Báo cáo khoa học: Donor and acceptor substrate selectivity among plant glycoside hydrolase family 32 enzymes doc

REVIEW ARTICLE
Donor and acceptor substrate selectivity among plant
glycoside hydrolase family 32 enzymes
Wim Van den Ende
1
, Willem Lammens
1,2
, Andre
´
Van Laere
1
, Lindsey Schroeven
1
and Katrien Le Roy
1
1 K.U.Leuven, Laboratory for Molecular Plant Physiology, Heverlee, Belgium
2 K.U.Leuven, Laboratory for Biocrystallography, Leuven, Belgium
Introduction
Fructans are polymers of fructose (Fru), and an exten-
sion of sucrose (Suc), and they occur in many plant
species mainly belonging to Asteraceae, Liliaceae and
Poaceae [1,2]. Apart from their function as a vacuolar
storage carbohydrate, fructans may protect plants
from drought and cold stress by stabilizing cellular
membranes [3–5]. Furthermore, they might participate
in vacuolar antioxidant mechanisms [6], consistent
with earlier [7] and more recent [8,9] observations on
transgenic plants carrying fructan-synthesizing genes.
Recently, substantial efforts have been made to under-
stand fructan metabolism and its regulation in impor-
tant cereals [10–12] and forage grasses [13]. Different

fructan types (inulin, levan, graminan, neoseries)
can be distinguished by the type of linkage present
between the fructosyl residues [b(2-1) versus b(2-6)]
and by the position of the glucose (Glc) residue
[14]. Depending on the plant species, fructans are
produced by different types of fructan biosynthetic
enzymes (a minimum of two enzymes) or fructosy-
ltransferases (FTs), which have probably evolved from
vacuolar-type invertases (VIs) [15]. Sucrose:sucrose 1-
fructosyltransferase (1-SST) seems to be a key enzyme
in both monocots and dicots for initiating the fructan
polymerization process. However, 1-SST might not ful-
fil a prominent role in species exclusively producing
levan-type fructans [16]. Critical-elongation FTs include
fructan:fructan 1-fructosyltransferases (1-FFTs; inulin
Keywords
fructan; fructosyl transferase; invertase;
structure–function; sucrose
Correspondence
W. Van den Ende, K.U.Leuven, Laboratory
for Molecular Plant Physiology, Kasteelpark
Arenberg 31, B-3001, Leuven-Heverlee,
Belgium
Fax: +32 16321967
Tel: +32 16321952
E-mail:
(Received 18 June 2009, revised 7 August
2009, accepted 19 August 2009)
doi:10.1111/j.1742-4658.2009.07316.x
Plant family 32 glycoside hydrolase enzymes include hydrolases (cell wall

invertases, fructan exohydrolases, vacuolar invertases) and fructosyltrans-
ferases. These enzymes are very similar at the molecular and structural
levels but are functionally different. Understanding the basis of the func-
tional diversity in this family is a challenging task. By combining structural
and site-directed mutagenesis data, Asp239 in AtcwINV1 was identified as
an amino acid critical for binding and stabilizing sucrose. Plant fructan
exohydrolases lack such an Asp239 equivalent. Substitution of Asp239 led
to the loss of invertase activity, while its introduction in fructan exohydro-
lases increased invertase activity. Some fructan exohydrolases are inhibited
by sucrose. The difference between the inhibitor (fructan exohydrolase) and
the substrate (invertase) binding configurations of sucrose can be explained
by the different orientation of Trp82. Furthermore, the evolutionary hydro-
lase ⁄ transferase transition could be mimicked and the difference between
S-type fructosyltransferases (sucrose as donor) and F-type fructosyltransfe-
rases (fructan as donor) could be unravelled.
Abbreviations
1-KEH, 1-kestose exohydrolase; CWI, cell wall invertase; FEH, fructan exohydrolase; FFT, fructan:fructan fructosyltransferase; Fru, fructose;
FT, fructosyltransferase; GH32, glycoside hydrolase family 32; Glc, glucose; SFT, sucrose:fructan fructosyltransferase; SST, sucrose:sucrose
fructosyltransferase; Suc, sucrose; VI, vacuolar invertase; WT, wild-type.
5788 FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS
synthesis), sucrose:fructan 6-fructosyltransferases (6-
SFTs; graminan synthesis) and fructan:fructan
6
G
-fructosyltransferases (6
G
-FFTs; inulin neoseries)
[15]. Plants degrade fructans using fructan exohydrolas-
es (FEHs). Similarly to FTs, many different types of
FEHs can be discriminated [17]. Intriguingly, these

FEHs seem to be present in all plants, including those
not accumulating fructans at all. Unravelling the
function of these ‘defective invertases’ (containing FEH
side activity) is a hot topic in this research area.
While fructans are degraded by FEHs, the cleavage of
Suc into Glc and Fru is catalyzed by two classes of acid-
type plant invertases: the cell wall invertases (CWIs) and
the VIs. Phylogenetically, two different groups can be
differentiated in plants: the CWI ⁄ FEH group on the one
hand and the VI ⁄ FT group on the other hand.
All the different types of FTs, FEHs, CWIs and VIs
are grouped together with microbial b-fructosidases
(degrading both sucrose and fructans) in the glycoside
hydrolase family 32 (GH32) ()
[18]. GH32 can be combined with GH68 in the clan
GH-J. GH68 harbours bacterial invertases, levansuc-
rases and inulosucrases. Recently, several 3D struc-
tures have been unraveled within GH32 [15]. All these
proteins consist of an N-terminal five-bladed b-propel-
ler domain (GH32 and GH68) followed by a C-termi-
nal domain formed by two b-sheets (only in GH32).
The active site is present within the b-propeller domain
and is characterized by the presence of three highly
conserved acidic groups (present in the WMNDPNG,
RDP and EC motifs). The Asp from the first motif
acts as a nucleophile, the Asp from the second motif
is believed to be a transition state stabilizer and the
Glu residue from the EC motif acts as an acid ⁄ base
catalyst, playing a crucial role in the catalytic mecha-
nism [19–21]. According to the )n to +n subsite

nomenclature proposed by Davies et al. [22], hydroly-
sis takes place between the )1 and +1 subsites. For
instance, when Suc binds as a donor substrate, the Fru
moiety positions at the )1 subsite and the Glc unit is
bound at the +1 subsite.
The availability of the 3D structures helped greatly
in the design of very specific site-directed mutagenesis
experiments, which were carried out with the aim of
understanding the molecular basis for the different
substrate specificities within plant GH32 enzymes.
Here we summarize the recent progress in this area.
Understanding substrate specificity
within the CWI
⁄
FEH group
Differences in the active sites of AtcwINV1 and
Ci1-FEHIIa
The first plant enzyme 3D structures that became
available within GH32 were Cichorium intybus 1-FEH
IIa (Ci1-FEHIIa) [23] and AtcwINV1 (a CWI from
Arabidopsis thaliana) [24]; both are hydrolases differing
strongly in donor substrate specificity. Indeed, Ci1-FE-
HIIa cannot degrade sucrose, while sucrose is the
preferential substrate for AtcwINV1. Multiple aligment
studies with other plant GH32 members revealed that
both AtcwINV1 and Ci1-FEHIIa contain eight
conserved motifs in the active-site region (Fig. 1A).
Additionally, one hypervariable loop is located very
close to the acid–base catalyst, and this loop is clearly
different between AtcwINV1 (KISLDDTKH) and

Ci1-FEHIIa (KADFEG H), the Ci1-FEHIIa showing
a double deletion in this area (Fig. 2). AtcwINV1
shows an Asp239⁄ Lys242 couple (strong hydrogen
A
B
Fig. 1. Schematic representation of
AtcwINV1 showing eight conserved regions
in the vicinity of the active site. The amino
acids involved in catalysis and in binding the
terminal fructose are in red. A hypervariable
loop that comes close to the acid–base
catalyst is shown in blue (A). Details of the
active sites of Ci1-FEHIIa and AtcwINV1 in
complex with sucrose [26]. The positions of
Fru ()1 subsite) and Glc (+1 subsite) are
indicated with arrows. Please note the
difference in the hypervariable loop and the
different orientation of both the Trp82 and
the Glc moiety in Ci1-FEHIIa compared with
AtcwINV1 (B). The figures were prepared
using
PYMOL [39].
W. Van den Ende et al. Donor and acceptor substrate selectivity in GH32
FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS 5789
bond), which is absent in Ci1-FEHIIa (Fig. 1B) and
also in all characterized FEHs (Fig. 2A). However,
such Asp ⁄ Lys or Asp ⁄ Arg couples are typically
observed among CWIs (Fig. 2B), although with some
exceptions (e.g. AtcwINV5).
Another striking difference is the presence of a Ser

residue in Ci1-FEHIIa (after the MLYTG motif),
while AtcwINV1 contains an Ile at this position
(Fig. 1). Intriguingly, among all plant FEHs character-
ized to date, we found a nice correlation between the
identity of the amino acid at this position and the
degree of inhibition by Suc. Indeed, all FEHs that
contain a small amino acid (such as Gly or Ser) at this
position are strongly inhibited by Suc (Fig. 3; [25]). A
last prominent difference in the active site between the
two structures is the different orientation of the Trp82
residue (Fig. 1B); this, of course, remained undetected
during multiple sequence alignments. Trp82 and
Ser101 are very close in space in the active site of
Ci1-FEHIIa (Fig. 1B).
A
B
Fig. 2. Multiple alignment of a selection of
FEHs (A) and CWIs (B) in the region
surrounding the hypervariable loop. The
position of the Asp ⁄ Lys or Asp ⁄ Arg couple
is indicated in bold in CWIs (B). FEHs (A)
and AtcwINV5 (B) contain alternative amino
acids at these positions (bold) or deletions
in this area. Functionally characterized
enzymes are marked by an asterisk.
Fig. 3. Comparison of the GWAS and
MLYTG motifs in FEHs and their inhibition
by sucrose.
Donor and acceptor substrate selectivity in GH32 W. Van den Ende et al.
5790 FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS

Asp239 fulfils a crucial role for binding Suc as a
substrate in the CWI
⁄
FEH group
Knowing these prominent differences in the active sites
of FEHs and invertases, we used a mutagenesis
approach to alter the donor substrate specificity within
plant GH32 enzymes. The Asp239 in AtcwINV1 was
changed into an Ala, a Phe and an Asn (Fig. 4; [21]).
Additionally, Lys242 was mutated into a Leu. The
invertase activity vanished in the D239A and D239F
mutants but remained more or less intact in the
D239N mutant. The kinetic parameters of the purified
mutant enzymes were determined and compared with
those obtained for the wild-type (WT) invertase
(Table 1; [21]). In both the D239A and D239F
mutants, the K
m
increased by 6–11-fold, respectively,
inferring an important role for the Asp239 residue in
substrate binding. The lower substrate affinity of
D239F compared with D239A can result from the
more extended steric hindrance of the bulky Phe. A
similar tendency was observed for k
cat
values, which
decreased 10–20-fold, suggesting that Asp239 is also
important for efficient catalysis. By contrast, the K
m
and k

cat
values of the D239N mutant differed only
slightly from those of the WT enzyme, convincingly
demonstrating that an acidic group is not essential at
this position and can be replaced by an Asn. Conclu-
sively, these data showed that the presence of an addi-
tional Asp or Asn residue, adjacent to the Glu203
proton donor, is important for optimal binding and
efficient catalysis of Suc.
To investigate whether the results obtained for the
Asp239 mutants were caused by the substitution of this
Asp residue itself, or were indirectly caused by the
disturbance of the tight Asp239–Lys242 interaction, a
K242L mutant was constructed and the kinetic proper-
ties of the purified enzymes were determined (Table 1).
The K
m
of the K242L mutant enzyme increased by a
factor comparable to those of the Asp239 mutants.
However, the k
cat
value only decreased by two-fold in
comparison with the WT enzyme. Therefore, a crucial
role for the Lys242 residue itself can be excluded.
However, regarding the conserved interaction between
Asp239 and Lys242, it is plausible that Lys242 is
necessary to keep Asp239 in the correct orientation
towards the active site.
When using 1-kestose as a substrate, the results pre-
sented in Table 1 show that the kinetics of the mutants

(D239A, D239F) did not change substantially com-
pared with the WT enzyme. So, the WT AtcwINV1
A
B
Fig. 4. Production of Glc and Fru from Suc by wild-type AtcwINV1
and several Asp239 mutants (D239A, D239F and D239N), as
revealed by high-pressure anion-exchange chromatography with
pulsed amperometric detection (HP-AEC-PAD; Dionex, Sunnyvale,
CA, USA). The reaction conditions were as follows: incubation of
50 ng of purified enzyme with 10 m
M Suc in 50 mM acetate buffer,
pH 5.0, for 30 min at 30 °C (A). Comparison of specific enzymatic
activities [s.a. in (mol fructose).(mol enzyme)
)1
.(s)
)1
] of the same
wild-type and mutant AtcwINV1 enzymes as a function of increas-
ing Suc concentrations (B). The figure is reproduced from that
presented in a previous publication [21].
Table 1. Kinetic parameters for the hydrolysis of sucrose and
1-kestose by purified WT AtcwINV1, D239A, D239F, D239N and
K242L mutant enzymes. The table was reproduced from a previous
publication [21].
Sucrose K
m
(mM) k
cat
(s
)1

) k
cat
⁄ K
m
(mM
)1
Æs
)1
)
AtcwINV1 0.35 ± 0.05 59 ± 4 168.6
D239A 2.1 ± 0.2 6 ± 0.3 2.9
D239F 4.5 ± 0.5 3 ± 0.2 0.7
D239N 0.6 ± 0.07 61 ± 5 101.7
K242L 3.7 ± 0.4 29 ± 1 7.8
1-Kestose
AtcwINV1 1 ± 0.1 21 ± 2 21.0
D239A 0.6 ± 0.03 21 ± 3 35.0
D239F 1.2 ± 0.1 20 ± 2 16.7
W. Van den Ende et al. Donor and acceptor substrate selectivity in GH32
FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS 5791
contains an intrinsic 1-kestose exohydrolase (1-KEH)
activity, which remains intact in the D239A mutant. It
can be concluded that the substitution of Asp239 very
selectively destroys the invertase activity, but not the
1-KEH activity.
The importance of a structurally equivalent Asp239
residue, and the presence of a conserved Asp ⁄ Lys or
Asp ⁄ Arg couple to bind and stabilize Suc as the prefer-
ential donor substrate in invertases, was confirmed by
generating several mutated AtcwINV1–Suc complexes.

These complexes nicely demonstrated the strong
interaction between Asp239 and the Glc moiety of Suc
(Fig 1B, [15,26]).
By using the presence of an intact Asp239 homo-
logue as a selective marker to discriminate between
real and defective invertases, we predicted that three
out of the six (i.e. 50%; AtcwINV3, 5 and 6; Fig. 5) of
the so-called CWIs from this species are not real inver-
tases but defective invertases (FEHs). This functional-
ity has already been proven for AtcwINV3 (termed a
6-FEH) and AtcwINV6 (termed a 6&1 FEH) [27].
For the next challenge, it was questioned whether
introduction of an Asp239 homologue in an FEH
would result in the introduction of sucrose hydrolyzing
activity. As the introduction of a structural and
functional Asp239 homologue in Ci1-FEH IIa is com-
plex because of the presence of a double deletion (see
also Fig. 2A), site-directed mutagenesis experiments
were performed on Beta vulgaris 6-FEH [28], which
lacks such a deletion and which is characterized by the
presence of a nonacidic Asp239 structural equivalent
(a bold Phe residue: see Fig. 2). Compared with the
WT sugar beet 6-FEH, a F233D mutant indeed
showed substantial invertase activity, especially at
higher Suc concentrations (Fig. 6A).
Trp82 fulfils a crucial role for binding Suc as
inhibitor in some FEHs
Some FEHs are strongly inhibited by Suc, and others
or not. Ci1-FEHIIa is a typical example of an FEH
that is strongly inhibited. The presence of a small

amino acid (Ser101) next to Trp82 in Ci1-FEHIIa is
believed to be important for determining the orienta-
tion of the Trp82 residue. The Glc moiety of Suc in
Ci1-FEHIIa occupied a position clearly different [25]
Fig. 5. Models of the active sites of the six putative CWIs from the model plant Arabidopsis thaliana. AtcwINV1, 2 and 4 contain a correctly
orientated Asp239 homologue (indicated with an arrow) close to the acid–base catalyst Glu203 (AtcwINV1 terminology). AtcwINV3, 5 and 6
lack the Asp239 equivalent. The active site residues are marked in red. Modelling was performed based on the known 3D structures of
Ci1-FEHIIa and AtcwINV1 using
SWISS MODEL (). The figures were prepared using PYMOL [39].
Donor and acceptor substrate selectivity in GH32 W. Van den Ende et al.
5792 FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS
from the ones observed in the bacterial levansucrase
[29] and in AtcwINV1 [26]. The Glc moieties are
moved away from each other by ± 3 A
˚
, but neverthe-
less they still stay in the same plane more or less
alongside each other (Fig. 1B).
The configuration of the Glc moiety of Suc present
in Ci1-FEHIIa can be considered as the ‘inhibitor con-
figuration or inhibitor-binding modus’, whereas the
Suc orientation in levansucrase or AtcwINV1 is termed
‘substrate configuration or substrate-binding modus’,
because Suc acts as an inhibitor in Ci1-FEHIIa and as
a substrate for the other enzymes. Interestingly, chang-
ing the Ser101 from Ci1-FEHIIa into a Leu also
resulted in an increased invertase activity (Fig. 6B),
strongly suggesting that the nature of this amino acid
is essential to discriminate whether Suc will bind as a
substrate rather than as an inhibitor. A W82L mutant

had invertase activity similar to that of the WT
enzyme (Fig. 6B). It is a challenging task to figure out
whether defective invertases occurring in nonfructan-
accumulating plants would also be able to bind Suc in
the inhibitor configuration [30].
Transforming a VI into a high-affinity FT
Destroying the hydrogen bond network in the
WMNDPNG motif is essential to create an FT
from TaVI
Both FTs (sucrose ⁄ fructan as acceptor substrate: trans-
ferase activity) and VIs (water as acceptor: hydrolase
activity) occur within plant GH32 members. A rather
limited increase in transfructosylation capability was
already realized by mutagenesis adjacent to the nucleo-
phile in the b-fructosidase or WMNDPNG motif of
onion VI [31], but the mutant enzyme still mainly
behaved as an invertase with fully saturable kinetics
for hexose production (hydrolytic reaction) and no full
saturation was observed for 1-kestose production
(transfer reaction). Triticum aestivum (wheat) is an eco-
nomically important species and an ideal model plant
for using to perform detailed structure–function work
on VIs and FTs. Moreover, it is a unique fact that
Ta1-SST, Ta6-SFT, Ta1-FFT and TaVI recombinant
enzymes (each of which were derived after hetero-
logous expression in Pichia pastoris) are all available
[32,33], which results in an excellent system for site-
directed mutagenesis, especially because the percentage
of identity is very high among TaVIs and TaFTs [34].
In an attempt to understand the evolution of an FT

from an ancestral VI, multiple sequence alignments of
VIs and Suc-splitting FTs (S-type FTs such as 1-SSTs
and 6-SFTs) revealed prominent differences in two
conserved VI regions, namely the WMNDPNG and
GWAS motifs [34]. Because 1-SSTs of Poaceae always
contain a Tyr instead of a Trp in the WMNDPNG
motif, the W23Y TaVI mutant was constructed. In
addition, because 1-SSTs of dicotyledonous plants and
6-SFTs from Poaceae contain a Ser instead of an Asn
adjacent to the nucleophile, the TaVI N25S and
W23Y+N25S mutants were designed.
Intriguingly, in all plant invertases (VIs as well as
CWIs), a well-defined hydrogen bond network is con-
served between the Trp23 and Asn25 equivalents and
the nucleophile. Strikingly, modelling studies indicated
that this hydrogen bond network is destroyed in all
plant FTs [34]. Therefore, we wanted to test the
hypothesis of whether the destruction of the hydrogen
bond network in TaVI would result in a mutant
enzyme with increased FT capability.
Comparing transfructosylation capabilities of WT
and mutant TaVIs
Reaction mixtures containing the purified heterolo-
gously expressed TaVI, Ta1-SST and Ta6-SFT WT and
A

B
Fig. 6. Specific sucrose-hydrolyzing activities [s.a. in (mol fruc-
tose).(mol enzyme)
)1

.(s)
)1
]ofBeta vulgaris wild-type 6-FEH and
F233D mutant enzymes as function of increasing sucrose concen-
trations (A). Specific sucrose-hydrolyzing activities [s.a. in (mol fruc-
tose).(mol enzyme)
)1
.(s)
)1
] of wild-type Ci1-FEHIIa and S101L
mutant enzymes as a function of increasing sucrose concentrations
(B). Results are the mean (± SE) for three replicates. The figure is
reproduced from that presented in a previous publication [30].
W. Van den Ende et al. Donor and acceptor substrate selectivity in GH32
FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS 5793
mutant TaVI enzymes were compared after incubation
with 500 mm Suc as a single substrate (Fig. 7). The WT
TaVI showed no significant production of 1-kestose.
The W23Y mutant synthesized a small amount of
1-kestose, comparable to the amount produced by the
WT Ta6-SFT. Additionally, the WT Ta6-SFT also
produced a small amount of 6-kestose. Compared with
the W23Y mutant, the N25S mutant produced a
greater amount of 1-kestose. Interestingly, a synergistic
effect is observed when the two mutations are com-
bined (W23Y+N25S), resulting in a strongly increased
1-kestose synthesis and a decreased Fru to Glc ratio.
The WT Ta1-SST, showing nearly zero hydrolytic
activity at 500 mm Suc, is shown for comparison
(Fig. 7; [34]).

The percentage of 1-kestose (transfer to Suc) over
the total of Fru (transfer to water: hydrolysis) and
1-kestose is a measure of the transfructosylation capac-
ity. Typically, the transfructosylation capacity increases
with increasing substrate concentration; this is the case
for all enzymes tested (Fig. 8; for details see [34]). A
clear shift from hydrolysis to transfructosylation was
observed for the mutant TaVIs, especially at higher
Suc concentrations. This change in acceptor substrate
specificity was most prominent for the double mutant
W23Y+N25S, showing a 17-fold increase in transfruc-
tosylation capacity and reaching a maximal transfruc-
tosylation capacity of more than 50% (Fig. 8). It can be
hypothesized that this capacity can be further increased
by changing the WGW motif (Fig. 1) into a WGY
motif, because the W to Y transition was also shown to
increase the transfructosylation capability [35].
We refer to Schroeven et al. [34] for a detailed kinet-
ical analysis on the WT and mutant enzymes. Conclu-
sively, it is demonstrated, for the first time, that it is
possible to create a FT (a 1-SST in this particular case)
with fully saturable kinetics (1-kestose production) and
a very high affinity for Suc (the K
m
is more than 10
times lower compared with that of the WT Ta1-SST).
Such mutated enzymes are extremely useful for bio-
technological applications, such as producing tailor-
made fructans in transgenic plants or in bioreactors.
Transforming a 6

G
-FFT into a 1-SST
1-SST and 6
G
-FFT enzymes in Lolium perenne
FTs might be crucial for plant survival under stress
conditions in species where fructans represent the
major form of reserve carbohydrate, such as in peren-
nial ryegrass (Lolium perenne). Perennial ryegrass is an
economically important species (the number 1 plant in
fructan research at this moment) and forms an ideal
system for conducting detailed structure–function work
on FTs and VIs. Indeed, many cDNAs have become
available from this species in the last 5 years [36,37].
Moreover, the overall identity between these cDNAs
is very high, and recombinant enzymes are available
through heterologous expression in P. pastoris. The
Fig. 7. High-pressure anion-exchange chromatography with pulsed
amperometric detection (HP-AEC-PAD; Dionex) chromatograms of
reaction mixtures of WT Triticum aestivum TaVI, Ta6-SFT, Ta1-SST
and mutant VI W23Y, N25S and W23Y+N25S enzymes. All
enzymes were incubated with 500 m
M Suc at 30 °C for 10 min (VI
and mutants derived thereof), 30 min (6-SFT) and 2 h (1-SST). 1-K,
1-kestose; 6-K, 6-kestose. The figure is reproduced from that pre-
sented in a previous publication [34].
Fig. 8. The percentage of transfructosyla-
tion at increasing concentrations of Suc (25–
1000 m
M). From left to right: WT Ta1-SST,

WT TaVI, TaVI mutants W23Y+N25S, N25S
and W23Y. The figure is reproduced from
that presented in a previous publication [34].
Donor and acceptor substrate selectivity in GH32 W. Van den Ende et al.
5794 FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS
very high identity at the sequence level between Lp6
G
-
FFT ⁄ 1-FFT (an enzyme from L. perenne showing both
6
G
-FFT and 1-FFT activities) and Lp1-SST inspired
us to study, in depth, the molecular differences
between these functionally different enzymes. Despite
their high identity, these enzymes differ greatly in sub-
strate specificity and product formation, the two most
crucial differences being (a) Lp1-SST uses Suc as a
donor substrate while Suc is a very poor donor sub-
strate for the recombinant Lp6
G
-FFT ⁄ 1-FFT, and (b)
Lp1-SST can only create a b(2-1) linkage between two
fructosyl residues while Lp6
G
-FFT ⁄ 1-FFT can create
both a b(2-1) linkage between two fructosyl residues
and a b(2-6) linkage between a fructosyl residue and a
glucosyl residue.
Designing specific mutants based on multiple
sequence alignments and modelling

In the mature proteins of Lp6
G
-FFT ⁄ 1-FFT and Lp1-
SST, 83% of the amino acids are identical. As a first
step to understand the difference at the molecular level
between these enzymes, a multiple sequence alignment
was made for the large subunits of Lp1-SST, Lp6
G
-
FFT ⁄ 1-FFT and all 6
G
-FFT types of enzymes charac-
terized to date [38]. Ten amino acids were unique in
6
G
-FFTs and could not be found in Lp1-SST. Among
those, only three amino acids (arrows in Fig. 9) were
located in the vicinity of the active-site region, as
revealed by modelling studies (Fig. 10). These amino
acids are Asn340, Trp343 and Ser415 in Lp6
G
-FFT ⁄ 1-
FFT (Fig. 10A), which are replaced by Asp349,
Arg352 and Asn424 in Lp1-SST (Fig. 10B). Both
Asn340 and Trp343 are present in a hypervariable
loop very close to the acid–base catalyst. Therefore,
the single mutants N340D, W343R and S415N, the
double mutants N340D+W343R, N340D+S415N
and W343R+S415N, and finally the triple mutant,
N340D+W343R+S415N, were constructed and trans-

formed into P. pastoris for functional characterization
and comparison with the WT enzymes Lp6
G
-FFT ⁄ 1-
FFT and Lp1-SST [38].
1-SST activity in mutant and WT enzymes:
transforming F-type into S-type
The WT Lp6
G
-FFT ⁄ 1-FFT showed a low, but intrin-
sic, 1-SST activity (Table 2; [37]). The N340D,
Fig. 9. Amino acid composition in the active-site motifs of WT Lp1-
SST and WT Lp6
G
-FFT ⁄ 1-FFT and mutants derived of the latter.
Amino acids near the active site that are unique in Lp6
G
-FFT ⁄ 1-FFT
compared with Lp1-SST are shown in bold. Two of these amino
acids (Asn340 and Trp343) are present in the hypervariable loop.
A
B
Fig. 10. Localization of the considered amino acids in the active
sites of the WT Lp6
G
-FFT ⁄ 1-FFT (A) and WT Lp1-SST (B). The
active site residues are marked in red, and the amino acids sub-
jected to mutagenesis are indicated with an arrow. Modelling was
performed based on the 3D structure of AtcwINV1 using
SWISS

MODEL
(). The figures were prepared
using
PYMOL [39].
W. Van den Ende et al. Donor and acceptor substrate selectivity in GH32
FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS 5795
W343R, S415N, N340D+S415N and W343R+S415N
mutants also showed very low 1-SST activity. Most
importantly, while the N340D+W343R mutant
showed 6
G
-FFT and 1-FFT activities that lay between
those of the N340D and W343R single mutants [38],
its 1-SST activity drastically increased (Table 2). The
1-SST activity increased further in the triple
N340D+W343R+S415N mutant, while its intrinsic
1-FFT activity was further decreased. At high Suc
concentrations (1.0 M), the triple Lp6
G
-FFT ⁄ 1-FFT
mutant produced the same amount of 1-kestose as the
WT Lp1-SST [38], demonstrating, for the first time,
the transformation of an F-type of enzyme into an
S-type enzyme by introducing a functional Asp ⁄ Arg
couple in the hypervariable loop. The interested reader
is referred to a previous publication to read further
details on the kinetics and substrate specificities of the
different mutant and WT enzymes [38].
Donor substrate selectivity: similarities between
the CWI

⁄
FEH and VI
⁄
FT subgroups
The results show that Asn340 and Trp343 are impor-
tant determinants for explaining the particular donor
substrate characteristics of the WT Lp 6
G
-FFT ⁄ 1-FFT
and WT Lp1-SST. What are the structural equivalent
amino acids in other FTs and in the structurally well
characterized Ci1-FEHIIa and AtcwINV1? Figure 11
shows a multiple alignment of a selection of FTs
together with Ci1-FEHIIa and AtcwINV1. Modelling
studies proposed that the N340 ⁄ W343 couple is equiva-
lent to the D239 ⁄ K242 couple in AtcwINV1. The
presence of an Asp ⁄ Lys or an Asp ⁄ Arg couple in the
hypervariable loop is believed to be essential for
binding Suc in the CWI ⁄ FEH group (see above). Simi-
larly, the same Asp ⁄ Arg couple seems to determine the
preference for Suc as the donor substrate within the
VI ⁄ FT subgroup. It can be concluded that the presence
of a functional Asp ⁄ Lys or Asp ⁄ Arg couple (not
disturbed by other amino acids in the area) determines
the Suc donor substrate selectivity in all plant GH32
enzymes. In particular, all enzymes that are able to use
Suc as a donor substrate (1-SSTs, 6-SFTs, VIs and
CWIs) contain this couple, while it is absent in all
typical F-type enzymes (FEHs, 1-FFTs and 6
G

-FFTs)
(Fig. 11). It is a matter of discussion whether this point
of view can be extended to all microbial GH32 enzymes
and to GH68 enzymes [38].
Conclusions
The availability of the 3D structures of Ci1-FEHIIa and
AtcwINV1 boosted structure–function research on plant
GH32 members. Enzymes preferentially using Suc as the
donor substrate (S-type enzymes) show a functional
Asp ⁄ Arg or Asp ⁄ Lys couple in a hypervariable loop
very close to the acid–base catalyst. Enzymes lacking
Table 2. 1-SST activity (1-kestose production from 200 mM
sucrose)* of Lp1SST, Lp6
G
-FFT ⁄ 1-FFT and the mutants derived
from Lp6
G
-FFT ⁄ 1-FFT. The table was reproduced from that shown
in a previous publication [38].
Enzyme
Activity
(nkat mg
)1
protein)
Lp6
G
-FFT ⁄ 1-FFT 5.38 ± 0.66
Lp1-SST 32.22 ± 0.69
N340D 1.71 ± 0.13
W343R 2.04 ± 0.67

S415N 3.32 ± 0.31
N340D+W343R 15.17 ± 0.23
N340D+S415N 2.69 ± 0.16
W343R+S415N 2.28 ± 0.04
N340D+W343R+S415N 21.17 ± 0.17
Fig. 11. Multiple alignment of a selection of plant GH32 VI ⁄ FT
members in the regions surrounding the amino acids under study
(bold). Both S-type and F-type enzymes can be discriminated. For
comparison, structurally characterized representatives of S-type
(AtcwINV1) and F-type (Ci1-FEH IIa) enzymes within the plant
GH32 CWI ⁄ FEH subgroup are also presented (underlined). We
refer the reader to a previous reference [30] for more extended
multiple alignments within the CWI ⁄ FEH subgroup. *The artificial
Cys in barley 1-SST probably represents a PCR mistake because
> 20 expressed sequence tags carrying a Tyr were detected.
Donor and acceptor substrate selectivity in GH32 W. Van den Ende et al.
5796 FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS
this couple have Fru as the preferential donor substrate.
By contrast, all FTs contain altered WMNDPNG and
WGW motifs (Fig. 11), while all hydrolases (invertases,
FEHs) show an intact hypervariable loop within the
WMNDPNG motif and an intact WGW motif
(Fig. 11). In all cases, only a few amino acids in the
vicinity of the active site seem to control the substrate
specificity. Using these simple rules, we are now close to
predicting the functionality of new GH32 plant
enzymes. However, it should be emphasized that more
heterologous expression and further structural and
mutagenesis studies are required to validate the hypoth-
eses described above. Without any doubt, there remain

many unanswered questions, such as which amino acid
residues are critical to explain the formation of b(2-1)
versus b(2-6) linkages in fructans. 3D structures of the
S- and F-types of plant FTs are now being created to
resolve this intriguing question.
References
1 Pollock CJ & Cairns AJ (1991) Fructan metabolism in
grasses and cereals. Annu Rev Plant Physiol Plant Mol
Biol 42, 77–101.
2 Hendry GAF (1993) Evolutionary origins and natural
functions of fructans – a climatological, biogeographic
and mechanistic appraisal. New Phytol 123 , 3–14.
3 Hisano H, Kanazawa A, Kawakami A, Yoshida M,
Shimamoto Y & Yamada T (2004) Transgenic perennial
ryegrass plants expressing wheat fructosyltransferase
genes accumulate increased amounts of fructan and
acquire increased tolerance on a cellular level to freez-
ing. Plant Sci 167, 861–868.
4 Hisano H, Kanazawa A, Yoshida M, Humphreys MO,
Iizuka M, Kitamura K & Yamada T (2008) Coordi-
nated expression of functionally diverse fructosyltrans-
ferase genes is associated with fructan accumulation in
response to low temperature in perennial ryegrass.
New Phytol 178, 727–740.
5 Hincha DK, Livingston DP, Premakumar R, Zuther E,
Obel N, Cacela C & Heyer AG (2007) Fructans from
oat and rye: composition and effects on membrane
stability during drying. Biochim Biophys Acta 1768,
1611–1619.
6 Van den Ende W & Valluru R (2009) Sucrose, sucrosyl

oligosaccharides, and oxidative stress: scavenging and
salvaging? J Exp Bot 60, 9–18.
7 Parvanova D, Ivanov S, Konstantinova T, Karanov E,
Atanassov A, Tsvetkov T, Alexieva V & Djilianov D
(2004) Transgenic tobacco plants accumulating
osmolytes show reduced oxidative damage under
freezing stress. Plant Physiol Biochem 42 , 57–63.
8 Li HJ, Yang AF, Zhang XC, Gao F & Zhang JR
(2007) Improving freezing tolerance of transgenic
tobacco expressing sucrose: sucrose 1-fructosyltransfer-
ase gene from Lactuca sativa . Plant Cell Tissue Organ
Cult 89, 37–48.
9 Kawakami A, Sato Y & Yoshida M (2008) Genetic
engineering of rice capable of synthesizing fructans and
enhancing chilling tolerance. J Exp Bot 59, 793–802.
10 Yoshida M, Kawakami A & Van den Ende W (2007)
Graminan metabolism in cereals: wheat as a model
system. In Recent Advances in Fructooligosaccharides
Research (Shiomi N, Benkeblia N & Onodera S, eds),
pp. 201–212. Research Signpost, Kerala, India.
11 Xue GP, McIntyre CL, Jenkins CLD, Glassop D, van
Herwaarden AF & Shorter R (2008) Molecular dissec-
tion of variation in carbohydrate metabolism related to
water-soluble carbohydrate accumulation in stems of
wheat. Plant Physiol 146, 441–454.
12 Zhang J, Huang S, Fosu-Nyarko J, Dell B, McNeil M,
Waters I, Moolhuijzen P, Conocono E & Appels R
(2008) The genome structure of the 1-FEH genes in
wheat (Triticum aestivum L.): new markers to track
stem carbohydrates and grain filling QTLs in breeding.

Mol Breeding 22, 339–351.
13 Prud’homme MP, Morvan-Bertrand A, Lasseur B,
Lothier J, Meuriot F, Decau ML & Noiraud-Romy N
(2007) Lolium perenne, backbone of sustainable develop-
ment, source of fructans for grazing animals and poten-
tial source of novel enzymes. In Recent Advances in
Fructooligosaccharides Research (Shiomi N, Benkeblia
N & Onodera S, eds), pp. 231–258. Research Signpost,
Kerala, India.
14 Lewis DH (1993) Nomenclature and diagrammatic
representation of oligomeric fructans – a paper for
discussion. New Phytol 124, 583–594.
15 Lammens W, Le Roy K, Schroeven L, Van Laere A,
Rabijns A & Van den Ende W (2009) Structural
insights into glycoside hydrolase family 32 and 68
enzymes: functional implications. J Exp Bot 60,
727–740.
16 Tamura K, Kawakami A, Sanada Y, Tase K, Komatsu
T & Yoshida M (2009) Cloning and functional analysis
of a fructosyltransferase cDNA for synthesis of highly
polymerized levans in timothy (Phleum pratense L.).
J Exp Bot
60, 893–905.
17 De Coninck B, Van den Ende W & Le Roy K (2007)
Fructan exohydrolases (FEHs) in plants: properties,
occurrence and 3-D structure. In Recent Advances in
Fructooligosaccharides Research (Shiomi N, Benkeblia
N & Onodera S, eds), pp. 157–179. Research Signpost,
Kerala, India.
18 Henrissat B (1991) A classification of glycosyl hydro-

lases based on amino-acid-sequence similarities.
Biochem J 280, 309–316.
19 Reddy A & Maley F (1996) Studies on identifying the
catalytic role of Glu-204 in the active site of yeast
invertase. J Biol Chem 271, 13953–13958.
W. Van den Ende et al. Donor and acceptor substrate selectivity in GH32
FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS 5797
20 Alberto F, Bignon C, Sulzenbacher G, Henrissat B &
Czjzek M (2004) The three dimensional structure of
invertase (beta-fructosidase) from Thermotoga maritima
reveals a bimodular arrangement and an evolutionary
relationship between retaining and inverting glyco-
sidases. J Biol Chem 279, 18903–18910.
21 Le Roy K, Lammens W, Verhaest M, De Coninck B,
Rabijns A, Van Laere A & Van den Ende W (2007)
Unraveling the difference between invertases and fructan
exohydrolases: a single amino acid (Asp239) substitution
transforms Arabidopsis cell-wall invertase 1 into a
fructan 1-exohydrolase. Plant Physiol 145, 616–625.
22 Davies GJ, Wilson KS & Henrissat B (1997) Nomencla-
ture for sugar-binding subsites in glycosyl hydrolases.
Biochem J 321, 557–559.
23 Verhaest M, Van den Ende W, Le Roy K, De Ranter
CA, Van Laere A & Rabijns A (2005) X-Ray Diffrac-
tion Structure of a plant Glycosyl Hydrolase family 32
protein: Fructan 1-Exohydrolase IIa of Cichorium inty-
bus. Plant J 41, 400–411.
24 Verhaest M, Lammens W, Le Roy K, De Coninck B,
De Ranter CJ, Van Laere A, Van den Ende W & Rabi-
jns A (2006) X-ray diffraction structure of a cell-wall

invertase from Arabidopsis thaliana. Acta Crystallogr D
62, 1555–1563.
25 Verhaest M, Lammens W, Le Roy K, De Ranter CJ,
Van Laere A, Rabijns A & Van den Ende W (2007)
Insights into the fine architecture of the active site of
chicory fructan 1-exohydrolase: 1-kestose as substrate
vs sucrose as inhibitor. New Phytol 174, 90–100.
26 Lammens W, Le Roy K, Van Laere A, Rabijns A &
Van den Ende W (2008) Crystal structures of Arabidop-
sis thaliana cell-wall invertase mutants in complex with
sucrose. J Mol Biol 377, 378–385.
27 De Coninck B, Le Roy K, Francis I, Clerens S, Vergau-
wen R, Halliday AM, Smith SM, Van Laere A & Van
den Ende W (2005) Arabidopsis AtcwINV3 and 6 are
not invertases but are fructan exohydrolases (FEHs)
with different substrate specificities. Plant Cell Environ
28, 432–443.
28 Van den Ende W, De Coninck B, Clerens S, Vergauwen
R & Van Laere A (2003) Unexpected presence of
fructan 6-exohydrolases (6-FEHs) in non-fructan plants:
characterization, cloning, mass mapping and functional
analysis of a novel ‘cell-wall invertase-like’ specific
6-FEH from sugar beet (Beta vulgaris L.). Plant J 36,
697–710.
29 Meng GY & Fu
¨
tterer K (2003) Structural framework of
fructosyl transfer in Bacillus subtilis levansucrase. Nat
Struct Biol 10, 935–941.
30 Le Roy K, Lammens W, Van Laere A & Van den Ende

W (2008) Influencing the binding configuration of
sucrose in the active sites of chicory fructan 1-exohy-
drolase and sugar beet fructan 6-exohydrolase. New
Phytol 178, 572–580.
31 Ritsema T, Hernandez L, Verhaar A, Altenbach D,
Boller T, Wiemken A & Smeekens S (2006) Developing
fructan-synthesizing capability in a plant invertase via
mutations in the sucrose-binding box. Plant J 48,
228–237.
32 Kawakami A & Yoshida M (2002) Molecular character-
ization of sucrose:sucrose 1-fructosyltransferase and
sucrose:fructan 6-fructosyltransferase associated with
fructan accumulation in winter wheat during cold
hardening. Biosci Biotechnol Biochem 66, 2297–2305.
33 Kawakami A & Yoshida M (2005) Fructan:fructan
1-fructosyltransferase, a key enzyme for the biosynthesis
of graminan oligomers in hardened wheat. Planta 233,
90–104.
34 Schroeven L, Lammens W, Van Laere A & Van den
Ende W (2008) Transforming wheat vacuolar invertase
into a high affinity sucrose:sucrose 1-fructosyltransfer-
ase. New Phytol 180, 822–831.
35 Altenbach D, Rudino-Pinera E, Olvera C, Boller T,
Wiemken A and Ritsema T (2009) An acceptor-
substrate binding site determining glycosyl transfer
emerges from mutant analysis of a plant vacuolar
invertase and a fructosyltransferase. Plant Mol Biol
69, 47–56.
36 Chalmers J, Johnson X, Lidgett A & Spangenberg G
(2003) Isolation and characterisation of a sucrose:

sucrose 1-fructosyltransferase gene from perennial
ryegrass (Lolium perenne). J Plant Physiol 160,
1385–1391.
37 Lasseur B, Lothier J, Djoumad A, De Coninck B,
Smeekens S, Van Laere A, Morvan-Bertrand A, Van
den Ende W & Prud’homme MP (2006) Molecular and
functional characterization of a cDNA encoding
fructan:fructan 6G-fructosyltransferase (6G-FFT) ⁄
fructan:fructan 1-fructosyltransferase (1-FFT) from
perennial ryegrass (Lolium perenne L.). J Exp Bot 57,
3961–3961.
38 Lasseur B, Schroeven L, Lammens W, Le Roy K,
Spangenberg G, Manduzio H, Vergauwen R, Lothier J,
Prud’homme MP & Van den Ende W (2009)
Transforming a fructan:fructan 6G-fructosyltransferase
from perennial ryegrass into a sucrose:sucrose
1-fructosyltransferase. Plant Physiol 149, 327–339.
39 DeLano WL (2002) The Pymol Molecular Graphics
System. DeLano Scientific, San Carlos, CA.
Donor and acceptor substrate selectivity in GH32 W. Van den Ende et al.
5798 FEBS Journal 276 (2009) 5788–5798 ª 2009 The Authors Journal compilation ª 2009 FEBS