On the mechanism of a-amylase Acarbose and cyclodextrin inhibition of barley amylase isozymes Naı¨ma Oudjeriouat 1 , Yann Moreau 2 , Marius Santimone 1 , Birte Svensson 3 , Guy Marchis-Mouren 1 and Ve ´ ronique Desseaux 1 1 IMRN, Institut Me ´ diterrane ´ en de Recherche en Nutrition, Faculte ´ des Sciences et Techniques de St Je ´ rome, Universite ´ d’Aix-Marseille, France; 2 IRD, Institut de Recherche pour le De ´ veloppement, UR081 Gamet c/o CEMAGREF Montpellier, France; 3 Carlsberg Laboratory, Department of Chemistry, Copenhagen Valby, Denmark Two inhibitors, acarbose and cyclodextrins (CD), were used to investigate the active site structure and function of barley a-amylase isozymes, AMY1 and AMY2. The hydrolysis of DP 4900-amylose, reduced (r) DP18-malto- dextrin and maltoheptaose (catalysed by AMY1 and AMY2) was followed in the absence and in the presence of inhibitor. Without inhibitor, the highest activity was obtained with amylose, k cat /K m decreased 10 3 -fold using rDP18-maltodextrin and 10 5 to 10 6 -fold using maltohep- taose as substrate. Acarbose is an uncompetitive inhibitor with inhibition constant (L 1i ) for amylose and maltodextrin in the micromolar range. Acarbose did not bind to the active site of the enzyme, but to a secondary site to give an abortive ESI complex. Only AMY2 has a second secon- dary binding site corresponding to an ESI 2 complex. In contrast, acarbose is a mixed noncompetitive inhibitor of maltoheptaose hydrolysis. Consequently, in the presence of this oligosaccharide substrate, acarbose bound both to the active site and to a secondary binding site. a-CD inhibited the AMY1 and AMY2 catalysed hydrolysis of amylose, but was a very weak inhibitor compared to acarbose. b-andc-CD are not inhibitors. These results are different from those obtained previously with PPA. However in AMY1, as already shown for amylases of animal and bacterial origin, in addition to the active site, one secon- dary carbohydrate binding site (s 1 ) was necessary for activity whereas two secondary sites (s 1 and s 2 )were required for the AMY2 activity. The first secondary site in both AMY1 and AMY2 was only functional when sub- strate was bound in the active site. This appears to be a general feature of the a-amylase family. Keywords: amylose; maltodextrin; acarbose; barley a-amy- lase; binding site. a-Amylase is a retaining glycoside hydrolase of family 13 acting on a-1,4 internal glycoside linkages in starch and related sugars [1]. a-Amylases occur widely in higher plants, animals, bacteria and fungi and are applied in several important industries, e.g. in starch processing, paper treatment, pharmaceutical and the food manufacturing [2–4]. Cereal a-amylases, such as barley isozymes AMY1 and AMY2, play an essential role during seed germination (malting) by hydrolysing the storage starch granules present in the endosperm. AMY1 and AMY2 have 80% sequence identity [5,6]. AMY1 was more active toward starch granules and more stable at low pH, while AMY2, the major isozyme, was more active toward nitrophenylated maltooligosaccharides and was inhibited by the proteina- ceous barley a-amylase/subtilisin inhibitor (BASI) to which AMY1 is insensitive [7,8]. Subsite mapping showed that the substrate binding cleft of both isozymes contains 10 consecutive subsites recogni- zing substrate glucose residue, i.e. six toward the nonreduc- ing end and four toward the reducing end relative to the bond to be cleaved [9]. The AMY1 and AMY2 active sites are twice as long as that of the human and porcine enzymes containing only five subsites [9–12]. In addition, a noncata- lytic site that facilitated adsorption onto starch granules (and most probably also hydrolysis of starch granules) has been found in barley a-amylase [7,13,14]. Binding of b-cyclodextrin at this site inhibits the a-amylase catalysed hydrolysis of starch granules, but no inhibition was observed with soluble substrate [15,16]. In AMY2, differ- ential labelling of tryptophan residues using b-CD for protection identified Trp276-Trp277 in this binding site [13]. Trp206 belongs to the active site where it is situated at subsite +2 [14]. Known crystal structures of a-amylases contain a central catalytic (b/a) 8 barrel domain (domain A) having an irregularly structured small domain B protruding between b-strand 3 and a-helix 3 of the barrel, and a C-terminal, domain C, folded as an antiparallel b-sheet [17–23]. Acarbose is a pseudotetrasaccharide inhibitor of a-amylase, that acts like a transition-state analogue [7] and Correspondence to V. Desseaux, IMRN case 342, Faculte ´ des Sciences et Techniques, Avenue. Esc. Normandie-Niemen, 13397 Marseille cedex 20, France. Fax: + 33 4 91 28 84 40, E-mail: veronique.desseaux@univ.u-3mr.fr Abbreviations:AMY,barleya-amylase; AMY1, barley a-amylase isozyme 1; AMY2, barley a-amylase isozyme 2; PPA, porcine pan- creatic a-amylase; CD, cyclodextrin; DP, degree of polymerization; rDP18, reduced DP18-maltodextrin; G7, maltoheptaose. Enzyme: a-amylase [a(1,4)-glucan-4-glucanohydrolase; EC 3.2.1.1]. Note: This paper is dedicated to the late Prof. E. Prodanov (Montevideo, Uruguay). (Received 17 March 2003, revised 23 May 2003, accepted 30 June 2003) Eur. J. Biochem. 270, 3871–3879 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03733.x binds to the active site [2,14,17]. The crystallography of AMY2/acarbose showed that both the active site, contain- ing Trp206, and the secondary so-called starch granule binding site at the surface, containing Trp276-Trp277, bind acarbose [14]. This surface binding site revealed a charac- teristic stacking of a disaccharide unit from acarbose onto the Trp residues [14]. The starch binding site was required when acting on insoluble substrates such as starch granules. Previously, acarbose was demonstrated to be a mixed noncompetitive-type inhibitor of the hydrolysis of amylose, rDP18-maltodextrin and maltopentaose catalysed by por- cine pancreatic [24–28] and human [29] a-amylases. Depending on the substrate, one or two secondary carbo- hydrate binding site(s) were found which became functional upon substrate binding. These sites may be involved in the catalytic process and/or in product release [24]. The same inhibition type using amylose as substrate was also reported using amylases from a fish (Tilapia) [30] and a bacterium (Lactobacillus) [31]. The a-amylase mechanism for hydro- lysis of soluble substrates includes several steps (a) internal binding to the amylose chain (b) splitting of the chain (c) and according to the multiple attack hypothesis [32] further hydrolysis near the reducing end of the nonreducing moiety of the initially cleaved amylose to liberate successively 1, 2, 3, etc. molecules of maltose or longer oligosaccharide(s). Secondary binding sites are probably required in such a mechanism for binding and sliding of the substrate chain. Barley AMY1 and AMY2 have a degree of multiple attack toward amylose of two (B. Kramhøft and B. Svensson, unpublished data). The goal of the present work is to characterize further the AMY1 and AMY2 function toward soluble sub- strates. The kinetics of hydrolysis of substrates of different length: i.e. DP 4900-amylose, rDP18-maltodextrin and maltoheptaose, in the presence and in the absence of the inhibitor acarbose, respectively, of the potential inhibitors a-, b-andc-cyclodextrin are reported. Using a statistical analysis of the data, the inhibitory mechanism is investi- gated. Moreover the present results are compared with those obtained recently in our laboratory using amylases from different species (porcine [24–28], human [29], Tilapia [30] and Lactobacillus [31]). The inhibitor and the inhibition type characterize the active site of the different enzymes and the secondary site(s) needed for soluble substrate(s) which appear(s) to be a general feature of a-amylases. Materials and methods Materials Barley a-amylases, AMY1 and AMY2, were purified from green and kilned malt, respectively, according to Svensson et al. [33] and Ajandouz et al.[9].PurifiedAMY1and AMY2 gave single bands in SDS/PAGE (not shown) in amounts corresponding to approximately 5 and 100 mgÆL )1 . The amylase concentrations were determined by measuring A 280 (A 1% 280 ¼ 24) [24]. Amylose (type III from potato) DP 4900 (794 kDa) [34], maltoheptaose, maltohexaose, maltopentaose, maltotetraose, maltotriose, maltose, glucose and neocuproin hydrochloride were from Sigma. Maltodextrin of average DP18 (2.9 kDa) was from Hayashibara Biochemical Laboratories (Okayama, Japan). Reduction of the DP18-maltodextrin to the corresponding alcohol was performed as earlier described [35] by using NaBH 4 . This was done to facilitate the reducing sugar assay by minimizing the contribution from the substrate to achieve low blank values. Acarbose (O-4,6-dideoxy-4-{[4,5,6-trihydroxy-3-hydroxymethyl-2-cyclo- hexen-1-yl]amino}-a- D -glucopyranosyl-(1 fi 4)-O-a- D -gluco- pranosyl-(1 fi 4)- D -glucose) was generously supplied by Bayer Pharma (France). a-, b-andc-cyclodextrins were from Sigma. Kinetics Kinetic experiments were performed at 30 °Cin20m M sodium acetate buffer (pH 5.5) containing 1 m M CaCl 2 and 1 m M sodium azide. Substrate, inhibitor and buffer were mixed and the reaction was initiated by adding the enzyme. When amylose or rDP18-maltodextrin was the substrate, the incubation volume was 400 lL and the enzyme volume 100 lL. More than 10 concentrations of the substrates, amylose (0.003–0.32 gÆL )1 or 0.038–0.4 l M for AMY1; 0.048–0.8 gÆL )1 or 0.06–1 l M for AMY2) and rDP18- maltodextrin (0.06–1.46 gÆL )1 or 20–500 l M for both isozymes) were used. The final concentration of AMY1 and AMY2 was 2.0 n M and 1.0 n M , respectively. Acarbose was used in the range 10–80 l M and a-, b-andc-CD were in the ranges 2–20 m M , 3.2–24 m M , and 1.6–13.6 m M , respectively. The reaction was stopped at appropriate time intervals (1, 3 and 5 min) by adding 500 lL of chilled 0.38 M sodium carbonate containing 1.8 m M cupric sulfate and 0.2 M glycine (500 lL)andkeptonice[36].Therate of hydrolysis of amylose and rDP18-maltodextrin was obtained from the increase in reducing power and using maltose as standard. When maltoheptaose was the substrate, the incubation volume was 900 lL and the enzyme volume 100 lL giving a final concentration of 100 n M . More than 10 concentrations of maltoheptaose (0.15–5 m M ) were used. Acarbose was in the range 0.75–5 m M .Samples(100lL) were removed at appropriate time intervals (0, 0.15, 0.30, 0.45 and 1.00 min), addedto0.1 M NaOH (300 lL) to stop the reaction, and kept on ice until analysis. The rate of hydrolysis was determined by measuring the produced maltooligosaccha- rides by high-performance anion-exchange chromatogra- phy (HPAEC) on a Carbopac PA-100 (4 mm · 250 mm) column with elution by a 5–500 m M sodium acetate linear gradient over 20 min in 100 m M NaOH, at a flow rate of 1.0 mLÆmin )1 . Detection of oligosaccharide and glucose in the eluate was performed by pulsed amperometric detection (PAD) using the Dionex DX-500 chromatograph as reported previously [25]. For quantification glucose, malt- ose, maltotriose, maltotetraose, maltopentaose, maltohexa- ose and maltoheptaose were used as standards. Values from either reductometry or HPAEC-PAD gave initial velocities as calculated from the slopes obtained by linear regression of the linear part of the progress curves, which in turn gave the number of glycoside bonds hydrolysed per minute or the amount of product (glucose and maltohexaose) released per minute, respectively. The experiments were repeated three or four times. 3872 N. Oudjeriouat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Statistical analyses of kinetics experiments Statistical analyses were performed using either the REG, NLIN, or GLM procedure from the SAS / STAT software package (Sas Institute Inc, Cary, NC, USA) [37]. A significance level of 0.05 was used in all statistical tests. The initial velocity was measured at fixed inhibitor and varying substrate concentration. In order to determine the type of inhibition, the kinetic data were analysed using a general initial velocity equation. As discussed earlier [9,27,28], Eqn (1) applies for the present type of data: v=½E 0 ¼ k cat ½S K m ð1 þ 1 K li ½Iþ 1 K li K 2i ½I 2 Þþ½Sð1 þ 1 L li ½Iþ 1 L li L 2i ½I 2 Þ ð1Þ In this equation v is the initial velocity, [E] 0 the enzyme concentration, [S] the substrate concentration, [I] the inhibitor concentration, K m the Michaelis constant and K 1i , K 2i , L 1i , L 2i the dissociation constants of the different abortive complexes, EI, EI 2 , ESI and ESI 2 , respectively, as shown in the scheme below. It should be noticed that this equation applies at steady state and at rapid equilibrium except in the case of noncompetitive inhibition with a random mechanism at steady state [24]. Equation (1) corresponds to the following reaction scheme, where Q and P are the products: A nonlinear statistical analysis was used. Equation (1) was modified by using the association constants K¢ 1i , K¢ 2i , L¢ 1i and L¢ 2i which are the inverse of the corresponding dissociation constants. In this equation, it was easier for a calculated constant to compare its value relative to zero rather than to obtain a large value. When the association constant value was close to zero, this meant that the corresponding abortive complex was not present in significant amounts. Actually, one will use the simplest equation which best matched the data and the actual inhibition type. Difference spectroscopy Difference spectra were determined using a double-beam Shimadzu UV-2401PC spectrophotometer. Double-com- partment cells (each 0.44 cm light path, 230-QS, from Hellma) were used for both control cell and sample cell. The cells were thermostated at 30 °C. First, both cells were filled with 20 m M sodium acetate buffer (pH 5.5) containing 1m M CaCl 2 and 1 m M sodium azide to define the baseline. Second, AMY1 (40 l M ) was introduced into one compart- ment of the control and one compartment of the sample cell and the reference line was determined (A 0 ). Then acarbose (1.7–6.5 m M ) was added to the buffer compartment of the control and to the compartment containing AMY1 in the sample cell. The AMY1 concentration in the control cell was adjusted accordingly by addition of buffer. Spectra were recorded in the 230–320 nm region at a rate of 0.2 nmÆs )1 . Results Determination of kinetic parameters with substrates of different sizes The AMY catalysed hydrolysis of DP 4900-amylose, rDP18-maltodextrin and maltoheptaose was first measured in the absence of inhibitor. Statistical analysis of the experimental initial rates (v) was performed using the general Michaelis–Menten initial velocity equation for determination of k cat and K m and calculation of the catalytic efficiency, k cat /K m . The kinetic parameters of AMY1 and AMY2 (Table 1) were rather similar, but depended import- antly on the substrate. With amylose and rDP18-maltodex- trin as substrates, under saturating conditions, no difference was observed between k cat of AMY1 and AMY2 for amylose or for rDP18-maltodextrin (Table 1). In contrast, however, for maltoheptaose, AMY2 had three times higher k cat than AMY1. The K m values were increasing with decreasing substrate length from around 0.2 l M for amy- lose, to around 215 l M for maltoheptaose. AMY1 and AMY2 (k cat /K m ) were 700 to 1000-fold more active toward amylose than rDP18-maltodextrin, which in turn was 170 to 690-fold superior as substrate than maltoheptaose (Table 1). Thus the longer the substrate, the higher was the activity. Inhibition by acarbose Inhibition of amylose hydrolysis occurred in the presence of 10–80 l M acarbose and the association constants K¢ 1i , K¢ 2i , L¢ 1i and L¢ 2i were determined according to the general equation (see Materials and methods). For both AMY1 and AMY2, the association constants K¢ 1i ,K¢ 2i and L¢ 2i were close to zero and could not be determined under these conditions while, L¢ 1i ¼ (62 ± 4)10 3 M )1 and L¢ 1i ¼ (28 ± 3)10 3 M )1 , respectively. The dissociation constants, calculated from the respective association constants in the corresponding equation (K 1i ,K 2i , L 1i and L 2i ), were given in Table 2. When the association constant values were close to zero, the significant values of the dissociation constants K 1i ,K 2i ,andL 2i could not be obtained (NS). The closest match to the experimental data corresponded to Eqn (2). Table 1. The enzyme kinetic parameters of hydrolysis of different sub- strates by barley a-amylase isozymes AMY1 and AMY2. Parameter values are given as ± SEM. Substrate Enzyme k cat (s )1 ) K m (l M ) k cat /K m (s )1 Æ M )1 ) Amylose AMY1 206 ± 12 0.21 ± 0.03 1.0 · 10 9 AMY2 202 ± 10 0.16 ± 0.02 1.3 · 10 9 Maltodextrin AMY1 129 ± 5 79.3 ± 9.8 1.6 · 10 6 AMY2 125 ± 4 71.4 ± 7.0 1.8 · 10 6 Maltoheptaose AMY1 2.02 ± 0.10 213 ± 46 9.5 · 10 3 AMY2 5.62 ± 0.5 217 ± 43 26 · 10 3 Ó FEBS 2003 Acarbose inhibition of barley a-amylases (Eur. J. Biochem. 270) 3873 v=½E 0 ¼ k cat ½S K m þ½Sð1 þ L 0 1i ½IÞ ð2Þ Eqn (2) represents the following reaction scheme of uncompetitive inhibition, in which the dissociation constant has been indicated: This model included only one abortive complex ESI (I bound at a secondary site s 1 ) and no significant amount of acarbose was bound to E as in an ES complex. The reciprocal plot drawn for AMY1 according to Eqn (2) illustrates this model: parallel straight lines intersect the ordinate axis as expected, the intercept increasing with increasing acarbose concentration (Fig. 1). A similar plot was obtained for AMY2 (not shown). Inhibition of the rDP18-maltodextrin hydrolysis occurred also in the presence of 10–80 l M acarbose. For both AMY1 and AMY2 the association constants K¢ 1i and K¢ 2i were close to zero and L¢ 1i ¼ (67 ± 9) 10 3 M )1 , L¢ 1i ¼ (42 ±7) 10 3 M )1 , respectively. For AMY1, L¢ 2i value was also close to zero. In this case Eqn (2) applies. For AMY2, L¢ 2i ¼ (11 ± 5) 10 3 M )1 and in this case, Eqn (3) accounted for the data. With both enzymes, the inhibition was as above ) the uncompetitive type. The resulting dissociation constants are given in Table 2. v=½E 0 ¼ k cat ½S K m þ½Sð1 þ L 0 1i ½IþL 0 1i L 0 2i ½I 2 Þ ð3Þ The corresponding reaction scheme is: indicating no EI complex in significant amount but two complexes, ESI (I bound at s 1 )andESI 2 (I bound at s 1 and s 2 ), to be present. The inhibition is still uncompetitive and the plot drawn with AMY1 illustrates this model: parallel straight lines intersected the ordinate (Fig. 2). A similar plot was obtained for AMY2 (not shown). To summarize, rDP18-maltodextrin hydrolysis by both AMY1 and AMY2 was uncompetitively inhibited by acarbose. For AMY1, however, only the ESI inhibition complex was present, while with AMY2 both ESI and ESI 2 were formed. In contrast, when maltoheptaose is the substrate in the presence of 0.75–5 m M acarbose, the experimental data most closely matched Eqn (4): v=½E 0 ¼ k cat ½S K m ð1 þ K 0 1i ½IÞ þ ½Sð1 þ L 0 1i ½IÞ ð4Þ For AMY1, calculation gave the association constants K¢ 1i ¼ (5.2 ± 1.4) 10 3 M )1 and L¢ 1i ¼ (0.25 ± 0.09) 10 3 M )1 , K¢ 2i and L¢ 2i were close to zero. Using AMY2, K¢ 1i ¼ (1.2 ± 0.3) 10 3 M )1 and L¢ 1i ¼ (1 ± 0.26) 10 3 M )1 , K¢ 2i and L¢ 2i were also close to zero. The dissociation constant values of K 1i and L 1i are shown in Table 2. This inhibition was of the mixed noncompetitive type for both isozymes and followed the reaction scheme: Fig. 1. Lineweaver–Burk plots. AMY1 with varying amylose and fixed acarbose concentration [I] as indicated. This plot was calculated by statistical analyses of initial rates of hydrolysis using Eqn (2). Gra- phical analysis was not possible with our data. For this reason no experimental points are reported. The plot is drawn from the corres- ponding rate equation determined by statistical analysis. Fig. 2. Lineweaver–Burk plots. AMY1 with varying rDP18-malto- dextrin concentration and fixed acarbose concentration [I] as indicated. This plot was calculated by statistical analyses of initial rates using Eqn (2). Graphical analysis was not possible with our data. For this reason no experimental points are reported. The plot is drawn from the corresponding rate equation determined by statistical analysis. Table 2. The inhibition constants and type of inhibition by acarbose for AMY1 and AMY2 acting on different substrates. K 1i , K 2i ,L 1i and L 2i are the EI, EI 2 ,ESIandESI 2 related dissociation constants. NS, not significant values. Substrate Enzyme K 1i (l M ) K 2i (l M ) L 1i (l M ) L 2i (l M ) Inhibition type Amylose AMY1 NS NS 16 NS Uncompetitive AMY2 NS NS 36 NS Maltodextrin AMY1 NS NS 15 NS AMY2 NS NS 24 95 Maltoheptaose AMY1 194 NS 4 10 3 NS Mixed AMY2 833 NS 1 10 3 NS Noncompetitive 3874 N. Oudjeriouat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 Two abortive complexes were present: EI (I bound at the active site) and ESI (I bound at a secondary site s 1 ). The reciprocal plot using estimated values drawn for AMY1 from Eqn (4) illustrated this model: straight lines intersected in the 2nd quadrant (Fig. 3) at a point close to, but distinct from the origin (see insert). A similar plot was obtained with AMY2 (not shown). This apparent discrep- ancy from the inhibition of the long chain substrate hydrolysis was associated with the weak affinity of malto- heptaose for the active site as well as an effect also of the high acarbose concentrations used. Inhibition by cyclodextrins In the presence of a-, b-orc-cyclodextrin (2–20 m M ,3.2– 24 m M and 1.6–13.6 m M , respectively) and using DP-4900 amylose, as substrate inhibition of AMY1 and AMY2 only occurred with a-CD which was a very poor inhibitor compared to acarbose. The inhibition constants (not given) were in the 10–100 millimolar range. No other substrates were investigated with the cyclodextrins. Inhibition of starch granule hydrolysis by b-cyclodextrin has previously been reported, however, in agreement with our result on amylose, soluble starch hydrolysis was not inhibited [15]. Difference spectra of acarbose binding The inhibition of the amylolytic activity of AMY1 by acarbose involved, as shown above, specific interactions at either the active site, as in EI, and/or at the secondary binding site. The binding of acarbose to AMY1 was also monitored by UV difference spectroscopy. A complete AMY2 study, however, has not been performed. The absorbance difference spectra of AMY1 produced by 1.7– 6.5 m M acarbose showed a major peak at 294–295 nm, except for the ÔdÕ spectrum (Fig. 4A) which for unknown reasons was slightly shifted toward a shorter wavelength. These spectra indicated that binding of acarbose perturbed at least one tryptophan residue [38,39]. The size of the peak increased with increasing acarbose concentration and was stable for up to 30 min. A shift at 294 nm occurred at longer incubation times and analysis by HPAEC of acarbose- AMY1 mixtures from the sample cell indicated that slow hydrolysis of acarbose took place. This showed that acarbose was bound at the active site. The UV difference spectra therefore were recorded within less than 30 min after mixing. The reciprocal of the normalized absorbance difference [E] 0 /DA([E] 0 ¼ AMY1 initial concentration; DA ¼ A ) A 0 ) measured at 294 nm was plotted against 1/[I] 0 ([I] 0 ¼ acarboseinitialconcentration) yieldingastraight line (Fig. 4B) indicating that one molecule of acarbose (I) binds to one molecule of AMY1 (E) to form the monitored AMY1-acarbose complex (EI) according to the reaction: as a consequence, the following equation applies: ½E 0 DA ¼ K d De  1 ½I 0 þ 1 De ð5Þ Fig. 4. UV difference spectroscopy of AMY1 with acarbose. (A) Scans from 270 to 320 nm are shown. The acarbose concentration (in m M ) was 0.00 (a), 1.70 (b), 2.70 (c), 4.60 (d), 6.5 (e). The AMY1 concen- tration [E] 0 was 38.8 l M decreasing to 37.3 l M by addition of acar- bose. A 0 is the AMY1 absorbance without acarbose, A is the absorbance measured at the above acarbose concentrations. (B) Reciprocal plot of the difference spectra [E] 0 /(A ) A 0 )vs.1/[I] 0 (acarbose initial concentration) measured at 294 nm upon adding acarbose to AMY1. Fig. 3. Lineweaver-Burk plots. AMY1 with varying maltoheptaose and fixed acarbose concentration [I] as indicated. This plot was cal- culated by statistical analyses of initial rates using Eqn (4). The insert enlarges the origin region. Graphical analysis was not possible with our data. For this reason no experimental points are reported. The plot is drawn from the corresponding rate equation determined by statistical analysis. Ó FEBS 2003 Acarbose inhibition of barley a-amylases (Eur. J. Biochem. 270) 3875 in which DA is the absorbance difference, De is the difference between the molar absorption coefficients of the inhibitor complex and the free enzyme, and K d is the dissociation constant of the EI complex. Equation (5) is of first order with respect to 1/[I] 0 and therefore fits a linear plot. It should be noted that Eqn (5) applies only when the concentration of the inhibitor, I, is much higher than that of the enzyme, which was the case in the present experiment. Moreover, if more than one molecule of inhibitor binds to the enzyme and perturbed the spectrum, then the resulting plot [E] 0 /DAvs.1/[I] 0 will not be linear [27]. Equation (5) andFig.4Bwereusedtodeterminethedissociation constant for EI to K d ¼ 0.6 m M which confirmed a previous determination of the binding constant to AMY1 [7]. However, in the light of the present data our interpret- ation of the data was somewhat different. It appeared that the EI complex was observed by difference spectroscopy when the concentration [I] was very high when compared to inhibitor concentrations used in the kinetics studies. The binding of inhibitor at the active center was supported by the fact that acarbose was slowly hydrolysed to release glucose in a reaction that followed linear kinetics (not shown). The question then arose, why the two sites, the active site and the surface site found by kinetic analysis, were not both revealed by the difference spectroscopy. Discussion As shown from the kinetic results obtained in the absence of inhibitor, amylose was by far the best substrate of barley amylase. Actually, rDP18-maltodextrin was hydrolysed at a 10 3 -fold lower rate and maltoheptaose at 10 5 )10 6 lower rate than DP 4900-amylose. AMY2 was only slightly more active than AMY1. This finding agreed with the generally accepted feature that a-amylases are mostly active on long chain substrate. The poor activity of AMY1 and AMY2, relatively speaking, using maltoheptaose as a substrate is due, on the one hand to the fact that maltoheptaose at most occupied 7/10 subsites of the active site in productive complexes and on the other hand because nonproductive complexes would inhibit barley a-amylase catalysed malto- heptaose hydrolysis [9]. When discussing the kinetic results obtained in the presence of inhibitors, the main question to be asked is why acarbose apparently did not occupy the active site of AMY1 and AMY2 when the substrates used are amylose and maltodextrin; while in all other situations, as shown by difference spectra, X-ray crystallography and for PPA, acarbose was bound at the active site. A second point is then how to explain that EI was formed when maltoheptaose was the substrate. How consistent were these corresponding data and what was the contribution to the knowledge of the barley isozymes and to the a-amylase family? As will be discussed further, amylose and rDP18- maltodextrin most probably have significantly higher affinity for the active site of AMY (E) than found for acarbose. When the substrates and the inhibitor compete for the active site, the high affinity of the substrates facilitates their binding whereas the inhibitor binding does not occur. Therefore, no significant amount of EI complex was formed and the acarbose inhibition was uncompeti- tive. The ES complex reacted to give either products or the abortive ESI complex (I bound at s 1 ). When AMY1 was used with the substrates amylose or rDP18-malto- dextrin, only one acarbose molecule was bound to ES as well as to ESI. In the case of rDP18-maltodextrin/AMY2, however, an additional acarbose molecule was bound to give ESI 2 suggesting that one more sugar binding site (s 2 ) was present on the enzyme surface (Fig. 5A). Such a site was found in PPA [24]. We suggest that this second surface site reflected a certain structural difference between AMY1 and AMY2. To summarize, we propose on the basis of the above kinetic results, that one secondary binding site (s 1 )inAMY1andtwo(s 1 and s 2 )inAMY2 were necessary for enzyme activity. It (they) became functional only when S was bound at the active site and were thus quite different from the starch granule binding site earlier characterized in cereal amylases. In the uncompetitive model, no inhibitor was present at the active site. This, however, did not contradict the X-ray data [14] and the present difference spectra. The kinetic results showed that acarbose was a poor inhibitor of AMY1 having a poor affinity for the active site. Conse- quently at the inhibitor concentrations [I] used, no EI complex was formed. At higher concentrations of acar- bose, as used for the difference spectroscopy, the EI complex could form and in accordance with the modest affinity, the dissociation constant was very high (0.6 m M ) (Fig. 5B). Also, the acarbose concentration (10 m M ), used for soaking crystals of AMY2 to get the acarbose/AMY2 complex, was very high [14]. In contrast to the uncompetitive inhibition with amylose and maltodextrin, the inhibition with maltoheptaose was of the mixed noncompetitive type. Thus, both the EI and ESI complexes were formed (Fig. 5A). The noncompetitive acarbose inhibition may result, firstly because maltohepta- ose was a poor substrate for which AMY1 and AMY2 showed, respectively, 10 5 and 10 2 -fold lower catalytic efficiency than for amylose and maltodextrin. With com- petitive binding to enzyme of the substrate and of the inhibitor, the weak affinity of maltoheptaose to enzyme (E) facilitated the binding of acarbose (I) to allow the formation of the abortive EI complex, the ES and ESI (I in s 1 ) complexes being also formed (Eqn 4 and Fig. 5). It can be concluded that the low affinity of both acarbose and maltoheptaose for the active site was associated with noncompetitive inhibition, while uncompetitive inhibition as a consequence of amylose and maltodextrin binding with high affinity for AMY. a-CD was a weak inhibitor of AMY catalysed amylose hydrolysis. b-andc-CD, however, were not inhibitory. In contrast a-, b-andc-CD were all inhibitors of PPA, and active at a slightly lower concentration in 0.25–5 m M range. Such difference most likely reflected the different structures at the active site of PPA [12] and AMY [9]. Two questions arose from the results of difference spectroscopy of acarbose binding: (a) in the observed EI complex, which binding site was then occupied? Our results support that in EI, acarbose occupied the active site as at prolonged incubation acarbose hydrolysis took place. This experiment was of major interest as it allowed determination of the K d (the dissociation constant) of EI which could not be obtained by the kinetics approach when amylose or rDP- 18 maltodextrin were used as substrates. The K d was 3876 N. Oudjeriouat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 actually in the same range as the K 1i obtained with maltoheptaose as substrate (0.2 m M ); (b) why do we not observe the secondary binding site demonstrated kinetic- ally? Two answers may be proposed: either this site was not functional (accessible) in the absence of substrate, as postulated in the conclusion, or acarbose did not bind to a Trp but to a different residue which could not be monitored by UV difference spectroscopy. As mentioned above, similar studies have been conduc- ted on PPA. The results were strikingly different from those obtained with barley AMY. Acarbose was a noncompetitive inhibitor for PPA and an uncompetitive for AMY when long chain substrates were used. In that case, the inhibitory complex ESI was formed with both enzymes, however, the EI complex was observed only with PPA. This discrepancy was explained by the higher affinity of acarbose for PPA as indicated by the lower dissociation constant of the acarbose-PPA complex (1.7 l M )[24].The dissociation constant of the acarbose-AMY complex cannot be determined kinetically but was obtained from the difference spectroscopy analysis (K d ¼ 0.6 m M ). Such a large difference probably reflects differences of the struc- ture and the energetics profiles of the respective active sites. The comparison of AMY and PPA active site showed large differences in the binding affinities of corresponding subsites [9,10,12]. The PPA active site, moreover, had five subsites and acarbose can occupy four of these, while the AMY active site had 10 subsites and this crevice was thus far from completely occupied by acarbose, and acarbose apparently binds with lower affinity. Acarbose was thus demonstrated to be a useful tool in describing active sites in different a-amylases. Fig. 5. Schematic mechanism for the AMY action of acarbose inhibition and binding. (A) Kinetics: S ¼ amylose, rDP18-maltodextrin or malto- heptaose with I ¼ acarbose; S ¼ amylose with I ¼ a-CD. K 1i , L 1i ,L 2i are dissociation constants. (B) Difference spectra. Kd is the dissociation constant. Ó FEBS 2003 Acarbose inhibition of barley a-amylases (Eur. J. Biochem. 270) 3877 Conclusion Barley isozymes AMY1 and AMY2 were thousand-fold more active toward amylose than toward maltodextrin and a million-fold more active than toward maltoheptaose. AMY2 was slightly more active than AMY1. AMY1 and AMY2 were inhibited by acarbose. a-CD was a weak inhibitor and b-andc-CD were not inhibitory. This is in contrast to the high inhibitory toward porcine [24–28] and human [29] a-amylases. Also the inhibitory mechanism by acarbose of the amylose and maltodextrin hydrolysis was of a different type in the barley, compared to the human and porcine enzymes. This different behaviour most probably reflects the individual active site structures. Moreover, in addition to the active site, the presence of one (s 1 )ortwo(s 1 and s 2 ) secondary carbohydrate binding sites already found in amylases from other species were demonstrated. Alto- gether three to four carbohydrate binding sites were postulated: (a) the starch granule binding site [40] [14]; (b) the active site; (c) and one and sometimes two secondary site (s) as deduced from the inhibition kinetics ([24] and the present work). The precise functions of each site are unknown but remarkably, the inhibition kinetics demon- strated that they became functional only when E was bound to S in the ES complex. Conformational changes very likely occurred that couple the function of these sites with that of the active site. The secondary site(s) might be involved in substrate hydrolysis and/or product release. This function was then clearly distinct from the barley a-amylase binding onto starch granules, which most probably occurred prior to hydrolysis of the substrate glycosidic bond. Acknowledgements We thank Drs E. H. Ajandouz and R. Koukiekolo for stimulating discussion, C. Villard for advice and excellent technical assistance, B. Dwisusilo for his help in the preparation of the illustration, and S. Ehlers for enzyme purification. References 1. Davies, G.J. & Henrissat, B. (1995) Structure and mechanism of glycosyl hydrolases. Structure 3, 853–859. 2. Thoma, J., Rao, G., Brothers, C. & Spradlin, J. (1971) Subsite mapping of enzymes. Correlation of products patterns with Michaelis parameters and substrate induced strain. J. Biol. Chem. 246, 6521–5635. 3. Fogarty, W.M. (1983) Microbial, Enzymes and Biotechnology (Fogarty, W.M., ed), 1–19. Applied Science Publishers, London. 4. Gottschalk, T.E., Fierobe, H.P., Mirgorodskaya, E., Clarke, A.J., Tull, D., Sigurskjold, B.W., Christensen, T., Payre, N., Frandsen, T.P.,Juge,N.,McGuire,K.A.,Cottaz,S.,Roepstorff,P.,Driguez, H., Williamson, G. & Svensson, B. (1998) Structure, function and protein engineering of starch-degrading enzymes. Biochem. Soc. Trans. 26, 198–204. 5. Rogers, J.C. & Milliman, C. (1983) Isolation and sequence analysis of a barley a-amylase cDNA clone. J. Biol. Chem. 258, 8169–8174. 6. Rogers, J.C. (1985) Two barley a-amylase gene families are regulated differently in aleurone cells. J. Biol. Chem. 260, 3731– 3738. 7. Søgaard, M., Kadziola, A., Haser, R. & Svensson, B. (1993) Site- directed mutagesis of histidine 93, aspartic acid 180, glutamic acid 205, histidine 290 and aspartic acid 291 at the active site and tryptophan 279 at the raw starch binding site in barley a-amylase 1. J. Biol. Chem. 268, 22480–22484. 8. Rodenburg, K.W., Juge, N., Guo, X., Søgaard, M., Chaix, J.C. & Svensson, B. (1994) Domain B protruding at the third b strand of the a/b barrel in barley a-amylase confers distinct isozyme-specific properties. Eur. J. Biochem. 221, 277–284. 9. Ajandouz, E.H., Abe, J., Svensson, B. & Marchis-Mouren, G. (1992) Barley malt a-amylase. Purification, action pattern and subsite mapping of isozyme 1 and two members of isoenzyme 2 sub-family using p-nitrophenylated maltooligosaccharide sub- strates. Biochim. Biophys. Acta 1159, 193–202. 10. Gye ´ mant,G.,Hovanszki,G.&Kandra,L.(2002)Subsitemap- ping of the binding region of a-amylase with computer program. Eur. J. Biochem. 269, 5157–5162. 11. Nahoum, V., Roux, G., Anton, V., Rouge ´ , P., Puigserver, A., Bischoff, H., Henrissat, B. & Payan, F. (2000) Crystal structures of human pancreatic a-amylase in complex with carbohydrate and proteinaceous inhibitors. Biochem. J. 346, 201–208. 12. Ajandouz. El, H. & Marchis-Mouren, G. (1995) Subsite mapping of porcine pancreatic a-amylase I and II using 4-nitrophenyl- a-maltooligosaccharides. Carbohydr. Res. 268, 267–277. 13. Gibson, R.M. & Svensson, B. (1987) Identification of tryptopha- nyl residues involved in binding of carbohydrate ligands to barley a-amylase 2. Carlsberg Res. Commun. 52, 373–379. 14. Kadziola, A., Søgaard, M., Svensson, B. & Haser, R. (1998) Molecular structure of a barley a-amylase inhibitor complex: implications for starch binding and catalysis. J. Mol. Biol. 278, 205–217. 15. Weselake, R.J. & Hill, R.D. (1982) Cycloheptaamylose as an affinity ligand of cereal a-amylases. Characteristics and possible mechanism of interaction. Carbohydr. Res. 108, 153–161. 16. Weselake, R.J., MacGregor, A.W. & Hill, R.D. (1983) Purifica- tion and characteristics of an endogenous a-amylase inhibitor from barley kernels. Plant. Physiol. 73, 1008–1012. 17. Brzozowski, A.M. & Davies, G.J. (1997) Structure of the Asper- gillus oryzae a-amylase complexed with the inhibitor acarbose at 2.0 A ˚ resolution. Biochemistry 36, 10837–10845. 18. Fujimoto,Z.,Takase,K.,Doui,N.,Momma,M.,Matsumoto,T. & Mizuno, H. (1998) Crystal structure of a catalytic-site mutant a-amylase from Bacillus subtilis complexed with maltopentaose. J. Mol. Biol. 277, 393–407. 19. Qian, M., Haser, R. & Payan, F. (1993) Structure and molecular model refinement of pig pancreatic a-amylase at 2.1 A ˚ resolution. J. Mol. Biol. 213, 785–799. 20. Brayer,G.D.,Sidhu,G.,Maurus,R.,Rydberg,E.H.,Braun,C., Wang, Y., Nguyen, N.T., Overall, C.M. & Withers, S.G. (2000) Subsite mapping of the human pancreatic a-amylase active site through structural, kinetic, and mutagenesis techniques. Bio- chemistry 39, 4778–4791. 21. Boel,E.,Brady,L.,Brzozowski,M.,Derewenda,Z.,Dodson, G.G., Jensen, V.J., Peteren, S.B., Swift, H., Thim, L. & Woldike, H.F. (1990) Calcium binding in a-amylase: an X-ray diffraction study at 2.1 A ˚ resolution of two enzymes from Aspergillus. Bio- chemistry 29, 6244–6249. 22. Brzozowski, A.M., Lawson, D.M., Turkenburg, J.P., Bisga ˚ rd- Frantzen, H., Svendsen, A., Borchert, T.V., Dauter, Z., Wilson, K.S. & Davies, G.J. (2000) Structural analysis of a chimeric bac- terial a-amylase. High-resolution analysis of native and ligand complexes. Biochemistry 39, 9099–9107. 23. Kadziola, A., Abe, J., Svensson, B. & Haser, R. (1994) Crystal and molecular structure of barley a-amylase. J. Mol. Biol. 239, 104–121. 24. Al Kazaz, M., Desseaux, V., Marchis-Mouren, G., Payan, F., Forest, E. & Santimone, M. (1996) The mechanism of porcine 3878 N. Oudjeriouat et al. (Eur. J. Biochem. 270) Ó FEBS 2003 pancreatic a-amylase. Kinetic evidence for two additional carbo- hydrate-binding sites. Eur. J. Biochem. 241, 787–796. 25. Al Kazaz, M., Desseaux, V., Marchis-Mouren, G., Prodanov, E. & Santimone, M. (1998) The mechanism of porcine pancreatic a-amylase. Inhibition of maltopentaose hydrolysis by acarbose, maltose and maltotriose. Eur. J. Biochem. 252, 100–107. 26. Koukiekolo, R., Le Berre-Anton, V., Desseaux, V., Moreau, Y., Rouge, P., Marchis-Mouren, G. & Santimone, M. (1999) Mechanism of porcine pancreatic a-amylase. Inhibition of amy- lose and maltopentaose hydrolysis by kidney bean (Phaseolus vulgaris) inhibitor and comparison with that by acarbose. Eur. J. Biochem. 265, 20–26. 27. Koukiekolo, R., Desseaux, V., Moreau, Y., Marchis-Mouren, G. & Santimone, M. (2001) Mechanism of porcine pancreatic a-amylase. Inhibition of amylose and maltopentaose hydrolysis by a-, b-andc-cyclodextrin.Eur. J. Biochem. 268, 841–848. 28. Desseaux, V., Koukiekolo, R., Moreau, Y., Santimone, M. & Marchis-Mouren, G. (2002) Mechanism of porcine pancreatic a-amylase (PPA). Inhibition of amylose and maltopen- taose hydrolysis by various inhibitors. Biologia, Bratislava 57, 163–170. 29. Ferey-Roux, G., Perrier, J., Forest, E., Marchis-Mouren, G., Puigserver, A. & Santimone, M. (1998) The human pancreatic a-amylase isoforms: isolation, structural studies and kinetics of inhibition by acarbose. Biochem. Biophys. Acta 1388, 10–20. 30. Moreau, Y., Desseaux, V., Koukiekolo, R., Marchis-Mouren, G. & Santimone, M. (2001) Starch digestion in tropical fishes. isolation, structural studies and inhibition kinetics of a-amylases from two tilapias Oreochromis niloticus and Sarother- odon melanotheron. Comparative Biochem. Physiol. Part B 128, 543–552. 31.Talamond,P.,Desseaux,V.,Moreau,Y.,Santimone,M.& Marchis-Mouren, G. (2002) Isolation, characterization and inhibition by acarbose of the a-amylase from Lactobacillus fermentum: comparison with Lb. manihotivorans and Lb. plan- tarum amylases. Comparative Biochem. Physiol. Part B 133, 351–360. 32. Robyt, J.F. & French, D. (1970a) Multiple attack and polarity of action of porcine pancreatic a-amylase. Arch. Biochem. Biophys. 138, 662–670. 33. Svensson, B., Mundy, F., Gibson, R.M. & Svendsen, I. (1985) Partial amino acid sequences of a-amylase isozymes from barley malt. Carlsberg Res. Commun. 50, 15–22. 34. Chen, Y., Fringant, C. & Rinaudo, M. (1997) Molecular char- acterization of starch by SEC: dependence of the performances on the amylopectin content. Carbohydr. Polymers 33, 73–78. 35. Schou, C., Rasmussen, G., Kaltoft, M.B., Henrissat, B. & Schu ¨ - lein, M. (1993) Stereochemistry, specificity and kinetics of the hydrolysis of reduced cellodextrins by nine cellulases. Eur. J. Biochem. 217, 947–953. 36. Dygert, S., Li, L.H., Don Florida, R. & Thoma, J.A. (1965) Determination of reducing sugar with improved precision. Anal. Chem. 13, 367–374. 37. Sas, Institute Inc. (1989) SAS/STATÒ. User’s Guides, Version 6 4th edn, Vol. 1, 943. Sas Institute Inc, Cary, NC, USA. 38. Ohnishi, M. (1971) Studies of interaction of the substrate ana- logues with bacterial liquefying a-amylase by means of spectro- photometry and steady state kinetics. J. Biochem. 69, 181–189. 39. Svensson, B. & Sierks, M.R. (1992) Roles of aromatic side chains in binding of substrates, inhibitors and cyclodextrins to gluco- amylase from Aspergillus niger probed by perturbation difference spectroscopy, chemical modification and mutagenesis. Carbohydr. Res. 228, 29–44. 40. MacGregor, E.A. & MacGregor, A.W. (1985) The action of cereal a-amylase on solubilized starch and cereal starch granules. In New Approaches to Research on Cereal Carbohydrates. (Hill, R.D. & Munch, L., eds), pp. 149–160, Elsevier, Amsterdam, the Nether- lands. Ó FEBS 2003 Acarbose inhibition of barley a-amylases (Eur. J. Biochem. 270) 3879 . On the mechanism of a -amylase Acarbose and cyclodextrin inhibition of barley amylase isozymes Naı¨ma Oudjeriouat 1 , Yann Moreau 2 , Marius Santimone 1 ,. Thus the longer the substrate, the higher was the activity. Inhibition by acarbose Inhibition of amylose hydrolysis occurred in the presence of 10–80 l M acarbose