(Received for publication, January 14, 1997, and in revised form, March 27, 1997)
From the Biochemical Research Laboratory, Ezaki Glico Co., Ltd., Utajima 4-6-5, Nishiyodogawa-ku, Osaka 555, Japan
The initial action of cyclodextrin
glucanotransferase (CGTase, EC 2.4.1.19) from an alkalophilic
Bacillus sp. A2-5a on amylose was investigated. Synthetic
amylose was incubated with purified CGTase then terminated in the very
early stage of the enzyme reaction. When the reaction mixture was
treated with glucoamylase and the resulting glucoamylase-resistant
glucans were analyzed with high performance anion exchange
chromatography, cyclic -1,4-glucans, with degree of polymerization
ranging from 9 to more than 60, in addition to well known
-,
-,
and
-cyclodextrin (CD), were detected. The time-course analysis
revealed that larger cyclic
-1,4-glucans were preferentially
produced in the initial stage of the cyclization reaction and were
subsequently converted into smaller cyclic
-1,4-glucans and into the
final major product,
-CD. CGTase from Bacillus macerans
also produced large cyclic
-1,4-glucans except that the final major
product was
-CD. Based on these results, a new model for the action
of CGTase on amylose was proposed, which may contradict the widely held
view of the cyclization reaction of CGTase.
Cyclodextrin glucanotransferase
(CGTase,1 EC 2.4.1.19), found in several
bacterial species, catalyzes the inter- and intramolecular transglycosylation of -1,4-glucan. Such activity of CGTase on inter-
and intramolecular transglycosylation of
-1,4-glucan is called the
disproportionation reaction and the cyclization reaction, respectively.
It is also known that CGTase catalyzes the transglycosidic linearization (coupling reaction) of cyclic
-1,4-glucan in the presence of a suitable acceptor molecule to produce linear
-1,4-glucan. The cyclization reaction of CGTase has been of great
interest since this is the only enzyme that can produce
-,
- and
-cyclodextrin (CD), which are generally known as the cyclic
-1,4-glucan with DP of 6, 7, or 8. These CDs all have a hydrophobic
central cavity, incorporate various inorganic or organic compounds, and
form inclusion complexes (1). Therefore, these CDs are widely used in
the pharmaceutical, food, agricultural, and cosmetic industries (2).
Extensive analyses on various CGTases indicated that all CGTases
convert amylose or amylopectin into a mixture of -,
-, and
-CD
and remaining dextrins; differences, however, were found in their
product specificities (
-,
-, and
-CD ratios). Thus, CGTase is
sometimes classified into three types (
-,
-, and
-CGTase), depending on the major CD produced. Since
-,
-, and
-CD all have a dimensionally distinct central cavity and different specificity for guest molecules, recent studies on CGTase have focused on trying to
understand the mechanism of the cyclization reaction and to find or
engineer a CGTase that produces a specific type of CD.
Several approaches have been carried out to obtain the structural
explanation of the cyclization reaction of CGTase. Analyses of the
three-dimensional structure of CGTase have been carried out using
several types of CGTases (3-6). Additionally, the structures of
CGTases with substrates (5, 7-9) and with inhibitor molecules (9, 10)
were also analyzed. From these studies, models of CGTase activity
cleaving the target -1,4-linkage were proposed. In the case of
CGTase from Bacillus circulans strain 251, three active site
residues, Asp-229, Glu-257, and Asp-328, which are conserved in all
CGTase primary sequences, play important roles for this step of
reaction (8, 10). A similar catalytic mechanism has been reported from
structural studies on amylases (11-14) that hydrolyze
-1,4-glucan.
Several amino acid residues involved in substrate binding or in the
determination of product specificity have also been proposed by
three-dimensional structure analysis (5, 6, 8, 9) or by protein
engineering approaches (15-17). However, it is less clearly understood
how CGTase catalyzes the following intramolecular transfer reaction to
produce cyclodextrins.
Although -,
-, and
-CD are the major products of CGTase, it
has been known that trace amounts of larger cyclic glucans (
-,
-,
-,
-, and
-CD) were also present in the reaction mixture of
CGTase on starch (18, 19). The structures of these larger cyclic
glucans are still not well understood, because they seem to be a
mixture of cyclic
-1,4-glucans, outer-branched cyclic
-1,4-glucans, and inner-branched cyclic glucans (18). Kobayashi and
Yamasaki (20) carried out further structural analyses on putative
-,
-,
-,
-, and
-CD fractions and reported that the
-CD
fraction contained a large amount of unbranched cyclic
-1,4-glucan
with DP 9. However, the proportion of unbranched cyclic
-1,4-glucan
in the following fractions decreased dramatically (50% in
-CD, 25%
in
-CD, and almost 0% in
- and
-CD fractions). From this
study, it is thought that the presence of cyclic
-1,4-glucan with DP
larger than 12 in the CGTase reaction products is unlikely. Recently,
however, we found that potato D-enzyme (disproportionating enzyme or
4-
-glucanotransferase, EC 2.4.1.25) catalyzes an intramolecular
transglycosylation reaction on amylose to produce cyclic
-1,4
glucans with DP range from 17 to several hundred (21). The time-course
analysis of D-enzyme action on amylose revealed that large cyclic
-1,4-glucans were preferentially produced in the initial stage of
the cyclization reaction, and subsequently converted into small cyclic
-1,4-glucans, although
-,
-, and
-CD were never produced
(21). D-enzyme also catalyzes disproportionating reactions on
malto-oligosaccharides (22) and transglycosidic linearization of cyclic
-1,4-glucans in the presence of a suitable acceptor (21). In all
these respects, D-enzyme and CGTase both seem to catalyze the same
reaction, with the major difference being the DP of the cyclic
-1,4-glucan produced. Furthermore, we also reported that the
glycogen-branching enzyme (EC 2.4.1.18) from Bacillus
stearothermophilus also catalyzes the intramolecular transglycosylation of amylose and amylopectin to produce branched cyclic glucans with DP larger than 18 (23, 24). From these studies on
glycosyltransferases other than CGTase, we are interested in the action
of CGTase in producing cyclic glucans larger than CDs. In this paper,
we investigated the initial action of CGTase from an alkalophilic
Bacillus sp. A2-5a (25) on synthetic amylose and found that
the CGTase also produced cyclic
-1,4-glucans with DP ranging from 9 to more than 60, in addition to
-,
-, and
-CD, from synthetic
amylose in the very early stage of the reaction.
Synthetic amylose with an average molecular mass of 30 kDa (amylose AS-30) and soluble starch were purchased from Nakano Vinegar Co., Ltd. (Aichi, Japan) and E. Merck AG (Darmstadt, Germany), respectively. Glucoamylase from Rhizopus sp. was purchased from Toyobo Co., Ltd. (Osaka, Japan). CGTase from an alkalophilic Bacillus sp. A2-5a was purified to a homogeneous state (25). CGTases from B. macerans was purchased from Amano Pharmaceutical Co., Ltd. (Aichi, Japan) and was used without further purification. The activity of CGTase was assayed using soluble starch as the substrate by measuring the decrease in iodine-staining power as described previously (25).
Preparation of Amylose SolutionAmylose AS-30 (8 mg) was dissolved in 200 µl of 1 N NaOH solution then neutralized by adding 200 µl of 1 M sodium acetate buffer (pH 5.5), 200 µl of 1 N HCl, and 400 µl of distilled water. The solution was used immediately after neutralization.
Analysis of Reaction Products of CGTase on AmyloseCGTase
(0.75 unit/ml) was incubated at 40 °C with amylose AS-30 (0.4%
(w/v)) in 0.2 M sodium acetate buffer, pH 5.5, and
reactions were terminated by boiling the solutions for 10 min. The
reaction mixture containing 20 µg of glucan was incubated with
glucoamylase (0.2 units) in 20 mM sodium acetate buffer (pH
5.5) for 16 h at 40 °C and then boiled for 5 min. The products
in the reaction mixture were determined with high performance anion
exchange chromatography (HPAEC, see below). The amounts of -,
-,
and
-CD were measured with high performance liquid chromatography
(HPLC, see below). The amount of glucoamylase-resistant molecules was
calculated by subtracting the amount of glucose released by
glucoamylase from that of total glucan in the reaction mixture. The
amount of glucose was measured by the glucose oxidase method (26).
HPAEC was carried out based on the DX-300 system
(Dionex) with a pulsed amperometric detector (model PAD-II, Dionex)
using a Carbopac PA-100 column (4 mm × 250 mm). A sample (25 µl) containing 40 µg of glucan was injected and eluted with a
gradient of sodium acetate (0-2 min, 50 mM; 2-37 min,
increasing from 50 mM to 350 mM with the
installed gradient program 3; 37-45 min, increasing from 350 mM to 850 mM with the installed gradient
program 7; 45-47 min, 850 mM) in 150 mM NaOH
with a flow rate of 1 ml min1.
HPLC was carried out based on the DX-300 system
(Dionex) using an Aminex HPX-42A column. To remove glucose from the
reaction mixtures, a sample (50 µl) containing 80 µg of glucan was
charged on a Waters Sep-Pak C18 cartridge (Millipore),
washed with 10 ml of H2O and eluted with 1.5 ml of 50%
methanol. The eluate was dried up in vacuo, and dissolved in
50 µl of water. The sample was then injected and eluted with water
with a flow rate of 0.6 ml min1 at 80 °C. The eluate
from the column was mixed with 0.3 M LiOH using an anion
micromembrane suppressor (model AMMS-II, Dionex), after which the
carbohydrate in the eluate was detected with a pulsed amperometric
detector (mode PAD-II, Dionex).
A reaction mixture (5 ml) containing 10 mg of amylose AS-30 and CGTase (0.7 unit) was incubated in 0.2 M sodium acetate buffer, pH 5.5, at 40 °C for 1 h, and then boiled for 10 min. The reaction mixture was incubated with 10 units of glucoamylase at 40 °C for 16 h, and then boiled for 5 min. After removing glucose with the Waters Sep-Pak C18 cartridge, the molecular masses of glucoamylase-resistant glucans were determined with a Kompact Maldi I TOF-MS system (Shimadzu, Kyoto, Japan).
Synthetic amylose AS-30 was incubated with CGTase from
alkalophilic Bacillus sp. A2-5a. The enzyme reaction was
terminated at the early stage of the reaction (10 min), and then the
reaction mixture was incubated with glucoamylase to digest the linear
amylose into glucose. When the glucoamylase-resistant molecules thus
obtained were analyzed with HPAEC, many peaks were detected (Fig.
1A). Note that these peaks were not found in
a control experiment with heat-inactivated CGTase (result not shown).
Most of these peaks were eluted in the region where cyclic
-1,4-glucans with DPs of over 17, produced by potato D-enzyme on
synthetic amylose (Fig. 1F), were eluted. This result
suggests that CGTase produced such cyclic
-1,4-glucans. The large
peak around 42 min may indicate the presence of glucoamylase-resistant
molecules with DP more than 60, since these molecules were not resolved
in this HPAEC condition and eluted together.
Then to prove the cyclic structure of glucoamylase-resistant molecules
produced by CGTase, their molecular masses were determined with TOF-MS
(Fig. 2). A non-cyclic glucan with DP of n
has a molecular mass of 162.1436n + 18.01534 Da, whereas a
cyclic glucan should have a molecular mass of 162.1436n Da.
The molecular mass of each glucoamylase-resistant molecule was compared
with the theoretical value for non-cyclic and cyclic glucans (Table
I). As shown in Table I, experimental values of
glucoamylase-resistant molecules were consistent with theoretical
values for cyclic glucan but not with those for non-cyclic glucan.
|
To confirm that the glucoamylase-resistant molecules produced by CGTase
are -1,4-glucans, the structure of these glucans were further
examined by treatment with several enzymes.
-Amylase from
Bacillus subtilis, an endo-type amylase, completely degraded these molecules to glucose and maltose, and isoamylase and pullulanase, which degrade
-1,6-linkage of glucans, did not (data not shown). These results indicate that the glucoamylase-resistant molecules produced by CGTase were
-1,4-glucans.
Based on all the results mentioned above, we concluded that the
molecules produced by the CGTase reaction on amylose, as shown in Fig.
1A, were cyclic -1,4-glucans with DPs ranging from 6 to
more than 60.
It is known that
the glucoamylase-resistant products of the CGTase reaction on amylose,
after a prolonged reaction time, are -,
-,
-CD and negligible
amounts of other glucoamylase-resistant glucans. Our result apparently
contradicts this widely held view since the major cyclic
-1,4-glucans produced in the initial stage of the CGTase reaction
were not
-,
-, and
-CD but were those with high DPs. To
investigate how large cyclic
-1,4-glucans, which were found in the
initial stage of CGTase reaction, were replaced by
-,
-, and
-CD, the time course of the reaction of CGTase were monitored.
Amylose AS-30 was incubated with CGTase from alkalophilic
Bacillus sp. A2-5a for up to 360 min. Each sample was
treated with glucoamylase and was analyzed by HPAEC (Fig. 1). As shown
in Fig. 1 (A-E), cyclic
-1,4-glucans with DPs of over 60 were most prevalent after a reaction time of 10 min (Fig. 1A). However, with prolonged reaction, cyclic
-1,4-glucans with high DPs gradually decreased, and those with low
DPs increased. At the end of the reaction, the main product of CGTase
was
-CD (Fig. 1E).
Fig. 3 shows the time course of the amount of cyclic
-1,4-glucans. The amount of total cyclic
-1,4-glucans increased
rapidly from the beginning of the reaction, and their yield reached
about 80% at 120 min. However, the amounts of
-,
-, and
-CD
increased at a more slower rate than those of total cyclic
-1,4-glucans. As a result, cyclic
-1,4-glucans, apart from
-,
-, and
-CD, increased at the early stage of the reaction and
reached a yield of about 52% at 50 min, but decreased thereafter.
Analysis of the Initial Action of Other CGTases
CGTases found
in many bacterial species are classified into three types, -,
-,
or
-CGTase, depending on the major product of the cyclization
reaction. The CGTase employed above is
-CGTase because it mainly
produces
-CD (25). To examine whether the production of a large
cyclic
-1,4-glucan is the specific feature only found in this CGTase
or is the common feature for others, similar experiments were carried
out by using CGTases from B. macerans, which is classified
as
-CGTase. This enzyme also produced large cyclic
-1,4-glucans
in the initial stage of reaction (Fig. 4, A
and B), which were subsequently converted into small cyclic
-1,4-glucans. However, the final major cyclic product was
-CD (Fig. 4E).
It is generally believed that the cyclization reaction of
CGTase on amylose is an exo-type attack (2), where the enzyme recognizes the 6-8 glucose units from the non-reducing end, attacks the adjacent -1,4-linkage, and transfers it to the C-4 position of
the non-reducing end to produce
-,
-, or
-CD (Fig.
5A). This view was only confirmed from the
analysis of CGTase action on 14C-labeled linear
-1,4-glucans with DP 7-12 (27), but not investigated in high
molecular weight glucans. If this view can be applied to the CGTase
action on high molecular weight glucans, cyclic products throughout the
reaction on amylose are expected to be only
-,
-, and
-CD.
However, the results presented in this paper clearly demonstrate that
the cyclic glucans produced in the initial stage of cyclization
reaction of CGTase are not only
-,
-, and
-CD, but are cyclic
-1,4-glucans with various DP ranging from 6 to more than 60. Large
cyclic
-1,4-glucans were preferentially produced in the initial
stage of cyclization reaction, which were subsequently converted into
small cyclic
-1,4-glucans and into the final major products,
-CD
or
-CD. Thus these findings apparently contradict the widely held
view of the action model of CGTase, and so we propose a new model for
the action of CGTase as shown in Fig. 5B. CGTase probably
attacks any
-1,4-linkage within the amylose molecule, and then
transfers the newly formed reducing end of the substrate either to the
non-reducing end of a separate linear acceptor molecule or glucose (the
intermolecular transglycosylation or disproportionation reaction), or
to its own non-reducing end (the intramolecular transglycosylation or cyclization reaction, Fig. 5B). This random cyclization
reaction produces wide ranges of cyclic
-1,4-glucans with DP 6 to
more than 60. The reversibility of these reactions allows large cyclic molecules to be linearized again by transglycosylation, and smaller cyclic molecules to be subsequently produced. The equilibrium of the
whole reaction tends toward the formation of
- or
-CD as the
final major products.
Both CGTase and D-enzyme catalyze the cyclization and
disproportionation of -1,4-glucan and transglycosidic linearization of cyclic
-1,4-glucan in the presence of a suitable acceptor molecule. During the cyclization reaction, large cyclic
-1,4-glucans were preferentially produced in the initial stage, but were
subsequently converted into smaller cyclic
-1,4-glucans in both
cases. Thus both enzymes seem to catalyze the same reaction, with the
major difference being in the smallest size of the cyclic
-1,4-glucans produced. The DP of the smallest cyclic
-1,4-glucan
produced by CGTase is 6. On the other hand, D-enzyme never produced
-,
-, and
-CD and the smallest cyclic
-1,4-glucan has DP of
17. It is very interesting to know how the specificities of these products are determined differently between D-enzyme and CGTase. In our
previous paper (23), we discussed that cyclic
-1,4-glucan with DP
6-8 (
-,
-, and
-CD) and those with DP more than 17 may have
different structures. The structures of cyclic
-1,4-glucan with DP
6-8 (
-,
-, and
-CD) may fit well to the active site of
CGTase, and those with DP more than 17 may fit well to D-enzyme. However, this idea seems to be unlikely because it is now understood that CGTase can also produce cyclic
-1,4-glucans with DP more than
17. One possible explanation for the different product specificity is
as follows. The equilibrium of the reaction catalyzed by both enzymes
tends toward the formation of smaller cyclic
-1,4-glucans; however,
the smallest cyclic
-1,4-glucan molecule to be produced by D-enzyme
has DP of 17, whereas that of CGTase is 6. At the moment, we do not
know the mechanism to determine the smallest cyclic product of each
enzyme, although we speculate that the difference in the smallest
products of each enzyme is attributable to the active site structure.
Despite the similarity found in their activities, D-enzyme and CGTase
show no similarity in their primary sequences (22). The tertiary
structure of CGTase has already been obtained by x-ray crystallographic
studies (3-6). Determination of the structure of D-enzyme and its
comparison with CGTase will be necessary to answer this question.
We especially thank Shimadzu Co. for TOF-MS analyses.