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
´

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,

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:
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

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

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

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

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

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

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

,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
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¢

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

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
,

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
)

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

Â
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

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

). 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

)
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

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

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

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

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
¨