Barley a-amylase Met53 situated at the high-affinity subsite )2
belongs to a substrate binding motif in the bfia loop 2
of the catalytic (b/a)
8
-barrel and is critical for activity
and substrate specificity
Haruhide Mori*, Kristian Sass Bak-Jensen and Birte Svensson
Carlsberg Laboratory, Department of Chemistry, Gamle Carlsberg Vej 10, Copenhagen Valby, Denmark
Met53 in barley a-amylase 1 (AMY1) is situated at the high-
affinity subsite )2. While Met53 is unique to plant a-amy-
lases, the adjacent Tyr52 stacks onto substrate at subsite )1
and is essentially invariant in glycoside hydrolase family 13.
These residues belong to a short sequence motif in bfia loop
2 of the catalytic (b/a)
8
-barrel and site-directed mutagenesis
was used to introduce a representative variety of structural
changes, Met53Glu/Ala/Ser/Gly/Asp/Tyr/Trp, to investi-
gate the role of Met53. Compared to wild-type, Met53Glu/
Asp AMY1 displayed 117/90% activity towards insoluble
Blue Starch, and Met53Ala/Ser/Gly 76/58/38%, but
Met53Tyr/Trp only 0.9/0.1%, even though both Asp
and Trp occur frequently at this position in family 13.
Towards amylose DP17 (degree of polymerization ¼ 17) and
2-chloro-4-nitrophenyl b-
D
-maltoheptaoside the activity
(k
cat
/K
m

and PNPG
and PNPG
2
in equal amounts from PNPG
5.
Met53Trp
affected the action pattern on PNPG
7
, which was highly
unusual for AMY1 subsite mutants. It was also the sole
mutant to catalyze substantial transglycosylation – promo-
ted probably by slow substrate hydrolysis – to produce up to
maltoundecaose from PNPG
6
.
Keywords: glycoside hydrolase family 13; plant a-amylases;
site-directed mutagenesis; binding subsite engineering;
oligosaccharide hydrolysis.
a-Amylases (a-1,4-
D
-glucan glucanohydrolase, EC 3.2.1.1)
catalyze hydrolysis of internal a-1,4-glucosidic linkages in
starch and related oligosaccharides and polysaccharides [1]
and belong to glycoside hydrolase family 13 (GH13) [2–5].
Family 13 and the closely related families 70 and 77
constitute glycoside hydrolase clan H (GH-H) [5] that
currently comprises 28 different enzyme specificities [2–5],
e.g. a-glucosidase (EC 3.2.1.20), maltotetraose-forming exo-
amylase (EC 3.2.1.60), cyclomaltodextrinase (EC 3.2.1.54),
isoamylase (EC 3.2.1.68), pullulanase (EC 3.2.1.41), oligo-

particular subsites of high or low affinity along the cleft
Correspondence to B. Svensson, Carlsberg Laboratory, Department of
Chemistry, Gamle Carlsberg Vej 10, DK-2500 Copenhagen Valby,
Denmark. Fax: + 45 33 27 47 08, Tel.: + 45 33 27 53 45,
E-mail: [email protected]
Abbreviations:AMY1,barleya-amylase 1; AMY2, barley a-amylase
2; Cl-PNPG
7
, 2-chloro-4-nitrophenyl b-
D
-maltoheptaoside; DP17,
degree of polymerization ¼ 17; GH13, glycoside hydrolase family 13;
GH-H, clan H of glycoside hydrolases; PNPG, 4-nitrophenyl
a-
D
-glucoside; PNPG
2
–PNPG
12
, 4-nitrophenyl a-
D
-maltoside
through 4-nitrophenyl a-
D
-maltododecaoside; TAA, Taka-amylase A.
Enzyme: a-Amylase (a-1,4-
D
-glucan glucanohydrolase, EC 3.2.1.1).
*Present address: Division of Applied Bioscience, Graduate School of
Agriculture, Hokkaido University, Sapporo 060-8589, Japan

onto the sugar ring at subsite )6 [44–46] and subsites
interacting with the leaving or aglycon part of substrates
to include sequence motifs at C-terminal extensions of
b-strands 4 and 5 as well as residues from the long b7fia7
segment of the (b/a)
8
-barrel [16,36–38,45,46].
Alignment of bfia loop 2 sequences that contributed to
subsites )1and) 2ina-amylases and other GH-H
members, identified a motif with an invariant Tyr stacking
onto inhibitor and substrate at subsite )1 [7,13,14,16–27].
This Tyr is succeeded by Trp in, e.g. CGTases, Taka-
amylase A (TAA), and maltogenic amylase [7,8,17,
20,21,23–25], and by Gln in porcine and human pancreatic
and B. subtilis a-amylases [14,18,19,22,24,27]. Glucosyl O6
at subsite )2 in enzyme/oligosaccharide complexes was
hydrogen bonded to NE1 of the indole ring of Trp
[7,17,20,21,23,25] or to NE2 of the carboxamide group in
Gln [18,19,22,24]. Asp was also common [9,13], while Phe
[10], Met [16], and Ala [47,48] were rarely seen among
known structures. Sequence alignment moreover identified
sporadic occurrence of Leu, Gly, Tyr, and His at the
position in question (see Table 1 below).
In barley a-amylase the invariant Tyr51 stacked onto the
valienamine ring (a sugar mimic) of acarbose bound at
subsite )1 [16]. Acarbose, however, only covered subsites )1
through +2 in this complex [16], but in a modeled AMY2/
maltodecaose complex, Met52 SD was 3.4 A
˚
from O6 of the

500, and > 40-fold variation in activity towards insoluble
Blue Starch, amylose DP17, and 2-chloro-4-nitrophenyl
b-
D
-maltoheptaoside, respectively.
MATERIALS AND METHODS
Materials
Escherichia coli DH5a and JM109 [53] (Life Technologies,
Inc., MD, USA) were used for propagation of the
expression plasmid derived from pPICZA (Invitrogen,
Carlsbad, CA, USA) carrying the Zeocin
R
marker gene
for selection of E. coli and Pichia pastoris transformants.
P. pastoris GS115 [54] (Invitrogen) was used for expression
of AMY1 cDNA inserted into pPICZA under the control of
the AOX1 promoter [55]. Standard culture media were used
for E. coli [56] and P. pastoris [50].
Construction of expression plasmids, transformation,
and screening
Derivatives were constructed of the expression plasmid
pPICZA harboring inserts encoding AMY1 flanked by
EcoRI and KpnI sites. For AMY1 wild-type, cDNA was
amplified using primers A; 5¢-TTT
GAATTCCATG
GGG AAG AAC GGC AGC-3¢ (pos. 87–114, sense
orientation), and B; 5¢-TTT
GGT ACC TCA GTT CTT
CTC CCA GAC GGC GTA-3¢ (pos. 1395–1363, antisense
orientation), to generate DNA with the EcoRI and KpnI

Plant Wheat (AMY3) 71–86 VS PEGYLPGQLYNLNS P08117
Plant Black gram 68–83 VS PEGYLPGRLYDLDA P17859
Mammal Hog 50–72
VVTN PSRPWWERYQPVSYKLCTR P00690
Bacterium Bacillus subtilis 90–113 KEGNQGDKSMS NWYWLYQPTSYQIGNR P00691
Mold Aspergillus oryzae 69–92 LPQTTAYG DAYHGYWQQDIYSLNE P10529
Yeast Saccharomycopsis fibuligera 96–119
IPDNTAYG YAYHGYWMKNIYKINE P21567
Bacterium Bacillus stearothermophilus 111–135 LDTLAGTDN TGYHGYWTRDFKQIEE P19531
Bacterium Bacillus acidopullulyticus 202–217
HS NHKYDTIDYMEIDP P32818
Bacterium Escherichia coli 48–66
ASGG YSVGYDSYDLFDLGE P26612
Bacterium Bacillus amyloliquefaciens 78–95
LSQS DNGYGPYDLYDLGE P00692
Bacterium Bacillus licheniformis 78–95 TSQA DVGYGAYDLYDLGE P06278
Bacterium Paenibacillus polymyxa 812–830 KSE YAYHGYHTYDFYAVDG P21543
Bacterium Escherichia coli 255–285 IHGWVGGGTKGDFPHYAYHGYYTQDWTNLDA P25718
G4-forming amylase Bacterium Pseudomonas saccharophila 84–109
FSSWTDGGKS GGGEGYFWHDFNKNGR P22963
G6-forming amylase Bacterium Bacillus sp. (strain 707) 84–101 ASQN DVGYGAYDLYDLGE P19571
G5-forming amylase Bacterium Pseudomonas sp. 68–96 EHNWVSSGDGAP YPWWMRYQPVSYSLDRS Q52516
Branching enzyme Bacterium Bacillus stearothermophilus 195–214 EHPFD RSWGYQGIGYYSATS P30538
Amylopullulanase Bacterium Thermoanaerobacter ethanolicus 477–494
QSPS NHRYDTTDYTKIDE P38939
Neopullulanase Bacterium Bacillus stearothermophilus 200–217 RSPS NHKYDTADYFEVDP P38940
CDase
a
Bacterium Thermoanaerobacter ethanolicus 196–213 LSHS THKYDTTDYYTIDP P29964
Tre 6-P hydrolase

DTG SCSSPYNSISSIALNP O84089
Bacterium Synechocystis sp. 51–69
PTG FGNSPYLCYSALAINP P72785
Plant Potato 124–146
PPGKR GNEDGSPYSGQDANCGNT Q06801
Plant Arabidopsis thaliana 128–146
PP DEGGSPYAGQDANCGNT Q9LV91
a
Cyclodextrinase.
b
Trehalose 6-phosphate hydrolase.
c
Cyclodextrin glycosyltransferase.
d
Enzymes from GH70.
e
Enzymes from GH77.
Ó FEBS 2002 Met53 mutants at subsite )2 in barley a-amylase 1 (Eur. J. Biochem. 269) 5379
precipitation [56] or a GFX plasmid purification column
(Pharmacia, Sweden). The entire sequence was subsequently
confirmed (Applied Biosystems 377 DNA Sequencer and
Taq DyeDeoxy Terminator Cycle Sequencing kit, Perkin-
Elmer) and the plasmid was used for P. pastoris transfor-
mation by electroporation [50] upon linearization at the
BstXI site. Screening was performed for Zeocin transform-
ants on YPDS plates (1% yeast extract, 2% peptone, 2%
glucose, 1
M
sorbitol, and 2% agar) containing 100 lgÆmL
)1

KTAexplorer automated chromatograph (Pharmacia)
[38]. The sample was applied to the column, equilibrated
in 10 m
M
sodium acetate buffer, 1 m
M
CaCl
2
, pH 7.0, and
eluted, at a flow rate of 3 mLÆmin
)1
, by a gradient (0–50%/
48 mL, 50–100%/24 mL) made from equilibration buffer
and 10 m
M
sodium acetate buffer, 1 m
M
CaCl
2
, 5m
M
NaCl, pH 5.3. The first eluted protein peak was collected,
dialyzed against 10 m
M
Mes, 25 m
M
CaCl
2
, pH 6.8,
concentrated (Centriprep YM10 or YM30, Millipore,

CaCl
2
, 0.5 mgÆmL
)1
BSA, pH 5.5. The reaction was
initiated by enzyme addition (around 1 U) to the suspension
(4 mL) and stopped after 15 min at 37 °Cby0.5
M
NaOH (1 mL). After centrifugation (2 min, 12 000 g)
supernatants were transferred (300 lL) to a microtiter
plate. A
620
values (Ceres UV900 HDI microplate reader,
Biotek Instruments, Inc., UK) in the range 0.8–1.2 were
used to calculate activity [36]. One unit was defined as the
amount of enzyme that during 15 min reaction resulted in
an increase in A
620
of 1 in the supernatant of the stopped
reaction mixture.
Amylose. Rates of hydrolysis of amylose DP17 (average
degree of polymerization 17, Hayashibara Chemical
Laboratories, Okayama, Japan) were determined in 20 m
M
sodium acetate buffer, 5 m
M
CaCl
2
, and 0.05 mgÆmL
)1

7
(GranutestÒ 3, Merck,
Darmstadt, Germany) at 30 °C was measured as described
[36] with 20.0–103 n
M
wild-type and Met53Ala/Gly/Asp/
Glu/Ser/Tyr AMY1 and k
cat
and K
m
were determined as
above using five to eight substrate concentrations (0.40–
8.0 m
M
).
Bond cleavage frequencies of 4-nitrophenyl a-
D
-malto-
oligosaccharides. Individual bond cleavage frequencies
were analysed for PNPG
7
(Boehringer Mannheim, Ger-
many), PNPG
6
, and PNPG
5
(both Calbiochem, Bad Soden,
Germany) in 20 m
M
sodium acetate buffer pH 5.5, 5 m

1-7
and 4-nitrophenol
were detected at 313 nm (Shimadzu SPD-10AU UV-VIS
detector) and quantified by using standard mixtures. The
bond cleavage frequencies were calculated for products
obtained at 4–17% substrate consumption.
Transglycosylation. In transglycosylation experiments the
same conditions as for hydrolysis were applied using 10 m
M
of PNPG
6
.
Molecular graphics
The structures of AMY2/acarbose and TAA/acarbose were
obtained from the protein data bank [59], entry codes 1BG9
and 7TAA, respectively. The figures were made using the
software program
INSIGHT II
(98.0) (Molecular Simulations
Inc., San Diego, CA).
5380 H. Mori et al. (Eur. J. Biochem. 269) Ó FEBS 2002
RESULTS
Choice of mutants
a-Amylases and other GH-H enzymes have a short
substrate glycon binding motif in the middle of the typically
35 residues bfia loop 2 in the (b/a)
8
-barrel (Table 1). In
barley AMY1 (for which the structure is not available) and
AMY2 this loop (Pro41-Gly65, AMY1 numbering)

TAA loop 2 in TAA appeared to hinder binding of the
substrate beyond subsite )3/)4 as illustrated by global
views of AMY2 and TAA complexes (Fig. 1B). The
enzymatically determined subsite maps agreed with different
length of the glycon binding region in AMY2 and TAA
[44,61]. Comparison of structures of AMY2 and other
a-amylases (not shown) also gave the impression that
AMY2 might accommodate larger parts of the substrate.
Thus porcine pancreatic a-amylase like TAA had a larger
loop 2 segment with the location of Gln63 resembling that
of Trp83
TAA
[18]. This holds true also for Trp101 in
CGTase [25]. Although almost one thousand GH-H
sequences were reported [5] Met53 occurred only in plant
a-amylases (Table 1), with the exception of a bacterial
isoamylase, which had a structural unit formed by bfia
segments 3 and 4 [11] but lacked domain B which together
with bfia loop 2 created the glycon binding site [3].
Subsite )2 had highest affinity of the 10 subsites in
AMY1 and AMY2 [44]. Subsites )6 and +1 were almost as
strong, while )5, )4, +2, and +4 contributed intermediate,
and )3 and + 3 very weak binding energy. The catalytic
subsite )1 had a large negative affinity [44] due to energy
spent to distort of the bound glucose residue in catalysis. SD
of Met52
AMY2
(Met53
AMY1
) in a computed AMY2/malto-

FVG and FGG variants which, although no reported
natural sequences contained Gly, were unusual [36–38,
61,62]. These mutants moreover were highly unusual among
the already described AMY1 subsite mutants [34–36,59,60]
by having improved activity towards Cl-PNPG
7
and less
than 10% activity for insoluble starch [37].
Inspection of Met53
AMY1
replacements in AMY2/
acarbose [16] indicated Asp, Glu, Ser, Asn, and Ala, as
readily accommodated, whereas Trp53
AMY1
and perhaps
Tyr53
AMY1
might obstruct the binding cleft.
Production and purification of AMY1 mutants
The P. pastoris transformants secreted the mutants at 14
(Met53Glu), 22 (Met53Ala), 16 (Met53Ser), 3.9
(Met53Gly), 1.9 (Met53Asp), 6.0 (Met53Tyr), and 20
(Met53Trp) mgÆL
)1
as calculated from the activity in the
culture supernatants towards insoluble Blue Starch and the
specific activity of the purified enzymes. All mutants gave as
a single band in SDS/PAGE after purification on b-cyclo-
dextrin-Sepharose, but resolved into two components of pI
4.8 and 4.7 in IEF (Fig. 2). The form of pH 4.8 eluting first

Met53TrpandMet53Tyrcomparedtowild-typeAMY1
were surprisingly poor catalysts showing 0.1 and 0.9%
activity, respectively, towards insoluble Blue Starch.
Moreover, the catalytic efficiency (k
cat
/K
m
) of these mutants
was reduced 10
3
-to10
4
-fold for both amylose DP17 and
Cl-PNPG
7
(Table 2). This was chiefly due a low k
cat
for
Ó FEBS 2002 Met53 mutants at subsite )2 in barley a-amylase 1 (Eur. J. Biochem. 269) 5381
5382 H. Mori et al. (Eur. J. Biochem. 269) Ó FEBS 2002
amylose DP17 of only 0.1 and 0.6% of the wild-type value
for Met53Trp and Met53Tyr, respectively, while the K
m
increased about 10-fold as for other Met53 mutants. For
Table 2. Activity and kinetic parameters of Met53 AMY1 mutants and wild-type towards insoluble Blue Starch, amylose DP17, and Cl-PNPG
7
. BS, Blue Starch; Amyl, amylose DP17. ND, not determined,
(K
m
too high).

)1
)
k
cat
(s
)1
)
K
m
(m
M
)
K
m
(s
)1
Æm
M
)1
)
Activ./k
cat
(UÆmg
)1
Æs)
Activ./(k
cat
/K
m
)

)1
.
Fig. 2. Isoelectric focusing in the pH range 4.0–6.5 of AMY1 wild-type
and mutants (60 ng each) produced in P. pastoris and purified on
b-cyclodextrin-Sepharose. (A) Protein silver staining. (B) Activity
staining. Lane 1: pI marker proteins; lane 2: wild-type AMY1; lanes
3–8: Met53Trp, Met53Tyr, Met53Asp, Met53Gly, Met53Ala, and
Met53Ser AMY1.
Fig. 1. Comparison of the structure of complexes of inhibitory substrate
analogues derived from acarbose and barley a-amylase 2 (AMY2 [16]);
and Taka-amylase A (TAA [17]). (A) Stereo view of interactions
involving segments of bfia loops 2 and 3 (i.e. domain B) from AMY2
(in green) and TAA (in black). The superimpositioning was guided by
the catalytic acids (D179
AMY2
, E204
AMY2
, and D289
AMY2
and
D206
TAA
, E230
TAA
, and D297
TAA
). The invariant Y51
AMY2
and
Y82

are in red
(indicated by arrow). Y51
AMY2
and Y82
TAA
are in orange. Other
binding residues (W9
AMY2
, H92
AMY2
, T94
AMY2
, A95
AMY2
,
Y130
AMY2
, A145
AMY2
, F180
AMY2
, K182
AMY2
, W206
AMY2
,
S208
AMY2
, Y211
AMY2

and K
m
could not be determined
due to low affinity and while the second order rate constant
(k
cat
/K
m
) of Met53Tyr was 0.025% of wild-type, it could not
be estimated for Met53Trp AMY1 as it had low activity
(Table 2).
The two other groups of Met53 mutants had consid-
erably reduced catalytic efficiency on amylose DP17 and
Cl-PNPG
7
, k
cat
/K
m
corresponding to 1.8–5.5% and
0.3–1.7%, respectively, of the wild-type values. On amy-
lose DP17 Met53Glu was the most active mutant with a
k
cat
of 84%, while K
m
increased 15 times compared to
wild-type. Met53Asp/Ser/Gly had a k
cat
of 38–45% and

for
Met53Glu/Ala of k
cat
/K
m
superior to Met53Asp AMY1
(Table 2). k
cat
of Met53Glu/Ala AMY1 was thus assessed
to be ‡ 30 s
)1
.
Remarkably, the Met53 mutants, except for Met53Trp/
Tyr, showed good activity towards insoluble Blue Starch of
38–117% compared to wild-type. The five most active
mutants also gave similar ratios of activity towards insol-
uble Blue Starch over k
cat
/K
m
for amylose DP17 in the
range of 100–140, while the ratios were around 60 for
Met53Trp/Tyr and 6 for wild-type. The noted expansion to
fourfold variation of the ratio of the activity towards starch
over k
cat
for amylose (10–43; Table 2) suggested that in
certain mutants reduced affinity for insoluble Blue Starch
accompanied the low affinity for amylose DP17. This
property, however, was not further investigated.

saccharides, as confirmed by quantitative analysis of
hydrolysis products from PNPG
7
, PNPG
6
, and PNPG
5
(Table 3).
Six Met53 mutants and wild-type AMY1 primarily
hydrolyzed the second glucosidic bond in PNPG
7
to release
PNPG and G
6
, but Met53Trp also released PNPG
2
and G
5
to constitute 30% of the products, PNPG and G
6
being
formed in 50%, and PNPG
5
and G
2
in 17% of its cleavages
(Table 3). Thus even subsites +4/+5 may be occupied in
productive Met53Trp–PNPG
7
complexes. The action pat-

for PNPG
7
as
indicated by the effect on Cl-PNPG
7
(Table 2) for which
kinetic parameters were not determined due to the high K
m
and/or low k
cat
or for both reasons. The rate of product
release was 8% of that of wild-type for the most active
mutants, Met53Glu/Ala, and 2% for the second most
active group, Met53Ser/Gly/Asp, whereas very low values
of 0.1% and 0.006% for Met53Tyr and Met53Trp,
respectively, presumably stemmed from a dominating loss
of rate of catalysis as suggested by the kinetics properties of
these mutants on amylose DP17 and Cl-PNPG
7
(Table 3).
Most remarkably binding of PNPG
6
at subsites )6
through +1 to release 4-nitrophenol was favored only by
wild-type AMY1. This mode reflected the high affinity in
AMY1 at subsite )6(7.68kJÆmol
)1
) compared to +2
(4.94 kJÆmol
)1

+5 and +6, i.e. exterior to the kinetically determined wild-
type binding cleft. Wild-type AMY1 in contrast released 1%
PNPG
4
and no PNPG
5
(Table 3). Met53Trp differed by
releasing as little as 1% 4-nitrophenol from PNPG
6
,
compared to 9–19% formed by the other mutants. The
two structurally similar mutants Met53Ala and Met53Gly
AMY1 showed closely related action patterns, which also
resembled that of Met53Asp, while Met53Glu/Ser/Tyr
shared a different trend in their action pattern (Table 3).
Remarkably, the Met53 mutants hydrolysed PNPG
5
and PNPG
6
at essentially the same rate, whereas wild-type
5384 H. Mori et al. (Eur. J. Biochem. 269) Ó FEBS 2002
AMY1 hydrolysed PNPG
6
about five times faster than
PNPG
5
. The mutant activity, however, still grouped for
PNPG
5
as for PNPG

3
surpassing that of PNPG
2
.
With PNPG
5
and PNPG
6
, all Met53 mutants thus
apparently disfavoured substrate glycon binding interac-
tions compared to wild-type AMY1 while this in PNPG
7
hydrolysis appeared only for Met53Trp AMY1.
Transglycosylation by Met53Ala/Tyr/Trp
Retaining glycoside hydrolases are able to catalyze
transglycosylation [3] as depicted in the schematics of
the reaction mechanism (Fig. 3). Under the present assay
Table 3. Action pattern for hydrolysis of PNPG
7
,PNPG
6
, and PNPG
5
by Met53 AMY1 mutant and wild-type. [PNPG
5-7
] ¼ 1.0 m
M
.
AMY1 Cleavage frequency (%) [E] (n
M

5
G–G–G–G–G–PNP
Wild-type 14 44 41 1 167 3.0 9.6 100 (0.019)
M53E 32 53 15 167 7.0 7.3 32.9
M53A 40 44 16 167 16 11.0 21.7
M53S 1 48 35 16 833 10 15.3 9.7
M53G 40 44 16 167 60 7.4 3.9
M53D 39 46 15 167 60 8.6 4.5
M53Y 2 32 54 11 1 833 90 7.4 0.52
M53W
a
3 43 42 11 1 5000 17 12.9 0.80
a
Transglycosylation was apparent by the formation of PNP malto-oligosaccharides longer than PNPG
5
;
b
the activity relative to wild-type
AMY1 as estimated from the substrate consumption, reaction time, and enzyme concentration. The wild-type AMY1 values given in
parenthesis are calculated as [product]/[enzyme] per minute corresponding to the entire period of incubation.
Fig. 3. Schematics of the double displacement mechanism of retaining
glycoside hydrolases [3,66]. In transglycosylation the covalent inter-
mediate is attacked at C1 by another sugar molecular, HO-R2, which
in hydrolysis would be replaced by water. R and R3 signify other
substrate chain parts.
Ó FEBS 2002 Met53 mutants at subsite )2 in barley a-amylase 1 (Eur. J. Biochem. 269) 5385
conditions, Met53Trp AMY1 formed 1.8 l
M
PNPG
7

(Figs 4 and 5; Table 4). In the case of Met53Trp
this amounted to 15% of the total products, compared to
£ 3% and £ 1% for Met53Tyr and Met53Ala AMY1,
respectively. It is noted that action patterns at 1 m
M
and
10 m
M
PNPG
6
were not completely identical (Tables 3
and 4; Fig. 4), for example 5–9% PNPG
5
was formed from
10 m
M
PNPG
6
(Table 4) but lacking in 1 m
M
PNPG
6
reaction mixtures (Table 3).
The degree of polymerization of the various transglyco-
sylation products from Met53Trp could not be confirmed
as proper reference compounds are not available. How-
ever, from the number of peaks in the HPLC chromato-
gram (Fig. 5B), Met53Trp presumably gave PNPG
7-11
,

(Table 3). The longest transglycosylation product from a
single catalytic event was PNPG
12
, which can only be
present in trace amounts (Fig. 5). The anticipated domin-
ant product was PNPG
10
generated by nucleophilic attack
of PNPG
6
as acceptor on the enzyme maltotetraose-
intermediate (Fig. 3), which arose by release of the major
product PNPG
2
(Fig. 4). Although PNPG
10
, however,
appeared in higher amounts than the products in neigh-
bouring peaks in the chromatogram (Fig. 5), the shorter
PNPG
8
predominated. Thus significant hydrolysis of the
longer products took place. Although monitoring of the
4-nitrophenyl chromophor fails to detect both substrate
glycon moieties after hydrolysis and – where such products
acted as acceptors – the transglycosylation products,
underivatized maltodextrins were assumed to arise in trace
amounts only.
Table 4. Product distribution from 10 m
M

6
catalyzed by AMY1
Met53Ala (A), Met53Trp (B), and Met53Tyr
(C) mutants. *, include PNPG
7-11
; j, PNPG
5
;
h, PNPG
4
; n, PNPG
3
; m, PNPG
2
; s,
PNPG; d, PNP. Enzyme concentrations are
given in Table 4.
Fig. 5. HPLC profiles of the reaction products from 10 m
M
PNPG
6
catalyzed by AMY1 Met53Ala (A), Met53Trp (B), and Met53Tyr (C)
mutants, and substrate before the reaction (D). Enzyme concentrations
and reaction times are given in Table 4. The arrows indicate presumed
PNPG
8-11
.
5386 H. Mori et al. (Eur. J. Biochem. 269) Ó FEBS 2002
DISCUSSION
Role of the Met53 region in AMY1 and GH-H

changes in barley compared to fungal a-amylase may stem
from the shorter and perhaps less adaptable bfia loop 2 in
the plant enzyme and a requirement of structural integrity at
a longer glycon binding crevice.
Noticeably, sequence variation is sparse in eukaryotes at
the position corresponding to Met53
AMY1
. Thus plants
have Met and occasionally Leu, animals Gln, yeast and
fungi Trp, whereas bacterial a-amylases have Gln, Trp,
Gly, Asp, His, or Tyr, and do not include the plant type.
Finally, in non-a-amylase GH-H members Phe, Gly, Asp,
Met, Trp, Ala, Gln, and Ser occur (Table 1). In TAA/
acarbose NE1 of Trp83 (corresponds to Met53
AMY1
)made
a hydrogen bond with O6 of the glucose ring at subsite )2
[17], and in porcine [18] and human pancreatic a-amylases
[19] NE2 of Gln63 participated in an analogous hydrogen
bond, as did Trp101 NE1 in cyclodextrin glycosyltransf-
erase from Bacillus circulans [20,21]. Of the non-a-amylase
members which do not utilize subsite )2, trehalose-6-
phosphate hydrolase, oligo-1,6-glucosidase, and a-glucosi-
dase have Asp, sucrose phosphorylase has Thr, and
different glucansucrases (GH13 and GH70) have Ala.
Interestingly, neopullulanase that produces panose from
pullulan and thus has an O6-substituted glucosyl ring at
subsite )2, also has Asp aligned to Met53 (Table 1).
However, because Trp, Asp, Gln, Leu, Gly, and His
corresponding to Met53

subsites similarly had high affinity at subsite-6 [65]. The
action pattern changes for the Met53 mutants produced in
bfia loop 2 showed that modification at subsite )2 could
importantly influence utilization of the outermost subsite-6.
Such long-range interactions in the substrate-mutant
enzyme complex between subsite )2 and other parts of
the binding cleft emphasized the intimate contact in
between bfia loop 2 and domain B and its importance
in activity [15].
The activity towards insoluble Blue Starch of seven
Met53
AMY1
mutants representing characteristic GH-H side
chains varied from being slightly superior (Met53Glu) to
less than 0.1% (Met53Trp) of wild-type AMY1. Interest-
ingly, Trp was present in many fungal and bacterial
a-amylases and certain other GH13 members. The muta-
genized position as evident from the crystal structure was
expected to play a major role in activity. Even with amylose
DP17, that filled the entire binding site, all Met53 mutants
displayed 5- to 20-fold higher K
m
accompanied by large
variation in k
cat
ranging from values similar to wild-type to
0.3% of its value. Indeed some of these mutants
(Met53Asp/Ala) had high activity for insoluble Blue Starch
and moderate k
cat

duction of an aromatic side chain in the middle of the
AMY1 binding cleft which apparently disturbed crucial
steps in the mechanism, perhaps involving contacts between
domains A and B. This effect on substrate transition state
stabilization suggested the presence of active site interac-
tions which would normally control development both of
substrate distortion and a negative electrostatic field at the
site of catalysis. The bulky side-chain in the cleft in
Ó FEBS 2002 Met53 mutants at subsite )2 in barley a-amylase 1 (Eur. J. Biochem. 269) 5387
Met53Trp AMY1 suppressed binding at subsites )3
through )6 as demonstrated in the action pattern analysis.
Further protein engineering, however, would be needed to
convert this endo-acting into an exo-acting a-amylase such
as the natural maltotetraose-forming exo-amylase [10] or
B. stearothermophilus maltogenic a-amylase [23].
Met53 was indicated in the modelled AMY2/maltodeca-
ose to contribute to the high affinity of subsite )2 [45,46],
perhaps via van der Waals’ interactions as a few plant
a-amylases had leucine at this position and a binding role of
SD Met53 seemed not adopted in Met53Asp/Glu AMY1
having high K
m
and low k
cat
/K
m
. Thus substrate hydrogen
bonding, in contrast to the situation seen for the corres-
ponding Trp and Gln in animal, fungal, and bacterial
a-amylases, may not play a critical role for this residue from

5
. Although the six other
mutants had essentially the same binding mode preference,
transglycosylation from PNPG
5
was not demonstrated,
probably due to a different balance between transglycosy-
lation and hydrolysis rates. From 10 m
M
PNPG
6
, how-
ever, both Met53Tyr and Met53Ala AMY1 catalyzed
transglycosylation to significant albeit small extent. The
earlier unique transglycosylation by the corresponding
Trp84Leu S. fibuligera a-amylase was explained by the
longer retention at the active site of the substrate glycon
part after cleavage [64]. The Trp84Leu thus enhanced the
transglycosylation/hydrolysis ratio of that enzyme. This
explanation may also apply to the AMY1 mutant,
although the shape of the binding cleft of S. fibuligera
a-amylase is very similar to that of TAA [60] and thus
different from AMY1 (Fig. 1A,B).
CONCLUSION
In AMY1, Met53 was required for wild-type kinetic
properties especially for affinity and in action on malto-
oligosaccharides and maltodextrins. Indeed substitution of
Met53 enabled modulation of activity and kinetic param-
eters for maltodextrins. Introduction of a bulky aromatic
group misguided the substrate glycon part to loose inter-

structures for the second and third (domain B) bfia
connecting segments that create the glycon binding region.
Although comprehensive sequence/specificity correlation
was not demonstrated for the short motif in bfia loop 2,
noticeably introduction of Asp, which is rare in a-amylases
but common in other GH-H members, maintained starch
hydrolysis at 90% of the wild-type AMY1 activity, whereas
introduction of Gly, another rare residue in a-amylases,
gave only 35%. It is proposed that combination of the
present mutations and mutations at other subsites can
accentuate the suppression of activity for shorter substrates
and further develop the enzyme specificity as done recently
for dual subsite mutants in AMY1 [37,38,61,62]. To this end
the present mutants may also be put through a directed
evolution programme.
ACKNOWLEDGEMENTS
The authors are grateful to C. Vincentsen for expert technical
assistance, L. H. Sørensen and the late B. Corneliussen for amino acid
analysis, and M B. Rask and the late J. Sage for DNA sequencing.
This work was supported by the EU 4th Framework Programme on
Biotechnology (CT98-0022) to the project AGADE.
REFERENCES
1. Yamamoto, T. (1995) Enzyme Chemistry and Molecular Biology of
Amylases and Related Enzymes. CRC Press, Inc., Boca Raton, FL.
2. Svensson, B. (1994) Protein engineering in the a-amylase family:
catalytic mechanism, substrate specificity, and stability. Plant Mol.
Biol. 25, 141–157.
3. MacGregor, E.A., Janecek, S. & Svensson, B. (2001) Relationship
of sequence and structure to specificity in the a-amylase family.
Biochim. Biophys. Acta 1546, 1–20.

maltotetraose-forming exo-amylase from Pseudomonas stutzeri.
J. Mol. Biol. 267, 661–672.
11. Katsuya, Y., Mezaki, Y., Kubota, M. & Matsuura, Y. (1998)
Three-dimensional structure of Pseudomonas isoamylase at 2.2 A
˚
resolution. J. Mol. Biol. 281, 885–897.
12. Kamitori, S., Kondo, S., Okuyama, K., Yokota, T., Shimura, Y.,
Tonuzuka, T. & Sakano, Y. (1999) Crystal structure of Thermo-
actinomyces vulgaris R-47 a-amylase II (TVAII) hydrolyzing
cyclodextrins and pullulan at 2.6 A
˚
resolution. J. Mol. Biol. 287,
907–921.
13. Kim, J.S., Cha, S.S., Kim, H.J., Kim, T.J., Ha, N.C., Oh, S.T.,
Cho, H.S., Cho, M.J., Kim, M.J., Lee, H.S., Kim, J.W., Choi,
K.Y., Park, K.H. & Oh, B.H. (1999) Crystal structure of a mal-
togenic amylase provides insights into a catalytic versatility. J. Biol.
Chem 274, 26279–26286.
14. Przylas, I., Tomoo, K., Terada, Y., Takaha, T., Fujii, K., Saenger,
W. & Stra
¨
ter, N. (2000) Crystal structure of amylomaltase from
Thermus aquaticus, a glycosyltransferase catalysing the production
of large cyclic glucans. J. Mol. Biol. 296, 873–886.
15. Kadziola, A., Abe, J., Svensson, B. & Haser, R. (1994) Crystal and
molecular structure of barley a-amylase. J. Mol. Biol. 239, 104–
121.
16. 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,

J. Mol. Biol. 277, 393–407.
23. Dauter, Z., Dauter, M., Brzozowski, A.M., Christensen, S.,
Borchert, T.V., Beier, L., Wilson, K.S. & Davies, G.J. (1999)
X-ray structure of Novamyl, the five domain ÔmaltogenicÕ
a-amylase from Bacillus stearothermophilus:maltoseandacarbose
complexes at 1.7 A
˚
resolution. Biochemistry 38, 8385–8392.
24. 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.
25. Uitdehaag, J.C.M., van Alebeek, G J.W.M., van der Veen, B.A.,
Dijkhuizen, L. & Dijkstra, B.W. (2000) Structures of malto-
hexaose and maltoheptaose bound at the donor sites of cyclo-
dextrin glycosyltransferase give insight into the mechanisms of
transglycosylation activity and the cyclodextrin size specificity.
Biochemistry 39, 7772–7780.
26. Yoshioka, Y., Hasegawa, K., Matsuura, Y., Katsube, Y. &
Kubota, M. (1997) Crystal structures of a mutant maltotetraose-
forming exo-amylase cocrystallized with maltopentaose. J. Mol.
Biol. 271, 619–628.
27. Machius, M., Ve
´
rtesy, L., Huber, R. & Wiegand, G. (1996) Car-
bohydrate and protein-based inhibitors of porcine pancreatic
a-amylase: structure analysis and comparison of their binding
characteristics. J. Mol. Biol. 260, 409–421.
28. Ramasubbu, N., Paloth, V., Luo, Y., Brayer, G.D. & Levine, M.J.

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.
36. Matsui, I. & Svensson, B. (1997) Improved activity and modulated
action pattern obtained by random mutagenesis at the fourth
b-a loop involved in substrate binding to the catalytic (b/a)
8
-barrel
domain of barley a-amylase 1. J. Biol. Chem. 272, 22456–22463.
37. Gottschalk, T.E., Tull, D., Aghajari, N., Haser, R. & Svensson, B.
(2001) Specificity modulation of barley a-amylase through biased
random mutagenesis involving a conserved tripeptide in bfia
loop 7 of the catalytic (b/a)
8
-barrel domain. Biochemistry 40,
12844–12854.
38. Mori, H., Bak-Jensen, K.S., Gottschalk, T.E., Motawia, M.S.,
Damager, I., Møller, B.L. & Svensson, B. (2001) Modulation of
Ó FEBS 2002 Met53 mutants at subsite )2 in barley a-amylase 1 (Eur. J. Biochem. 269) 5389
activity and substrate binding modes by single and double subsites
+1/+2 and )5/)6 mutation of barley a-amylase 1. Eur. J. Bio-
chem. 268, 6545–6558.
39. Svensson, B., Mundy, J., Gibson, R.M. & Svendsen, I. (1985)
Partial amino acid sequences of a-amylase isozymes from barley
malt. Carlsberg Res. Commun. 50, 15–22.
40. Jones, R.L. & Jacobsen, J.V. (1991) Regulation of synthesis and
transport of secreted proteins in cereal aleurone. Int. Rev. Cytol.
126, 49–88.
41. Bertoft, E., Andtfolk, C. & Kulp, S E. (1984) Effect of pH,
temperature, and calcium ions on barley malt a-amylase isoen-

from the a-amylase family. J. Biol. Chem. 276, 25273–25278.
48. Monchois, V., Willemot, R M. & Monsan, P. (1999) Glucansu-
crases: Structure-function relationships. FEMS Microbiol. Rev. 23,
131–151.
49. Søgaard, M. & Svensson, B. (1990) Expression of cDNAs
encoding barley a-amylase 1 and 2 in yeast andcharacterization of
the secreted proteins. Gene 94, 173–179.
50. Juge, N., Andersen, J.S., Tull, D., Roepstorff, P. & Svensson, B.
(1996) Overexpression, purification, and characterization of
recombinant barley a-amylase 1 and 2 secreted by the methylo-
trophic yeast Pichia pastoris. Protein. Expr. Purif. 8, 204–214.
51. Søgaard, M., Olsen, F.L. & Svensson, B. (1991) C-terminal pro-
cessing of barley a-amylase 1 in malt, aleurone protoplasts, and
yeast. Proc.Natl.Acad.Sci.USA88, 8140–8144.
52. Søgaard, M., Andersen, J.S., Roepstorff, P. & Svensson, B. (1993)
Electrospray mass spectrometry characterization of post-transla-
tion almodifications of barley a-amylase 1 produced in yeast. Bio/
Technology 11, 1162–1165.
53. Yanisch-Perron, C., Vieira, J. & Messing, J. (1985) Improved M13
phage cloning vectors and host strains: nucleotide sequences of the
M13mp18 and pUC19 vectors. Gene 33, 103–119.
54. Cregg, J.M., Barringer, K.J., Hessler, A.Y. & Madden, K.R.
(1985) Pichia pastoris as a host system for transformations. Mol.
Cell. Biol. 5, 3376–3385.
55. Ellis, S.B., Brust, P.F., Koutz, P.J., Waters, A.F., Harpold, M.M.
& Gingeras, T.R. (1985) Isolation of alcohol oxidase and two
other methanol regulatable genes from the yeast Pichia pastoris.
Mol. Cell. Biol. 5, 1111–1121.
56. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular
Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor

64. Matsui, I., Ishikawa, K., Miyairi, S., Fukui, S. & Honda, K.
(1991) An increase in the transglycosylation activity of Saccharo-
mycopsis a-amylase altered by site-directed mutagenesis. Biochim.
Biophys. Acta 1077, 416–419.
65. Mori, H., Tatematsu, A., Saito, A., Matsui, H., Kimura, A. &
Chiba, S. (1995) Substrate specificitiy and subsite affinities of
a-amylase form germinating cotyledons of Phaseolus vulgaris L. cv
Toramame. Oyo Toshitsu Kagaku (J. Glycosci.) 42, 387–394.
66. Uitdehaag, J.C.M., Mosi, R., Kalk, K.H., van der Veen, B.A.,
Dijkhuizen, L., Withers, S.G. & Dijkstra, B.W. (1999) X-ray
structures along the reaction pathway of cyclodextrin glycosyl-
transferase elucidate catalysis in the a-amylase family. Nature
Struct. Biol. 6, 432–436.
67. 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.
5390 H. Mori et al. (Eur. J. Biochem. 269) Ó FEBS 2002