INTRODUCTION

Top Abstract Introduction Procedures Results & Discussion Conclusions References

Cyclodextrins (CDs)¹ are cyclic molecules composed of 6, 7, or 8 glucose units linked via (1,4)-glycosidic bonds (-, -, and -CD, respectively). The ability of cyclodextrins to form inclusion complexes with small hydrophobic molecules has provided a number of practical uses in the food, pharmaceutical, and agrochemical industries (1-3). Cyclodextrins are produced from starch by the action of the enzyme cyclodextrin glycosyltransferase (CGTase; EC 2.4.1.19). However, apart from the cyclization reaction, the enzyme can also catalyze disproportionation, coupling, and hydrolysis reactions. All known CGTases produce a mixture of -, -, and -cyclodextrins. For the industrial production of pure cyclodextrins, -CD is selectively crystallized and - and -CD are complexed with organic solvents. The industrial production of cyclodextrins might be improved by the construction of mutant CGTases with improved product specificity (2, 3).

In the conventional commercial production of cyclodextrins, starch is first liquefied by the action of a thermostable -amylase, whereafter cyclodextrins are produced using the mesophilic CGTase from Bacillus macerans. Two highly thermostable CGTases have been characterized that can directly be used for starch liquefaction, eliminating the need for -amylase pretreatment (2, 3). These enzymes are produced by thermophilic anaerobic bacteria belonging to the genus Thermoanaerobacter (9, 10) and Thermoanaerobacterium thermosulfurigenes EM1 (11). The overall amino acid compositions of both enzymes show relatively minor differences, and also the biochemical characteristics of both enzymes are very similar (11). The T. thermosulfurigenes EM1 CGTase displays maximum cyclization activity at 80-85 °C and maximum starch hydrolyzing activity at 90-95 °C. The pH optimum for cyclization is broad, in the range of pH 4.5-7.0.

The three-dimensional structures of several CGTases have been solved (4, 5, 7). CGTase belongs to family 13 of the glycosyl hydrolases, a group of homologous (/)₈-barrel proteins, to which also the -amylases belong (8). Recently, the three-dimensional structure of the CGTase from T. thermosulfurigenes EM1 (Tabium CGTase) was solved at 2.3-Å resolution (6). In the present study we describe the three-dimensional structure of an enzyme-substrate complex of Tabium CGTase. The x-ray structure of a maltohexaose inhibitor complexed with Tabium CGTase was solved at 2.6-Å resolution. The detailed information thus obtained allowed rational engineering of the cyclodextrin product specificity of Tabium CGTase.

Residues Glu²⁵⁸, Asp³²⁹, and Asp²³⁰ are directly involved in catalysis in CGTase (20). Glu²⁵⁸, together with Asp²³⁰, is believed to cleave the substrate's (1,4)-glycosidic bonds and to form the product's (1,4)-glycosidic bonds by a double displacement mechanism (15). In the first step of the reaction the general acid Glu²⁵⁸ protonates the oxygen of the glycosidic bond to be cleaved. After cleavage of this scissile bond, an oxocarbonium transition state is formed, which is believed to collapse into a covalently linked intermediate by nucleophilic attack of Asp²³⁰ on the anomeric C1. Subsequently, the reducing end diffuses out of the active site and an acceptor comes in, which can be a water molecule (in case of hydrolysis) or a carbohydrate C4-hydroxyl group (in case of transglycosylation). In the second step of the reaction, the acceptor hydroxyl group is activated through deprotonation by Glu²⁵⁸, after which the acceptor performs a nucleophilic attack on the covalent intermediate. Through another oxocarbonium transition state, the product ((1,4)-linked) is then formed (16, 17).

It appears from the double displacement mechanism that for optimal catalysis the nucleophile Asp²³⁰ must be deprotonated in the first step of the reaction, whereas the general acid Glu²⁵⁸ must be protonated in the first step, but deprotonated in the second step of the reaction. A third catalytic residue, Asp³²⁹, has been found to be hydrogen bonded to Glu²⁵⁸ in the unliganded CGTase, thereby elevating the pK_a of Glu²⁵⁸ and assuring its protonation. After substrate binding, this hydrogen bond is lost, making deprotonation of Glu²⁵⁸ possible. It was suggested that this Glu²⁵⁸-Asp³²⁹ interaction is responsible for the broad pH optimum exhibited by CGTases (16). The importance of the pK_a of Glu²⁵⁸ in the different reactions catalyzed by Tabium CGTase was further studied. By changing the electrostatic environment of Glu²⁵⁸ by site-directed mutagenesis, we could drastically shift the Tabium CGTase pH optimum.

The biochemical characteristics of the various mutant CGTases are presented with emphasis on the effects of these mutations on the CGTase cyclization and hydrolytic activities, pH optima, and product formation from starch.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results & Discussion Conclusions References

Structure Determination-- Crystals of the Tabium CGTase were grown from 21% saturated ammonium sulfate in 100 mM Tris buffer, pH 7.6. The crystals were stable in the presence of carbohydrates.

Bacterial Strains, Plasmids, and Growth Conditions-- Escherichia coli JM109 (35) was used for recombinant DNA manipulations. E. coli PC1990 (36), known to leak periplasmic proteins into the supernatant because of a mutation in its tolB locus, was used for (extracellular) production of CGTase (mutant) proteins. Plasmid pCT2, a derivative of pUC18 containing the amyA (cgt) gene of T. thermosulfurigenes EM1 (37), was used for site-directed mutagenesis, sequencing, and expression of the CGTase (mutant) proteins (Fig. 1). Plasmid-carrying bacterial strains were grown on LB medium with 100 µg/ml ampicillin. When appropriate, isopropyl--D-thiogalactopyranoside was added at a concentration of 0.1 mM for induction of protein expression.

View larger version (25K): [in this window] [in a new window]   Fig. 1.   Restriction map of the plasmid pCT2.

DNA Manipulations-- DNA manipulations and transformation of E. coli were essentially as described by Sambrook et al. (38). Electrotransformation of E. coli was performed using the Bio-Rad gene pulser apparatus (Bio-Rad, Veenendaal, The Netherlands). The selected conditions were 2.5 kV, 25 µF, and 200 .

Site-directed Mutagenesis-- Mutant CGTase genes were constructed via a double PCR method using Pfu DNA polymerase (Stratagene, Westburg, Leusden, The Netherlands). A first PCR reaction was carried out with the mutagenesis primer for the coding strand plus a primer 195-715 base pairs downstream on the template strand. The reaction product was subsequently used as primer in a second PCR reaction together with a primer 295-815 base pairs upstream on the coding strand. The product of the last reaction was cut with NcoI and MunI and exchanged with the corresponding fragment (900 base pairs) from the vector pCT2 (Fig. 1). The resulting (mutant) plasmid was transformed to E. coli JM109 for sequencing and to E. coli PC1990 for production of the (mutant) proteins. The following oligonucleotides were used to produce the mutations: D197H, 5'-CGTAACTTATTTCATTTAGCAGATCTAAATCAACAG-3'; F284K, 5'-GTCTTTTGGACAAGAGGTTTTCTC-3'; N327D, 5'-GGTTACTTTTATTGATGATCATGATATGG-3'; D371R, 5'-GACAGGCAATGGACGTCCTTATAATAGAGC-3'. The bold codons indicate the changed amino acids. Successful mutagenesis resulted in appearance of the underlined restriction sites (BglII for D197H, BclI for N327D and AatII for D371R), which allowed rapid screening of potential mutants. It was not possible to find a convenient restriction site for mutant F284K. Mutations were verified by DNA sequencing (39). All 900 base pairs on the MunI-NcoI fragment obtained by PCR were checked by DNA sequencing.

Production and Purification of CGTase Proteins-- For production of CGTase proteins, E. coli PC1990 (pCT2) was grown in a 2-liter fermentor at pH 7.0 and 30 °C. The medium contained 2% (w/w) trypton (Oxoid, Boom BV, Meppel, The Netherlands), 1% (w/w) yeast extract (Oxoid), 1% (w/w) sodium chloride, 1% (w/w) casein hydrolysate (Merck, Darmstadt, Germany), 100 µg/liter ampicillin, and 0.1 mM isopropyl--D-thiogalactopyranoside. Growth was monitored by measuring the absorbance (A) at 450 nm. At an A_450 nm of 2-3, an extra amount of 50 g of trypton was added to the fermentor. Cells were harvested after 20-24 h of growth (8000 × g, 30 min, 4 °C), at absorbance values of 8-12. The supernatant was directly applied to an -CD-Sepharose-6FF affinity column (40) for further purification of the CGTase proteins. After washing the column with 10 mM sodium acetate, pH 5.5, the CGTase was eluted with the same buffer supplemented with 1% (w/w) -CD. Purity and molecular weight of the CGTase (mutant) proteins were checked on SDS-polyacrylamide gel electrophoresis (11). Protein concentrations were determined by the method of Bradford (42), using the Coomassie protein assay reagent of Pierce (Pierce Europe bv, Oud-Beijerland, The Netherlands).

Enzyme Assays-- Specific assays were used to determine the activities (initial rates) of the four different reactions catalyzed by CGTases (14). In the cyclization reaction the reducing end of a sugar is transferred to another sugar residue in the same oligosaccharide chain, resulting in the formation of cyclic compounds. Coupling is the reverse reaction in which a cyclodextrin molecule is linked to a linear oligosaccharide chain, producing a longer oligosaccharide chain. In the disproportionation reaction, part of a linear donor-oligosaccharide is transferred to a linear acceptor chain. The saccharifying activity is the hydrolysis of starch into linear oligosaccharides. All assays were standardly performed at pH 6.0 and 60 °C. Cyclization and saccharifying assays were performed as described by Penninga et al. (14). Coupling activity was measured essentially as described by Nakamura et al. (41). -CD (2.5 mM) was used as donor substrate and methyl -D-glucopyranoside (100 mM) as acceptor substrate. The linear oligosaccharide formed in the reaction was converted to single glucose units by the action of amyloglucosidase (Sigma, Darmstadt, Germany). Glucose was detected with the glucose/GOD-Perid method of Boehringer Mannheim (Almere, The Netherlands). Disproportionation activity was measured as described by Nakamura et al. (18). EPS, 4-nitrophenyl--D-maltoheptaoside-4-6-O-ethylidene (3 mM, Boehringer Mannheim), was used as donor substrate and maltose (10 mM) as acceptor substrate. The reaction product containing the nitrophenyl group was cleaved by the action of -glucosidase (Boehringer Mannheim). For each reaction units were defined as the amount of enzyme producing/converting 1 µmol of product/substrate at pH 6.0 and 60 °C.

HPLC Product Analysis-- Formation of cyclodextrins was measured under industrial process conditions by incubation of 0.1 unit/ml CGTase (-CD forming activity) with 10% Paselli WA4 (pregelatinized drum-dried starch with a high degree of polymerization; AVEBE) in 10 mM sodium citrate buffer, pH 6.0, at 60 °C for 45 h. Samples were taken at regular time intervals and boiled for 10 min. Products formed were analyzed by HPLC, using a 25-cm Econosil-NH₂ 10-µm column (Alltech Nederland bv, Breda, The Netherlands) eluted with acetonitrile/water (65:45) at 1 ml/min. Products were detected by a refractive index detector (Waters 410, Waters Chromatography Division, Milford, MA). The temperature of the flow cell and column was set at 50 °C, to avoid possible precipitation of starch. Formation of linear products was directly analyzed. Formation of CDs was analyzed after incubation of the samples with an appropriate amount of -amylase (type I-B from sweet potato, Sigma, Boom BV, Meppel, The Netherlands), degrading linear sugars (but not CDs) to glucose. The retention times for -, -, and -CD were the same as those for G4, G5, and G6 linear oligosaccharides, respectively.

	RESULTS AND DISCUSSION

Top Abstract Introduction Procedures Results & Discussion Conclusions References

Binding of the Maltohexaose Inhibitor

The maltohexaose inhibitor complexed with the CGTase from T. thermosulfurigenes EM1 was bound at subsites 3 to +3 (Fig. 2). In complexes with BC251 CGTase (13) or other CGTases, binding at subsite 3 has never been observed. The present study thus reveals the nature of subsite 3 for the first time. The glucose at subsite 3 (overall B factor: 58 Å²) has long distance interactions with Glu²⁶⁴ O-1 and Thr²⁶² N, both at 3.6 Å. A better contact is formed at 3.4 Å with the Asn⁵⁹¹ O1 from a symmetry related molecule. This stabilization by a crystal contact may explain why in BC251 CGTase crystals a glucose at subsite 3 has never been observed. The interactions that do not result from a crystal contact are very weak, indicating that subsite 3 is not of large relevance.

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Conformation of the maltohexaose inhibitor in the active site of the CGTase from T. thermosulfurigenes EM1. The inhibitor is occupying subsites 3 to +3 in domains A and B of the CGTase.

In the vicinity of the scissile bond the active site architecture is identical in Tabium and BC251 CGTase. It is therefore not surprising that at subsites 2 to +2, the maltohexaose inhibitor is bound in the same fashion to Tabium CGTase as the maltononaose inhibitor to BC251 CGTase (13). Even though the acarbose binding mode has been modeled differently, the same amino acids are providing similar interactions from subsites 2 to +2, suggesting that acarbose can adapt its conformation easily to that required by the active site. It is clear that differences in characteristics between Tabium and BC251 CGTase must originate from interactions at more distant subsites.

In contrast to subsites 2 to +2, at subsite +3 the binding mode of the maltohexaose inhibitor is radically different from the maltononaose binding mode, the former being more bent. In Fig. 3, an overlay of the two inhibitor conformations can be seen. All enzyme-substrate interactions are given in Table II. The glucose at subsite 3 of the maltohexaose inhibitor occupies a position more bent toward Phe¹⁹⁶. This conformation at subsite 3 is stabilized by Lys⁴⁷. In BC251 CGTase this residue is Arg⁴⁷, which so far has never been found to be involved in substrate binding. Furthermore, in the BC251 enzyme Tyr⁸⁹ has strong interactions at subsite 3, but in the Tabium CGTase this residue is an Asp. The conformation of Asp⁸⁹ does not allow any interactions with substrate. Apart from these two differences at subsite 3, residues in Tabium CGTase at subsites 4-7 might be unfavorable to the maltononaose binding mode (straight), although we could not find evidence for that from the structure of Tabium CGTase. The maltohexaose conformation observed is only stabilized by the protein through the contact with Lys⁴⁷. The maltohexaose conformation, however, might have less internal strain because it allows the O2-O3 interglucoside hydrogen bond between subsites 2 and 3 to be formed at 2.7 Å, whereas in the straight conformation the O2-O3 distance is 4.0 Å (13) prohibiting hydrogen bond formation. In addition, the lack of interactions with the enzyme might allow for more flexibility of the maltohexaose chain, which would thus be entropically stabilized. On the basis of these data, we thought it possible to favor one binding mode over the other by site-directed mutagenesis, thereby investigating whether it could be related to one of the Tabium CGTase characteristics.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Superposition of the maltohexaose (sticks) and maltononaose (lines) inhibitor structures. At subsite +3 the conformation of the maltohexaose inhibitor is more bent toward Phe196 and is stabilized by Lys47, which is Arg47 in the CGTase from B. circulans strain 251. Moreover, the replacement of Tyr89 (B. circulans CGTase) by Asp89 (T. thermosulfurigenes EM1 CGTase) makes that the "straight" maltononaose conformation at subsite +3 is not as stably bound in T. thermosulfurigenes EM1 CGTase than as in B. circulans CGTase.

The Tabium and BC251 CGTases are different in many respects. The molecular basis for thermostability of the Tabium CGTase has been extensively discussed (6). Furthermore, Tabium CGTase displays a relatively high hydrolytic activity (24 units/mg), compared with BC251 CGTase (3.5 units/mg), which results in formation of substantial amounts of linear sugars besides cyclodextrins from starch (11). Tabium CGTase produces a mixture of -, -, and -CD at a ratio of 28:58:14, respectively, whereas BC251 CGTase has a product ratio of 13:64:23 (14). A mutant BC251 CGTase with a substantially higher -cyclodextrin production also showed preference for a bent maltohexaose inhibitor over a straightly bound maltononaose inhibitor (43). This suggests a relation between the bent conformation and -cyclodextrin production. The maltohexaose and maltononaose binding modes thus may reflect (part of) specific intermediates for -CD and -CD production by Tabium CGTase and BC251 CGTase, respectively. Further experimental evidence for this was sought by modifying relevant residues in Tabium CGTase, using site-directed mutagenesis. Since the maltohexaose and maltononaose inhibitors are synthesized by CGTase in situ in the crystal, no sufficient quantities were available to determine their binding or inhibitor constants. Therefore we designed our mutants only by qualitative arguments.

Our first approach was to design a mutation that would hinder the maltononaose binding mode and possibly bind a substrate in the maltohexaose conformation. We found that the replacement of Asp¹⁹⁷ by His fulfilled these requirements, since modeling of the mutation D197H (Fig. 4) shows that the His ring cannot assume a conformation in which all the atoms are more than 2.0 Å away from the atoms in the maltononaose inhibitor. So His¹⁹⁷ is likely to block the straight conformation. Moreover, the His¹⁹⁷ N-2 atom could potentially stabilize the bent conformation by forming a hydrogen bond with the glucoside O6 at subsite +3. The D197H mutation changes the electrostatic field of the active site, which could have long distance effects by changing charge-charge interactions. However, since the substrate is uncharged, effects on the substrate binding mode will have to be indirect, in contrast to the van der Waals and hydrogen bonding interactions.

View larger version (139K): [in this window] [in a new window]   Fig. 4.   Model of the mutations D371R and D197H in the CGTase from T. thermosulfurigenes EM1 (green and yellow) together with the maltohexaose structure. The mutations (red) have been given the most frequently occurring side chain conformation (with O; Ref. 28) that doesn't involve steric clashes with the enzyme. The maltohexaose chain is depicted in light blue; the modeled maltononaose chain in the B. circulans 251 CGTase structure is colored dark blue.

Our second approach was to construct a mutant that would stimulate the maltononaose conformation over the maltohexaose conformation, which is easiest achieved, not by constructing a specific interaction with the straight conformation, but by destabilizing the bent conformation. With the mutation Asp³⁷¹ to Arg we aimed at introducing a bulky residue that would clash with the maltohexaose conformation, but be less hindering to the straight maltononaose conformation (Fig. 4). The long and flexible side chain of Arg³⁷¹ could probably extend its effect to subsite +3, where the bent and straight conformations differ most. The mutation D371R changes again the electrostatics of the active site, with possible indirect consequences for substrate binding. However, the mutation conserves the polarity of residue 371.

Of the two mutants D197H, designed to relatively stabilize the maltohexaose conformation, is expected to produce more -cyclodextrin. Mutant D371R, designed to prefer the maltononaose conformation, is expected to produce less -cyclodextrin. The experimental data show that these mutants display altered cyclodextrin product specificity according to expectation (Table III).

View this table: [in this window] [in a new window]   Table III Starch conversion of T. thermosulfurigenes wild-type and mutant CGTase proteins Proteins (0.1 unit/ml -CD forming activity) were incubated for 45 h at pH 6.0 and 60 °C with 10% Paselli WA4.

Production and Purification of CGTase (Mutant) Proteins

Mutants of Tabium CGTase were successfully constructed by site-directed mutagenesis via PCR; all mutations were verified by restriction site analysis (except for mutant F284K) and DNA sequencing. Amounts of 0.4-1.2 mg of pure protein were obtained in a single fermentor run depending on the construct used (Table IV). Purification yields varied between 11% for mutant D371R and 58% for mutant D197H. Purity and molecular weight of the (mutant) CGTases were checked on SDS-polyacrylamide gel electrophoresis. All proteins were purified to apparent homogeneity and displayed a molecular mass of 68 kDa.

Characterization of CGTase Mutant Proteins

Cyclodextrin Product Specificity-- Mutant D197H has a cyclodextrin production profile different from wild-type CGTase (Fig. 5). At the initial stages of the reaction, the mutant has an increased preference for -cyclodextrin production, mostly due to a collapse of the production of -cyclodextrin. At later stages of the reaction, the production of -cyclodextrin increases, but the product ratio is then a result of a subtle equilibrium between cyclization and cyclodextrin breakdown specificities, as well as the solubilities of the diverse cyclodextrins. However, the cyclodextrin production ratio after 45 h still shows a small preference for -cyclodextrin (Table III), proving that the total product ratio can be modified by changing the initial reaction rates. The initial reaction's preference for -cyclodextrin is according to the expectations we had upon designing the mutant (see above). An additional effect of the mutation D197H is that coupling activity is reduced by a factor 4 when compared with the activity of the wild-type enzyme (Table V), possibly because coupling activity was measured with -cyclodextrin whereby a maltononaose was formed as an intermediate. The mutation was, however, designed to make binding of maltononaose less favorable.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Cyclodextrins formed during incubation of (mutant) CGTase proteins from T. thermosulfurigenes EM1 (0.1 unit/ml -CD forming activity) with 10% (w/v) Paselli WA4 starch for 50 h at pH 6.0 and 60 °C. , -CD; , -CD; , -CD. A, wild-type CGTase; B, mutant D197H; C, mutant D371R; D, mutant F284K; E, mutant N327D.

Site-directed Mutations Close to the Proton Donor Glu²⁵⁸: Implications for pH Optima of CGTases-- To further investigate the role of the environment of Glu²⁵⁸ on the pK_a of Glu²⁵⁸ in the different reactions catalyzed by CGTases we replaced Phe²⁸⁴ by Lys (F284K) in Tabium CGTase. Residue Phe²⁸⁴ is located in a hydrophobic cavity immediately above the active site, close to Glu²⁵⁸ (Fig. 6). At most physiological pH values Lys can be expected to bear a positive charge. Positioning of a positive charge near Glu²⁵⁸ stabilizes its deprotonated form, so decreasing its pK_a. This would predict a shift of the pH optimum for this mutant toward acidity. Fig. 7, A and B, show that this indeed happens for both the cyclization and hydrolysis activity. Apart from this effect, the optimum curves also become more narrow, an observation that was also made for a F184L mutant in Bacillus sp. 1011 CGTase (18). The mutation F184L does not introduce any charges; therefore, the narrowing effect can be best explained by a reduction of the hydrophobicity of the environment of Glu²⁵⁸ (see below).

View larger version (106K): [in this window] [in a new window]   Fig. 6.   Model of the mutations near Glu258. Residues in the CGTase from T. thermosulfurigenes EM1 are shown in green and yellow. The maltohexaose structure is given in blue. The mutations N327D and F284K are modeled in red, as described in the legend to Fig. 4. The red sphere is the water as mentioned in the text.

View larger version (18K): [in this window] [in a new window]   Fig. 7.   A, effect of pH on the cyclization activity of CGTase (mutant) proteins from T. thermosulfurigenes EM1. B, effect of pH on the saccharifying activity of the CGTase (mutant) proteins from T. thermosulfurigenes EM1.

pH Optimum Curve for Cyclization and Hydrolysis-- In Tabium CGTase residue Glu²⁵⁸ is the proton donor in the first step of catalysis, so a protonated state of Glu²⁵⁸ is essential. In the second step of catalysis, Glu²⁵⁸ is suggested to activate the acceptor by deprotonation, in this case a deprotonated state of Glu²⁵⁸ would be essential. Furthermore, for optimal activity the catalytic nucleophile, Asp²³⁰, must remain deprotonated in the first step of catalysis. The Tabium CGTase displays a broad pH optimum for cyclization in the range of pH 4-7 (Fig. 7A). Other CGTases have an efficient cyclization reaction from pH 5 to 7 (16, 18, 21, 22). It might be expected that the drop in enzyme activity at high pH is caused by deprotonation of Glu²⁵⁸ in the first step of the reaction and that the activity drop at low pH is caused by protonation of Asp²³⁰. The combination of these two effects would result in a pH optimum curve. In Fig. 7A, however, it can be seen that the mutations near Glu²⁵⁸ shift both the slopes at high and low pH, while we had expected only to see effects on the high pH slope of the optimum curve, which is affected by Glu²⁵⁸. Similar observations have been described before in literature. The mutation F284L in Bacillus sp. 1011 CGTase shifts the pH optimum both over acidic and alkaline pH ranges (18). This mutation is close to the proton donor and is unlikely to have long distance effects. Mutation of Asp³²⁹ in BC251 CGTase to Asn had a most pronounced effect on the low pH slope of the optimum curve (20). Furthermore, mutation of Asp²³⁰ to Asn in the BC251 CGTase did not result in any shift of the pH optimum, whereas a shift was observed with the Glu²⁵⁸ to Gln mutation (20). Therefore, it is likely that the protonation state of Glu²⁵⁸ determines both slopes of the pH optimum curve. At high pH, the first step of the double displacement mechanism is hindered by lack of a proton donor, at low pH the second step is inhibited by incomplete substrate activation. If the protonation state of Glu²⁵⁸ determines both slopes of the pH profile, it is unexpected that the pH optimum is so broad. This broad pH range of activity must be explained by a different environment for Glu²⁵⁸ in the first and second step of catalysis. Indeed, in Tabium CGTase (6) the third catalytic residue Asp³²⁹ probably forms a hydrogen bond with Glu²⁵⁸ initially, increasing the pK_a of Glu²⁵⁸ and ensuring its protonation. This bond is broken after substrate binding, facilitating again deprotonation of Glu²⁵⁸ in the ensuing second catalytic step. The impact of changes near Glu²⁵⁸ during catalysis is enhanced by its hydrophobic environment, in which electrostatic interactions are much stronger. Changes in the environment of Glu²⁵⁸ before and after substrate binding have also been observed in the structure of BC251 CGTase (13, 16).

	CONCLUSIONS

Top Abstract Introduction Procedures Results & Discussion Conclusions References

At subsite 3 the binding mode of the maltohexaose inhibitor in Tabium CGTase is radically different from the maltononaose binding mode in the BC251 CGTase. The bent conformation of the maltohexaose inhibitor, in contrast to the straight conformation of the maltononaose inhibitor, was found to be correlated to enhanced -cyclodextrin production. Mutations stabilizing the bent conformation but hindering the straight conformation resulted in enhanced production of -CD, whereas mutations hindering the bent conformation but stabilizing the straight conformation resulted in decreased production of -CD. The Tabium CGTase can hence be changed from an /-cyclodextrin producer to a /-cyclodextrin producer by a single mutation, illustrating the feasibility of CGTase protein engineering.

Mutations near the proton donor Glu²⁵⁸ suggest that the pH optimum curve of CGTase may be determined only by the protonation state of residue Glu²⁵⁸. Both the high and low slopes of the pH optimum curve could be manipulated by site-directed mutations close to Glu²⁵⁸. Changes in the environment of Glu²⁵⁸ before and after substrate binding can account for its broad pH optimum.