD l Lm MnN Catalytic CHO cis 570 580 590 600 F GF TGRLP RTWFKS VDQLP MNVGDAHYDPLF RLGYGLTTNATKKY

cis CHO Structure

An ALSCRIPT [52] representation of the sequence of barley β-glucan and arrows (β strands), and lines designate the connecting coils. The exohydrolase isoenzyme ExoI, colour-coded and labelled as described putative catalytic acids (white up-arrows), N-glycosylation sites (black in Figure 1, to represent the secondary structural elements of the up-arrows; CHO represents carbohydrate), cis-proline residues (white three-dimensional fold of the molecule. The secondary structural down-arrows) and non-proline cis peptide bonds (black down-arrows) elements are displayed below the sequence as cylinders (α helices) are also shown.

β α β for residue assignments). The long axis of the enzyme is helices (G1, G2); the last strand, h, completes the ( / )8 vertical in Figure 1 and the barrel axis points out of the barrel. Exiting β strand h at the top of the barrel is a long paper and 45° to the right. There are two helices, A–2 and helix H1, which runs up along the top of the molecule β A–1, near the N terminus before the first strand a of the before turning down at the top of the molecule as the long α β ( / )8 barrel is entered about halfway up the disk. The helix H2. The chain continues along the underside of the chain exits on the upper-middle surface of the disk and molecule in a helix-like, extended conformation that con- winds itself into an (α/β) pair, with helix A on the outside nects the first domain to the second domain (Figure 1). and β strand b towards the left side of the disk axis. This motif is repeated with helix B and β strand c, the β barrel Most of the interactions between the domains involve winding anti-clockwise. Exiting c is a long strand that hydrogen bonds or are ionic in nature. Interactions ends with a small helix C1 and turns back under the barrel between mainchain peptide atoms and sidechain groups β to enter strand d via a long helix C2. This loop and the predominate and there are a number of charged groups two helices are the main contacts with the second domain. (His98, His110, Arg526, Arg569) buried in the interface The polypeptide exits β strand d as an extended strand region. Hydrophobic interactions between the domains are and loop, before entering β strand e via helix D (now on restricted to the lower region of the interface (Figure 1). the right-hand side of the disk and above helix C2) and a α β long loop under the disk. This is repeated for the strand The ( / )6-sheet domain β exiting strand e, except that there are two closely placed The chain enters the second domain via a long helix H3, cis peptide bonds between Lys206 and His207 and which lies underneath the molecule as it is orientated in Phe208 and Val209 in the extended strand that enters a Figure 1. It subsequently forms the sixth (antiparallel) β β long helical segment (E1, E2) prior to entering strand f outer strand i (to the right of domain 2) of a six-stranded (see Figure 3 for an electron-density map of this region). β-sheet structure. The chain then loops across the bottom The typical (α/β) motif is repeated for helix F and of the molecule and enters the second β strand j, after β β strand g (Figure 1). Between strands g and h are two which, via the two helices J1 and J2 that make up the left Research Article Barley β-glucan exohydrolase Varghese, Hrmova and Fincher 183

Figure 3

A MOLSCRIPT [50] stereo representation of the region of amino acid residues 206–211 of domain 1 of the barley β-glucan exohydrolase isoenzyme ExoI. The cis peptide linkages between Lys206 and His207 and between Phe208 and Val209 are clearly observed in the electron density of the 2Fobs–Fcal map, and are represented by a green caged mesh contour (s = 2.1), where Fobs and Fcal are the observed and refined structure factors. Carbon, oxygen, sulphur and nitrogen atoms are represented by black, red, yellow and blue spheres, respectively.

side of domain 2, the chain enters β strand k at the bottom Asn498 and Asn600 appear to have extra electron density left-hand side of the domain. The chain exits strand k, emanating from their sidechain amide nitrogen atoms. which forms the outermost strand of the sheet, and forms These are the only potential N-glycosylation sites on the a helix K that runs down the bottom edge of the domain enzyme and no O-linked carbohydrate was observed. The and on the other side of the sheet, into β strand l. The molecular mass deduced from the cDNA sequence is chain then forms a long loop that approaches domain 1 65,390 Da. The overall molecular mass as determined by and, via helix L on the upper surface of the domain and matrix-assisted laser desorption time-of-flight (MALDI- pointing down the molecule, enters β strand m, which TOF) mass spectrometry was 68,592 Da (± 46 Da; data now runs parallel to the long axis of the molecule towards not shown). Thus, it can be calculated that N-linked car- domain 2. This structure is repeated for helix M and bohydrate constitutes approximately 4.7% by weight of β strand n, except than the loop connecting β strand m the glycosylated protein. and helix M is short. Strand n forms the final strand in a six-stranded β sheet consisting of strands k, j, l, m, n and i, A single N-acetyl-D-glucosamine (GlcNAc) residue where the first five strands are parallel and the sixth is attached to Asn221 can be discerned in the electron- antiparallel (Figure 1). Three helices K, L and M lie on density map, but no further monosaccharide residues can the underside of this sheet, and H3, J1, J2 and N are on the be seen at this position. Similarly, electron density at topside of the sheet (Figure 1). Asn600 reveals that this residue is glycosylated but the resolution is very poor and the structure of the carbohy- The C terminus drate could not be modelled. However, the data on the The final secondary structural element in the second carbohydrate attached to Asn498 are sufficiently clear to domain is helix N, which runs into two antiparallel strands enable the definition of several glycosyl residues (residues 560–602) that proceed up the right-hand side of (Figure 4). The oligosaccharide consists of a core structure the interface between domains 1 and 2 and are perpendic- of two (1→4)-β-linked GlcNAc residues, followed by a ular to the β-sheet structure (Figure 1). The antiparallel branching mannosyl residue. The (1→3)-β-linked branch strands form two pseudo β sheets that terminate at the of this mannosyl residue, which would project away from underside of the molecule (Figure 1). As mentioned the enzyme’s surface, is not observed, but the C6 atom is earlier, the last three residues at the C terminus are not linked to another mannosyl residue, which, in turn, is visible in the electron-density map, presumably because (1→2)-β-linked to a terminal GlcNAc residue. A (1→3)-α- of high thermal mobility in this region. It is also possible linked fucosyl residue is also linked to the GlcNAc that is that the residues have been removed before or during attached to Asn498 (Figure 4). There is a suggestion in the purification by carboxypeptidases that are present in the electron density of a (1→2)-linked xylosyl residue germinated barley grain [14]. The complete sequence attached to the branching mannosyl residue. The struc- overlayed by the secondary structure elements of ExoI is ture is represented diagrammatically as follows: shown on Figure 2. GlcNAcβ1→2Manα1→6Manβ1→4GlcNAcβ1→4GlcNAcβ1→N-Asn498 Carbohydrate structure 23 From the electron-density map, carbohydrate has been ↑↑ identified at three sites. Amino acid residues Asn221, Xylβ1(?) Fucα1 184 Structure 1999, Vol 7 No 2

Figure 4

A MOLSCRIPT [50] ball-and-stick stereo representation of the oligosaccharide chain (yellow bonds) that is linked to Asn498. The protein molecular surface (GRASP) [51] is in transparent green and the backbone-chain atoms are in standard colours. The Cα backbone is represented by a green ‘worm’ tube. Hydrogen-bond interactions within the oligosaccharide are represented by dashed lines. Carbon, oxygen and nitrogen atoms are represented by black, red and blue spheres, respectively, and water molecules are represented as larger red spheres.

where GlcNAc represents N-acetyl-D-glucosaminyl groups (Lys206 and His207, linked by a cis peptide bond) residues, Man are mannosyl residues, Xyl is a xylosyl interact with the C3 hydroxyl group of the glucose, and residue and Fuc is a fucosyl residue. This structure is very Arg158 interacts with the C2 hydroxyl group. Further- similar to the N-linked oligosaccharides that are associated more, this C2 hydroxyl interacts with Asp285, which is with many plant [15–17], including the barley located on the top surface of the pocket and which also (1→3,1→4)-β-glucan endohydrolase (enzyme classifica- forms a close contact (about 3 Å) with the C1 anomeric tion [EC] 3.2.1.73) isoenzyme EII [18]. Although it is not carbon atom of the glucopyranose ring. There is another possible at this stage to assign a function to the oligosac- acid (Glu491), located on the bottom surface of the charide chain on Asn498, it protrudes from an otherwise pocket, that lies on the other side of the glucose ring with flat molecular surface of domain 2 and could possibly act its Oε2 about 3.5 Å from the C1 carbon atom. The next to guide substrates into the binding site. closest acidic residue is Glu220, in which one of the car- boxylate oxygens is 5.8 Å and 6.6 Å from the O1 and C1 A glucose molecule is bound to the enzyme atoms of the glucose moiety. Another acidic residue, A glucose molecule is observed in the bottom of a deep Glu161, which has its carboxylate oxygens 8.0–8.5 Å from α β pocket at the interface between the ( / )8-barrel and the C1 carbon atom, is occluded by the guanidinium α β ( / )6-sheet domains (Figure 5a). The electron density of group of Arg158. The carboxylate group of Asp95 at the the MIR map was of sufficient quality to show that the lower left side of the pocket interacts with both the C4 4 glucopyranose molecule is in the C1 conformation; no dis- and the C6 hydroxyl groups of the glucose moiety. Adja- tortion of the glucopyranose ring is apparent. Furthermore, cent to this at the lower side of the pocket is Trp430, the glucopyranose is present in the β-anomeric configura- which lies perpendicular to the glucopyranose ring and tion, and it appears that physical constraints within the interacts with its lower, hydrophobic face. The hydroxyl pocket (Figure 5a) would prevent mutarotation to the group of Tyr253 on the right side of the slot points α-anomeric configuration. The refined structure revealed towards the C1 of the glucosyl moiety at a distance of that the glucose has an occupancy of approximately one in approximately 3.7 Å (Figure 5a). the crystal. The glucose moiety lies near the centre of mass of the enzyme and is affected less by libration motion The location of the acid groups Asp285 and Glu491 on of the molecule. This results in the glucose having a lower either side of the C1 carbon of the glucose moiety site is thermal motion than that of the enzyme molecule overall. consistent with these residues being the catalytic nucle- ophile and the catalytic proton donor of the enzyme, The entrance to the pocket that encloses the glucose mol- respectively. The Asp285 has been confirmed as the cat- ecule is shaped like a coin slot and is about 13 Å deep. alytic nucleophile by specific labelling with conduritol B The entrance is bound by Trp286 at the top and Trp434 epoxide (MH and GBF, unpublished results). On the basis at the bottom on the narrower side (approximately 8 Å) of of suggestions that catalytic-acid residues in retaining gly- the slot, by the triple glycine loop (Gly56–Gly57–Gly58) cosyl hydrolases are usually 3–4.5 Å from the anomeric C1 on the left side, and by Arg291 on the wider (11.5 Å), right carbon [19], we conclude at this stage that Glu491 is more side of the slot. The top left-hand corner of the pocket is likely to be the catalytic acid than Glu220 is. The aglycon lined with the hydrophobic residues Met250, Phe144, can be easily accommodated in the mouth of the active-site Met316 and Leu54. At the bottom of the pocket, two basic pocket, with the pyranose ring sandwiched between Research Article Barley β-glucan exohydrolase Varghese, Hrmova and Fincher 185

Figure 5

(a) (b)

Structure

Glucose trapped in the active site of β-glucan exohydrolase isoenzyme exohydrolase isoenzyme ExoI. The dotted lines indicate hydrogen ExoI. (a) A MOLSCRIPT [50] stereo representation of the active-site bonds and ionic interactions between the glucose and acidic and pocket of barley β-glucan exohydrolase isoenzyme ExoI. Nearest- basic amino acid residues of the enzyme, and a water molecule neighbour hydrogen-bond interactions between the bound glucose (distances are in Å). Represented on an arc at the top left-hand side of moiety (cyan) and the contact amino acid residues (standard colours) the diagram are a group of four hydrophobic residues (Leu54, Met316, in the active site pocket are shown as dashed lines. The molecular Phe144 and Met250), which interact with the C3, C5 and C6 atoms surfaces (GRASP) [51] of domains 1 and 2 are represented by on the hydrophobic face of the bound glucopyranose ring. transparent red and green surfaces, respectively. Carbon, oxygen and Represented on an arc on the other, lower, side of the ring are nitrogen atoms are represented by black, red and blue spheres hydrophobic interactions of Trp430 with carbon atoms C2 and C4. respectively, and water molecules are represented as larger red Note that the carboxylate oxygen of Asp285, which is not H-bonded to spheres. (b) A schematic drawing of the bonding interactions of the the C2 hydroxyl of the sugar, has a close contact (less than 3 Å) with 4 β bound glucose molecule, shown as C1 -glucopyranose with atomic the anomeric C1 carbon of the glucopyranose ring (not shown). numbering of C atoms, in the active site of the barley β-glucan

Trp286 and Trp434, and some of the hydroxyl groups of The presence of the glucose moiety in the active site of the sugar could be stabilized by Arg291. It might be argued the enzyme raises questions regarding its origin. One pos- that the glucose-binding pocket of the barley β-glucan exo- sibility is that the glucose was present in tissue extracts hydrolase is not associated with the active site, because and was loosely bound in the active-site region of the ‘non-catalytic’ glucose-binding sites are found on a number enzyme during its purification. The protein underwent of other proteins. However, the glucose-binding site on the several cycles of purification prior to crystallization, barley enzyme (Figure 5a) is quite different from non-cat- including fractional precipitation with ammonium sul- alytic glucose-binding sites on proteins such as glycogen phate, anion and cation ion-exchange chromatography, phosphorylase [20], glucose/galactose-binding protein [21] chromatofocusing, hydrophobic-interaction chromatogra- and xylose [22], where acidic amino acid phy and gel filtration [1]. Given the relatively low binding residues on either side of the C1 carbon are invariably constant of glucose to the enzyme [1], one would expect absent. In addition, the glucose bound to the barley that any glucose in the original tissue extracts that might enzyme interacts exclusively with charged and hydropho- bind weakly to the active site would have diffused away bic residues, whereas the glucose bound to the non-cat- by the time the enzyme was purified and crystallized. No alytic sites of the other glucose-binding proteins interacts glucose was used in buffers during the purification proce- with neutral residues such as asparagine. Finally, the sug- dure [1]. It would appear therefore that the glucose is gestion that the pocket represents the active site of the tightly bound to the barley β-glucan exohydrolase. Consis- enzyme has been confirmed by the observation that a com- tent with this conclusion is the observation from the petitive inhibitor binds to the enzyme in exactly the same refined structure that the glucose has thermal parameters region, and simultaneously displaces the bound glucose similar to those of the surrounding amino acid residues (unpublished results). Deeper, funnel-shaped pockets are (data not shown). This indicates a very high binding affin- found at the active sites of β- and 6-phospho- ity for the glucose molecule at this site. β- [23–25], but these other pockets are much longer, are narrower at the exit in comparison with the It might also be possible that loosely bound glucose origi- barley β-glucan exohydrolase and can accommodate sub- nally present in the active site has been chemically modi- strates with up to eight glucosyl residues [25,26]. fied to form a lactone or some other derivative that binds 186 Structure 1999, Vol 7 No 2

to the enzyme more strongly. In the case of viral neu- not appear to belong to the 4/7 superfamily [29] of β α raminidase, it has been shown that in the ( / )8-barrel glycosyl hydrolases, where the catalytic active site is slowly converted to deoxydehydroneuraminic acid/base and nucleophile active site residues are found at acid, which is a more potent inhibitor of the carboxy ends of the fourth and seventh β strands of the than is neuraminic acid [27]. Because the barley β-glucan barrel, respectively. exohydrolase was purified over a period of several weeks and crystals grew for 100 days or more [9], it is possible The putative active-site pocket is lined by 17 amino acid that any weakly bound glucose originally present could residues, three of which (Trp430, Trp434 and Glu491) are have been oxidised to gluconolactone during the part of the second domain. The site is characterised by the prolonged purification and crystallization procedures. presence of several acidic and basic amino acid residues, D-Glucono-1,5-lactone is a considerably stronger inhibitor and by two hydrophobic patches that interact with the of the enzyme than glucose [1]. If the bound moiety in the bound glucose moiety (Figure 5). Among these residues active site were indeed a lactone, enzyme activity would are the putative active-site residues Asp285 and Glu491; be measurably lowered. However, on redissolving the the latter arises from domain 2. Such an arrangement of crystals, the specific activity was the same as the native, amino acid residues in the active-site region is reminiscent freshly purified enzyme [9]. of the neuraminidase [27], which is a glycan exohydrolase from family 34 [7,30]. In the neuraminidase, On the basis of these considerations and data, we con- Glu276 forms short H bonds with hydroxyl groups of the clude that the glucose probably represents the product of triol group of neuraminic acid, Asp151 is likely to be the the enzyme-catalysed reaction and that the glucose catalytic acid, and the nucleophile is probably Tyr406 [27]. product has not been released from the surface of the binary enzyme–product complex after hydrolysis is com- In the active-site region of the barley β-glucan exohydro- pleted. This further implies that the glucose remains lase, the hydroxyl groups of the bound glucose are all bound to the enzyme until another, incoming substrate involved in multiple hydrogen-bonding and ionic interac- molecule displaces it. But what would be the structural tions with seven active-site amino acid residues basis for this release mechanism as the incoming substrate (Figure 5b). The C1 hydroxyl of the glucose forms two molecule binds to the binary enzyme–product complex? hydrogen bonds with Glu491 and one with Tyr253, the C2 The two-domain structure of the active site could allow a hydroxyl interacts with both Arg158 and Asp285, the C3 change in the spatial orientation between domains when hydroxyl forms three tetrahedrally directed interactions substrate binds to the enzyme or when the aglycone disso- with Arg158, Lys206 and His207, the C4 hydroxyl inter- ciates after hydrolysis. The long loop (residues 358–373) acts with both Asp95 and Lys206, and the C6 hydroxyl that connects the two domains (Figure 1) has a partially interacts with Asp95 and a water molecule (Figure 5b). formed helical structure and could, when an incoming Hydrophobic interactions of Leu54, Met316, Phe144 and substrate molecule is bound, become fully helical. Thus, Met250 with the C3H, C5H and C6H regions on the changes in the conformation of the loop could allow it to hydrophobic face of the glucopyranose ring, together with act as a hinge that would, in turn, allow the two domains to interactions between Trp430 and the C2H and C4H on separate in such a way that conformational changes in the the other face, further stabilize the glucose molecule in region of the pocket would release the bound glucose. the active-site pocket of the enzyme.

Substrate binding The likely locations of the catalytic and binding sites on The barley β-glucan exohydrolase is inhibited by glucose, the barley enzyme can also be gleaned from analysis of the product of hydrolysis that is released from nonreducing residues that are conserved among closely related family 3 termini of a range of substrates, and by glucose analogues glycosyl hydrolases [1, 10, 11; AJ Harvey and GBF, unpub- such as conduritol B epoxide, D-glucono-1,5-lactone and lished results, GenBank accession number AF102868; 2,4-dinitrophenyl 2-deoxy-2-fluoro-glucopyranoside [1] N Koizumi, unpublished results, GenBank accession (MH and GBF, unpublished results). On this basis, the number G3582436]. Apart from the substitution of Tyr253 pocket containing the noncovalently bound glucose moiety in the barley enzyme with a Phe residue in the corre- in the crystal structure is likely to represent the active site sponding position of the bacterial enzyme [11], the α β of the enzyme. The site lies at the junction of the ( / )8 arrangement of amino acid residues in the active-site α β domain and the ( / )6 domain, above the pseudo-eightfold pockets (Figure 5) is totally conserved. α β axis of the ( / )8 barrel. According to the structural classifi- cation of proteins (SCOP), a location for the active site The active site and the hydrolytic mechanism above the pseudo-eightfold axis is typical of enzymes with On the basis of the bell-shaped pH dependence of barley β α β β ( / )8 folds and, in particular, of members of the -gly- -glucan exohydrolase activity, Hrmova and Fincher [6] canase family of the glycosyltransferase superfamily [28]. concluded that at least two ionizable amino acid residues On the other hand, the barley β-glucan exohydrolase does participate in catalysis, and ionization constants for the Research Article Barley β-glucan exohydrolase Varghese, Hrmova and Fincher 187

two catalytic amino acid residues were estimated at glucopyranose ring is consistent with the enzyme having a approximately 4.7 and 5.9. Furthermore, proton NMR retaining mechanism. Proton NMR has been used previ- showed that the anomeric configuration of released ously to demonstrate that β-glucose is indeed released glucose is retained during hydrolysis [1]. Enzymic hydrol- during hydrolysis of a variety of substrates by the barley ysis of glycosidic linkages in oligosaccharides and polysac- enzyme [1]. Further support for a role for Asp285 and charides is widely believed to be initiated by protonation Glu491 in catalysis is provided by the observation that of the glycosidic O atom, via proton transfer from an these residues are amongst the most highly conserved active-site-located catalytic acid [19,31]. Following proto- residues in family 3 glycosyl hydrolases, although the nation, the aglycone is released from the enzyme surface neighbouring Glu220 is also highly conserved among the and the reducing terminal residue of the bound substrate plant exohydrolases (data not shown). It should now be is converted to a positively charged oxycarbonium ion. possible to use active-site-directed inhibitors [25] and site- The oxycarbonium ion is stabilized, either electrostatically directed mutagenesis [34] to confirm the role of these or by a covalent enzyme–substrate intermediate, by a residues in catalysis. nucleophilic amino acid residue, which usually carries a carboxylate group. Hydration of the oxycarbonium or As mentioned earlier, the 16-residue helix-like loop that hydrolysis of the covalent intermediate results in the connects the two domains (Figure 1) could act as a hinge regeneration of the glycosyl residue and reprotonation of that would allow the two domains to move relative to each the catalytic acid. other. Because three residues of domain 2, including the putative catalytic acid Glu491, are located close to the active In the barley β-glucan exohydrolase structure, the two site and are likely to contribute to catalysis, movement of acidic amino acid residues Asp285 and Glu491, which arise the second domain away from, or closer to, the active site on from domains 1 and 2 of the enzyme, respectively, are domain 1 would almost certainly affect catalytic activity and appropriately positioned in relation to the C1 hydroxyl could be used to regulate enzyme activity. group of the bound glucose molecule and have been iden- tified as likely catalytic residues (Figure 5). One of the car- A potential role for domain 2 in binding (1→3,1→4)-β-glucan boxylate oxygens of Asp285 has a close contact (less than Kinetic data previously obtained for the barley β-glucan 3 Å) with the C1 carbon of the glucose moiety. This car- exohydrolase isoenzyme ExoI suggested that there is boxylate oxygen could stabilize the positive charge on the more than one binding site on the enzyme for the C1 carbon of the oxycarbonium ion during catalysis, iden- (1→3,1→4)-β-glucan substrate [6]. Given that domain 1 is tifying it as the most likely candidate for the nucleophile likely to contain the catalytic site, the additional in the catalytic mechanism. This situation is similar to the (1→3,1→4)-β-glucan-binding site could be located on the short O…C=O contacts found in small organic compounds, second domain. Consistent with this possibility is the in which the oxygen atom acts as the nucleophile in an identification of a potential binding site, based on the incipient nucleophilic addition reaction with the carbon topology of the open β-sheet domain (Figure 3). Brändén atom [32]. Similar short contacts have been observed in [35] has described carbohydrate-binding sites in a number the structures of neuraminidase/ complexes [27]. of proteins where two adjacent loops that emerge on oppo- In addition, in bovine trypsin complexed with the bovine- site sides of a β sheet form a binding site between them trypsin inhibitor, the oxygen atom of the putative serine [36]. The most likely region for (1→3,1→4)-β-glucan nucleophile (Ser195) has a short contact (2.7 Å) with a binding would be on the loops that emerge from the peptide carbonyl (Lys15) of the inhibitor [33]. C-terminal ends of β strands j and l (Figure 1). This would involve residues 417–440 (loop A) emerging from β strand In the barley β-glucan exohydrolase, Glu491 is positioned j and residues 484–504 (loop B) emerging from β strand l. on the other side of the glucopyranose ring (in comparison In addition, the 17-residue loop (residues 429–445) con- with the catalytic nucleophile Asp285), approximately 3.5 Å necting helices J1 and J2 has characteristics similar to cellu- from the anomeric C1 carbon (Figure 5), and is a strong lose-binding domains from bacterial [37], where candidate for the catalytic acid. Although Glu491 appears relatively large numbers of hydroxy amino acids, a low solvated in the structure, it could facilitate the activation of abundance of charged amino acid residues and conserved a water molecule that would subsequently act as the proton Trp and Gly residues are observed (Figures 1 and 2). It donor in the catalytic mechanism. The presence of the should now be possible to experimentally confirm the hydrophobic residues Trp430 and Trp434 in close proxim- location of the (1→3,1→4)-β-glucan-binding site by first ity to Glu491 (Figure 5a) might be expected to increase the blocking enzyme activity by soaking crystals with pKa of Glu491, although interactions with Arg158 would inhibitors such as conduritol B epoxide [25], and subse- → → β tend to deprotonate the acid or decrease its pKa. quently adding (1 3,1 4)- -glucooligosaccharides to the crystals. The (1→3,1→4)-β-glucooligosaccharides would The location of Asp285 in the barley β-glucan exohydro- be expected to bind to the second binding site identified lase (Figure 5) above the hydrophobic face of the bound by Hrmova and Fincher [6], and enzyme inactivation 188 Structure 1999, Vol 7 No 2

would ensure that the (1→3,1→4)-β-glucooligosaccharides Of particular importance in describing the enzyme’s would not be hydrolyzed. mechanism of action will be the reconciliation of its three-dimensional structure with previous indications Comparison with other family 3 glycosyl hydrolases from kinetic analyses that a specific, non-catalytic Currently, there are over 60 members of the family 3 (1→3,1→4)-β-glucan-binding site is present. The second group of glycosyl hydrolases [7,30]. Multiple amino acid binding site for (1→3,1→4)-β-glucans could serve to sequence alignments and hydrophobic-cluster analyses attach the enzyme to cell walls. It could simply anchor show that these members fall into approximately six sub- the enzyme to (1→3,1→4)-β-glucans during hydrolysis groups that appear as clearly distinguishable branches on or it could form a more integral component of the an unrooted phylogenetic radial tree (data not shown). hydrolytic mechanism, for example, by feeding the The four plant members, two from barley [1] (AJ Harvey nonreducing terminus of the polysaccharide into the cat- and GBF, unpublished results; GenBank accession alytic site. If the specific (1→3,1→4)-β-glucan-binding number AF102868), and one each from nasturtium [10] site enables the enzyme to attach to fibrillar or more and tobacco (N Koizumi, unpublished results; GenBank refractory regions of the wall, then the enzyme might accession number G3582436), are closely clustered. The disrupt these regions physically to facilitate overall related (1,4)-β-glucan glucohydrolase from Pseudomonas enzyme penetration and hydrolysis of the wall. This [11] belongs to a different cluster. Our preliminary analy- could also provide opportunities for engineering the ses indicate that each of the plant enzymes has a domain enzyme to allow it to bind to other insoluble substrates, organization that is similar to the barley β-glucan exo- or to express the (1→3,1→4)-β-glucan binding domain hydrolase isoenzyme ExoI described here, and that amino for use as an ‘antibody’ to locate the polysaccharide in acid residues implicated in catalysis are highly conserved. histochemical studies. The availability of the 3D structure for the barley enzyme (Figure 1) will now allow a detailed comparison of the Materials and methods structural features and evolutionary relationships of other Enzyme purification and characterization β members of family 3 glycosyl hydrolases. The barley -glucan exohydrolase isoenzyme ExoI was purified from young seedlings as previously described [1,6]. The purity of the preparation was confirmed by the presence of a single band during Biological implications sodium dodecyl sulphate (SDS) polyacrylamide gel electrophoresis The presence of the β-glucan exohydrolases in young and by N-terminal amino acid sequence analysis of the first 100 seedlings, as well as their broad substrate specificities, residues, which demonstrated that the protein corresponded to β-glucan exohydrolase isoenzyme ExoI. In addition, recoveries of raise important questions regarding the biological func- amino acid residues during sequencing were close to theoretical tions of these enzymes. One possible role for the values and no secondary sequences were observed. The molecular β-glucan exohydrolases could be in the auxin-induced mass of the purified enzyme was estimated on a Finnigan Lasermat elongation of cells in growing coleoptiles. In this process, 2000 MALDI-TOF mass spectrometer (Finnigan MAT, Hemel Hemp- stead, UK), using sinapinic acid as matrix. it has been suggested that matrix-phase polysaccharides of cell walls, including (1→3,1→4)-β-glucans and het- Crystallization and data collection eroxylans, could be partially hydrolysed, and that the Crystals were obtained as described by Hrmova et al. [9]. Data were × × resultant ‘loosening’ of noncovalent cross-links between collected from a single large crystal (0.8 mm 0.6 mm 0.4 mm) in a glass capillary at 18°C on a Rigaku R-axis II Image Plate X-ray detector, cellulose microfibrils could allow osmotically induced mounted on a MAC Science SRA M18XH1 rotating anode X-ray gen- turgor pressure to stretch the cell wall and cause the cell erator, which was operated at 47 kV and 60 mA with focussing mirrors to expand. The β-glucan exohydrolases are certainly for Cu Kα. Data-collection and heavy-atom refinement statistics are associated with cell walls in elongating coleoptiles of given in Table 1. A native data set of 148,526 observations (> 1σ) were collected and reduced to 40,310 unique reflections. The crystal both barley and maize, and decreases in cell wall had a tetragonal space group P41212 or P43212, with cell edges (1→3,1→4)-β-glucan levels are observed in both coleop- 102.09 Å × 102.09 Å × 184.50 Å. Data from a PIP [di-µ-iodo bis(ethyl- tiles and young leaves as development proceeds. enediamine) di-platinum (II) nitrate] derivative were collected from the same crystal by removing the seal at one end of the capillary and dis- However, it is not clear how exohydrolases would solving a grain of PIP into the mother liquor. The capillary was resealed for 18 h, which allowed diffusion of PIP into the crystal. The crystal was mediate the ‘loosening’ process via partial exohydrolysis subsequently pushed out of the mother liquor and data were collected. of cell wall (1→3,1→4)-β-glucans, unless an endohydro- A total of 106,488 intensities (> 1σ) were collected, reducing to lase first released a new nonreducing end. It should be 42,372 unique reflections. A third data set was collected on a second emphasised that cell elongation in higher plants is not smaller crystal, obtained by micro-seeding techniques [38,39], soaked in 5 mM MHG (methyl mercury chloride) for 18 h. A total of 83,154 well understood and a more thorough knowledge of the observations, reducing to 23,992 unique observations, were collected action mechanisms of the enzymes involved in wall loos- to 2.5 Å resolution. All data were processed using the HKL software ening could provide clues as to the mechanism of cell [40] (see Table 1 for data collection statistics). elongation and, hence, to the crucial and as yet unan- Heavy-atom location and refinement swered questions regarding the more general mecha- For the PIP derivative, four platinum-binding sites were located by differ- nisms of plant growth. ence Patterson methods and vector-verification methods using the Research Article Barley β-glucan exohydrolase Varghese, Hrmova and Fincher 189

Table 2 have been deposited with the Brookhaven [49] under codes 1EX1 and R1EX1SF, respectively. Final model statistics*. Acknowledgements Resolution (Å) 20–2.2 We thank Peter M Colman for his continuing and high level support, Brian Number of reflections 39,456 (1633) Smith for useful discussions and ideas relating to catalytic mechanisms, Completeness (%) 83.3 (37.8) Andrew J Harvey for providing amino acid sequence data, Ross Degori for help with refinement of the structure and Bert van Donkelaar for skilled tech- Rwork 17.0 (20.2) nical assistance. We are grateful to Bruce Stone for his advice and ongoing Rfree 21.1 (28.3) Rms bonds (Å) 0.013 interest. The work was supported by grants (to GBF) from the Australian Research Council and the Grains Research and Development Corporation Rms angles (°) 2.1 of Australia. *The numbers in parentheses are the values in the high resolution shell. References 1. Hrmova, M., et al., & Fincher, G.B. (1996). Barley β-D-glucan PROTEIN software [41]. These sites were also located in the anomalous exohydrolases with ß-D-glucosidase activity. J. Biol. Chem. difference Patterson and were indicative of a strong anomalous signal. 271, 5277-5286. Two mercury-binding sites were located in the MHG derivative by similar 2. Kotake, T., Nagawa, N., Takeda, K. & Sakurai, N. (1997). Purification methods and, using difference Fourier methods, each derivative gave a and characterization of wall bound exo-1,3-β-glucanase from barley consistent set of heavy-atom sites for the other derivative in the space (Hordeum vulgare L.) seedlings. Plant Cell Physiol. 38, 194-200. group P4 2 2. Using anisotropic temperature factors, the platinum sites 3. Sakurai, N. & Masuda, Y. (1978). Auxin-induced changes in barley 3 1 Plant Cell Physiol. 19 were split into two sites along the direction of maximum anisotropy. coleoptile cell wall composition. , 1217-1223. 4. Inouhe, M., McClellan, M. & Nevins, D.J. (1997). Developmental regulation of polysaccharide and growth in the primary cell Heavy-atom refinement and phasing (see Table 1 for statistics) were walls of maize. Int. J. Biol. Macromol. 21, 21-28. carried out by the maximum likelihood Bayesian methodology, using the 5. Cline, K. & Albersheim, P. (1981). Host-pathogen interactions. XVI. program SHARP [42] with data from the PIP derivative to 2.3 Å and Purification and characterization of a glucosyl hydrolase/ from the MHG derivative to 2.5 Å. A 2.3 Å map using the SHARP MIR present in the walls of soybean cells. Plant Physiol. 68, 207-220. phases was solvent-flattened by the Solomon method [43] using 55% 6. Hrmova, M. & Fincher, G.B. (1998). Barley β-D-glucan exohydrolases. solvent (calculated solvent content of 63%) and the phases extended Substrate specificity and kinetic properties. Carbohydr. Res. to 2.2 Å. A reliability index based on calculated structure factors from 305, 209-221. 7. Henrissat, B. & Bairoch, A. (1993). New families in the classification of the back-transformed solvent-flattened map was 0.167. glycosyl hydrolases based on amino acid sequence similarities. Biochem. J. 293, 781-788. Structure determination and refinement 8. Varghese, J.N., Garrett, T.P.J., Colman P.M., Chen, L., Høj, P.B. & The protein was clearly interpretable in the solvent-flattened map. The Fincher, G.B. (1994). Three-dimensional structures of two plant high quality of the map was probably due to the lack of systematic errors β-glucan endohydrolases with distinct substrate specificities. Proc. in the differences between the native and the PIP data (collected on the Natl Acad. Sci. USA 91, 2785-2789. same crystal), the low mosaic spread (< 0.1°) of the large crystal, and 9. Hrmova, M., Varghese, J.N., Høj, P.B. & Fincher, G.B. (1998). β improvements in heavy-atom refinement [42]. The map was skeletonised, Crystallization and preliminary X-ray analysis of -D-glucan exohydrolase isoenzyme ExoI from barley (Hordeum vulgare). Acta and this was used as a guide for tracing the Cα chain of the protein, Cryst. D 54, 687-689. using the graphics program O version 6.1 [44]. All residues in the chain 10. Crombie, H.J., Chengappa, S., Hellyer, A. & Reid, J.S.G. (1998). A were identified unambiguously, apart from the last three C-terminal amino xyloglucan oligosaccharide-active, transglycosylating β-D-glucosidase acids (residues 603–605). An atomic model was built using Lego frag- from the cotyledons of nasturtium (Tropaeolum majus L.) seedlings — ments in the program O, including N-linked carbohydrates at Asn221 purification, properties and characterization of a cDNA clone. Plant J. and Asn498, and a glucose molecule bound in the putative active site of 15, 27-38. the molecule. Water molecules could be seen in this map, but were not 11. Rixon, J.E., Ferreira, L.M.A., Durrant, A.J., Laurie, J.I., Hazlewood, G.P. built. The atomic model was refined with the maximum-likelihood refine- & Gilbert, H.J. (1992). Characterization of the gene celD and its β Pseudomonas ment program REFMAC [45], and SIGMAA-weighted maps [46] were encoded product 1,4- -glucan glucohydrolase D from fluorescens subsp. cellulosa. Biochem. J. 285, 947-955. used for model building. One hexosyl residue attached to Asn221 and 12. Cutfield, S., Brooke, G., Sullivan, P. & Cutfield, J. (1992). seven hexosyl residues attached to Asn498 were refined. A putative Crystallization of the exo-(1,3)-β-glucanase from Candida albicans. xylosyl residue attached to the C2 hydroxyl of the branching mannosyl J. Mol. Biol. 225, 217-218. residue in the carbohydrate attached to Asn498 was observed, but not 13. MacKenzie, L.F., Brooke, G.S., Cutfield, J.F., Sullivan, P.A. & Withers, modelled. Two hundred and twenty water molecules were added to the S.G. (1997). Identification of Glu330 as the catalytic nucleophile of structure, using the program ARPP [47], by adding 20 water molecules Candida albicans exo-β-(1,3)-glucanase. J. Biol. Chem. at a time to peaks (> 5σ) in the difference Fourier. Refinement cross vali- 272, 3161-3167. dation was carried out by selecting, at random, 5% of the data (2128 14. Mikola, J. (1983). Proteases, peptidases, and inhibitors of endogenous proteinases in germinating seeds. In Seed Proteins. reflections), which were used throughout the refinement to calculate free (Daussant, J., Mosse, J. & Vaughan, J., eds), pp. 35-52, Academic R factors. Final refinement statistics are presented in Table 2. The esti- Press, London. mated overall coordinate error based on R value, Free R value and 15. Takahashi, N., et al., & Arata, Y. (1986). Xylose-containing common maximum likelihood is 0.195, 0.176 and 0.106 Å, respectively. structural unit in N-linked oligosaccharides of laccase from sycamore cells. Biochemistry 25, 388-395. The stereochemical quality of the model was assessed with the 16. Sturm, A., Johnson, K.D., Szumilo, T., Albein, A.D. & Chrispeels, M.J. program PROCHECK [48]. A Ramachandran plot indicated that (1987). Subcellular localization of glycosidases and 90.3% of the amino acid residues are in the ‘most-favoured’ regions, glycosyltransferases involved in the processing of N-linked 9.1% are in ‘additionally allowed’ regions, 0.4% are in ‘generously oligosaccharides. Plant Physiol. 85, 741-745. allowed’ regions and 0.2% in ‘disallowed’ regions. The ‘disallowed’ 17. Imberty, A., Delage, M-M., Bourne, Y., Cambillau, C. & Perez, S. (1991). Data bank of three-dimensional structures of disaccharides: Part II, residue is Ile432, but this residue is well-defined by electron density. N-acetyllactosaminic type N-glycans. Comparison with the crystal structure of a biantennary octasaccharide. Glycoconjugate J. 8, 456-483. Accession numbers 18. Harthill, J.E. & Thomsen, K.K. (1995). Analysis of glycan structures of The atomic coordinates and structure factors of the refined protein barley (1,3;1,4)-β-D-glucan 4-glucanohydrolase isoenzyme EII. Plant model of barley β-glucan exohydrolase isoenzyme ExoI determined here Physiol. Biochem. 33, 9-18. 190 Structure 1999, Vol 7 No 2

19. White, A. & Rose, D.R. (1997). Mechanism of catalysis by retaining β- 45. Murshudov, G.N., Vagin, A.A. & Dodson, E.J. (1997). Refinement of glycosyl hydrolases. Curr. Opin. Struct. Biol. 7, 645-651. macromolecular structures by maximum-likelihood method. Acta Cryst. 20. Martin, J.L., Johnson, L.N., & Withers, S.G. (1990) Comparison of the D 53, 240-255. binding of glucose and glucose-1-phosphate derivatives to T-state 46. Read, R.J. (1986). Improved Fourier coefficients for maps using glycogen phosphorylase. Biochemistry 29, 10745-10757. phases from partial structures with errors. Acta Cryst. A 42, 140-149. 21. Mowbray, S.L. (1992). Ribose and glucose-galactose receptors. 47. Lamzin, V.S. & Wilson, K.S. (1993). Automated refinement of protein Competitors in bacterial chemotaxis. J. Mol. Biol. 227, 418-440. models. Acta Cryst. D 49, 129-147. 22. Carrell, H.L., Glusker, J.P., Burger, V., Manfre, F., Tritsch, D. & 48. Laskowski, R.A., MacArthur, M.W., Moss, D.S. & Thorton, J.M. (1993). Biellmann, J.-F. (1989) X-ray analysis of D-xylose isomerase at 1.9 Å: PROCHECK: a program to check the stereochemical quality of native enzyme complexed with substrate and a mechanism designed protein structures. J. Appl. Cryst. 26, 283-291. inhibitor. Proc. Natl Acad. Sci. USA 86, 4440-4444. 49. Bernstein, F.C., et al., & Tasumi, M. (1977). The protein data bank: a 23. Barrett, T., Suresh, S.G., Tolley, S.P., Dodson, E.J. & Hughes, M.A. computer-based archival file for macromolecular structures. J. Mol. (1995). The crystal structure of cyanogenic, β-glucosidase from white Biol. 112, 535-542. clover, a family 1 glycosyl hydrolase. Structure 3, 951-960. 50. Kraulis, P.J. (1991). MOLSCRIPT: a program to produce both detailed 24. Weismann, C., Beste, G., Hengstenberg, W. & Schultz, G.E. (1995). and schematic plots of protein structures. J. Appl. Cryst. 24, 946-950. The three-dimensional structure of 6-phospho-β-galactosidase from 51. Nicholls, A., Sharp, K. & Honig B. (1991). Protein folding and Lactococcus lactis. Structure 3, 961-968. association: insights from the interfacial and thermodynamic 25. Hrmova, M., MacGregor, E.A., Biely, P., Stewart, R.J. & Fincher, G.B. properties of hydrocarbons. Proteins 4, 281-296. (1998). Substrate binding and catalytic mechanism of a barley 52. Barton, G.J. (1993). ALSCRIPT: a tool to format multiple sequence β-D-glucosidase/(1,4)-β-D-glucan exohydrolase. J. Biol. Chem. alignments, Protein Eng. 6, 37-40. 273, 11134-11143. 26. Burmeister, W.P., Henrissat B., Bosso, C., Cusack, S. & Ruigrok, R.W.H. (1993) neuraminidase can synthesize its own inhibitor. Structure 1, 19-26. 27. Varghese, J.N., McKimm-Breschkin, J., Caldwell, J.B., Kortt, A.A. & Colman, P.M. (1992). The structure of the complex between influenza virus neuraminidase and sialic acid, the viral receptor. Proteins 14, 327-332. 28. Murzin, A.G., Brenner, S.E., Hubbard, T. & Chothia, C. (1995). SCOP: a structural classification of proteins database for the investigation of sequences and structures. J. Mol. Biol. 247, 536-540. 29. Jenkins, J., Leggio, L.L., Harris, G. & Pickergill, R. (1995). β-glucosidase, β-galactosidase, family A cellulases, family F and two barley glucanases form a superfamily of enzymes with 8-fold β/α architecture and with two conserved glutamates near the carboxy- terminal ends of β-strands four and seven. FEBS Lett. 362, 281-285. 30. Henrissat, B. (1998). Glycosidase families. Biochem. Soc. Trans. 26, 153-156. 31. McCarter, J.D. & Withers, S.G. (1994). Mechanism of enzymatic glycoside hydrolysis. Curr. Opin. Struct. Biol. 4, 885-892. 32. Burgi, H.B., Dunitz, J.D. & Shefter, E. (1974). Chemical reaction paths. IV. Aspects of O…C=O interactions in crystals. Acta Cryst. B 30, 1517-1527. 33. Marquart, M., Walter, J., Deisenhofer, J., Bode, W. & Huber, R. (1983). The geometry of the reactive site and the peptide groups in trypsin, trypsinogen and its complexes with inhibitors. Acta Cryst. B 39, 480-490. 34. Chen, L., Garrett, T.P.J., Fincher, G.B. & Høj, P.B. (1995). A tetrad of ionizable amino acids is important for catalysis in barley β-glucanases. J. Biol. Chem. 270, 8093-8101. 35. Brändén, C.I. (1980). Relation between structure and function of α/β proteins. Q. Rev. Biophys. 13, 317-338. 36. Gilliland, G.L. & Quicho, F.A. (1981). Structure of the L-arabinose- binding protein from Escherichia coli at 2.4 Å. J. Mol. Biol. 146, 341-362. 37. Gilkes, N.R., Henrissat, B., Kilburn, D.G., Miller, J.C. Jr. & Warren, R.A.J. (1991) Domains in microbial β-1,4-glycanases: sequence conservation, function, and enzyme families. Microbiol. Rev. 55, 303-315. 38. McPherson, A. (1990). Current approaches to macromolecule crystallization. Eur. J. Biochem. 189, 1-23. 39. Chayen, N.E. (1998). Comparative studies of protein crystallization by vapour diffusion and microbatch techniques. Acta Cryst. D 54, 8-15. 40. Otwinowski, Z. & Minor, W. (1996). A processing of X-ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307-326. 41. Stiegeman, W. (1974). Protein: Program System for the Crystal Structure Analysis of Proteins. Ph.D. Dissertation, Technical University Munich, Germany. 42. De la Fortelle, E. & Bricogne, G. (1997). A maximum-likelihood heavy- atom parameter refinement for multiple isomorphous replacement and multiwavelength anomalous diffraction methods. Methods Enzymol. 276, 472-494. Because Structure with Folding & Design operates a 43. Abrahams, J.P. & Leslie, A.G.W. (1996). Methods used in the structure determination of bovine mitochondrial F1 ATPase. Acta ‘Continuous Publication System’ for Research Papers, this Cryst. D 52, 30-42. paper has been published on the internet before being printed 44. Jones, T.A., Zou, J.Y., Cowan, S.W. & Kjeldgaard, M. (1991). Improved methods for building protein models in electron density maps and the (accessed from http://biomednet.com/cbiology/str). For location of errors in these models. Acta Cryst. A 47,110-119. further information, see the explanation on the contents page.