JOURNAL OF BIOSCIENCE AND BIOENGINEERING
Vol. 100, No. 4, 380–390. 2005
DOI: 10.1263/jbb.100.380
© 2005, The Society for Biotechnology, Japan
Two Novel Cyclodextrin-Degrading Enzymes Isolated from
Thermophilic Bacteria Have Similar Domain Structures
but Differ in Oligomeric State and Activity Profile
Pernilla Turner,1* Antje Labes,2 Olafur H. Fridjonsson,3 Gudmundur O. Hreggvidson,3
Peter Schönheit,2 Jakob K. Kristjansson,3 Olle Holst,1 and Eva Nordberg Karlsson1
Department of Biotechnology, Lund University, P.O. Box 124, SE-221 00 Lund, Sweden,1 Institute for General
Microbiology, Christian-Albrechts University Kiel, Am Botanischen Garten 1-9, D-24118 Kiel,
Germany,2 and Prokaria Ltd., Gylfaflöt 5, IS-112 Reykjavik, Iceland 3
Received 12 January 2005/Accepted 17 May 2005
In this paper, we present the expression and characterization of two novel enzymes from the
α-amylase family exhibiting cyclomaltodextrinase specificity. The nucleotide sequences encoding
the enzymes were isolated from the genomic DNA of two thermophilic bacterial strains originating from Icelandic hot springs and belonging to the genera Anoxybacillus (AfCda13) and
Laceyella (LsCda13). The genes were amplified using a consensus primer strategy utilizing two of
the four conserved regions present in glycoside hydrolase family 13. No identifiable signal peptides were present in open reading frames encoding the enzymes, indicating an intracellular location of both enzymes, and their physiological function to be intracellular cyclodextrin degradation. The domain structures of both enzymes were also similar, including an N-terminal domain,
the catalytic module composed of the A- and B-domains, and a C-terminal domain. Despite the
similarity in domain composition, the two enzymes displayed differences in the oligomeric state
with AfCda13 being a dimeric protein, whereas LsCda13 was monomeric. The two enzymes also
displayed significantly different activity profiles, despite being active on the same range of substrates. It was shown that the enzyme displaying the highest activity on cyclodextrin was dimeric
(AfCda13). Moreover, a fraction of the dimeric enzyme could be converted to a monomeric state
in the presence of KCl and this fraction retained only 23% of its activity on α-cyclodextrin while
its activity on starch was not significantly affected, indicating that the oligomeric state is an important factor for a high activity on cyclodextrin substrates.
[Key words: cyclomaltodextrinase, glycoside hydrolase family 13, Anoxybacillus, Laceyella]
The enzymes involved in the degradation and conversion
of starch are often termed the α-amylase family. These
structurally and functionally related enzymes have been
classified into three families (GH13, 70 and 77) (1) in the
sequence/structure-based glycoside hydrolase family system (2). Glycoside hydrolase family 13 (GH13) includes the
majority of the enzymes, such as α-amylases, which are
conventionally used in various industrial applications, e.g.,
in starch conversion applications (3, 4). Enzymes with a
spectrum of other starch-modifying and hydrolyzing activities also belong to the family, including glycogenase, pullulanase, 1,6-glucosidase, branching and debranching enzymes,
maltogenic amylase (MAase), neopullulanase (NPase) and
cyclomaltodextrinase (CDase) (1, 5). Despite the multitude
of activities, four regions in the primary structure close to
the active site of the enzymes are conserved within the family (5, 6). The conserved regions were utilized in a search
for novel enzymes using two Icelandic thermophilic bacterial strains as the target organisms. Using the CODEHOP
consensus primer strategy (7) the genes encoding two
cyclomaltodextrin-degrading enzymes were isolated. The
enzymes have a domain organization typical of CDases
(EC 3.2.1.54), MAases (EC 3.2.1.133), and NPases (EC
3.2.1.135), which despite having different EC numbers are
found to share a similar domain organization, and to hydrolyze cyclodextrins very efficiently (8).
The novel enzymes isolated in this work have been expressed in Escherichia coli and characterized, and on the
basis of our results, we recommend to classify them as
CDases. The reasons for this and the interesting differences
in activity profile found between the two enzymes will be
discussed here.
* Corresponding author. e-mail: [email protected]
phone: +46-(0)46-222-4626 fax: +46-(0)46-222-4713
The costs of publication of this article were supported in part by
Grants-in-Aid for Publication of Scientific Research Results from the
Japan Society for the Promotion of Science (JSPS) (no. 173048).
380
VOL. 100, 2005
TWO CDases DIFFERING IN OLIGOMERIC STATE AND ACTIVITY PROFILE
MATERIALS AND METHODS
Amplification and analysis of nucleotide sequences encoding
the enzymes
For the primer construction, the amino acid sequences of various amylolytic enzymes were retrieved from protein sequence databases and aligned using CLUSTALX ver. 1.8
(9). Primers were designed according to the CODEHOP strategy
(7). Forward and reverse primers were constructed, aimed to complement the DNA encoding the sequences of two of the conserved
regions (A [or I] and B [or II]) in the glycoside hydrolase family
13 (6) as primer target sites for the first amplification. The primers were C-F, corresponding to region A, and C-R, to region B
(Table 1). The template for the amplification of each of the genes
was genomic DNA from two bacterial strains isolated from Icelandic hot springs, which were found to be closely related to Anoxybacillus flavithermus (Af) (previously Bacillus flavothermus) and
Laceyella sacchari (Ls) (previously Thermoactinomyces sacchari)
by 16S rRNA sequence analysis (Hreggvidsson, G.O., unpublished
results). PCR amplification was carried out with DyNAzyme DNA
polymerase (Finnzymes, Espoo, Finland) using a MJ Research
thermal cycler PTC-0225. The reaction mixture was first denatured
at 95°C for 5 min, followed by 30 cycles of denaturing at 95°C for
50 s, annealing at 52°C for 50 s, extension at 72°C for 3 min, and
a final extension for 7 min at 72°C to obtain A-overhangs. PCR
products were separated on gels and purified using GFX spin columns (Amersham Biosciences, Piscataway, NJ, USA).
The gene fragments were cloned into conventional pUC-based
sequencing vectors (10) by the TA-cloning method (11). Inserts
were sequenced with M13 forward and reverse primers on ABI
3700 DNA sequencers using a BigDye terminator cycle sequencing ready reaction kit according to the instructions of the manufacturer (PE Applied Biosystems, Foster City, CA, USA). Following
the sequencing of the obtained target gene fragments, their upstream and downstream flanking regions were amplified from the
corresponding genomic DNA in a series of nested PCR amplifiTABLE 1. Primers for PCR amplification
Primer
Nucleotide sequence
designation
5′-GCATGTTATGCTGGATGCAGTNTTYAAYCA
C-F
C-R
5′-GATCAACTTAATTAGCAACATCCATTCKCCANCCRTC
1-F
5′-TATTATTACATATGCTTATTGAAGCACTTTTTC
1-R
5′-TATTATTACTCGAGAACCAGCACT ATTCCCTG
2-F
5′-ATGCTGTTCCATGGATATGTTAAAGGAAGCTGTA
2-R
5′-ATTATTTATTGCGGCCGCTTCAACTAAGAACATAAG
Af1
5′-GAAACTCGTGAATATGGAACC
Af2
5′-GCCTTCGTGCCAACTATGCC
Af3
5′-TTTCAGCACATCTTGAAATGGAGCG
Af4
5′-GATGTAGCAACATATTGGATTCGCG
Af5
5′-TATCGTATTTATGATTGGACGATGCT
Af6
5′-GAAAGCGGTGAAACAAGACGT
Af7
5′-AATAAATCCCGTTAATTCCTAGCT
Af8
5′-TTGGCGCGAGTTTCGTCAAGCAG
Af9
5′-CCCAACAATACGGATCACCATGCA
Af11
5′-TTCGCAACCGAATGTGCAATG
Ls1
5′-GCTCGATGAAAAACCAGTCTCTGTA
Ls2
5′-CAGGTCGTCCGCTACTGGACAG
Ls3
5′-CCCTTTTCCAGTACATCGCG
Ls4
5′-GATCGATGGATGGCGAATGG
Ls5
5′-GCCAAACTGGGGATCAATTCTCAGA
Ls6
5′-GAATCTGCTTGATTCCCATGATACC
Ls7
5′-TGTGATTGGATGGGGAGGCAA
Ls8
5′-CTTTTTGACCTCTTGCGGTGAA
Ls9
5′-GGCGTTTTGCAGTGCACTCCAGCA
Ls11
5′-GCTTCAAAATAATCAAACCGCTGG
Restriction sites for cloning are underlined.
381
cations using one gene-specific, 5′-biotin-labeled primer, and one
arbitrary primer targeting the unknown flanking sequence. The
PCR product was purified using QIAquick PCR purification spin
columns (Qiagen, Hilden, Germany) prior to the next PCR amplification with a nested gene-specific primer upstream the previous
gene-specific primer. The PCR product of the latter amplification
was cloned and sequenced as described above and the sequence
information was used to generate new gene-specific primers for
the next nested PCR amplification until the complete genes were
obtained. The primers used are listed in Table 1 (Af1-11 corresponding to AfCda13 and Ls1-11 corresponding to LsCda13).
Even numbered primers are designed for upstream amplification
whereas odd numbered primers are designed for downstream amplification. Similarity searches using BLAST were performed on
the NCBI server (http://www.ncbi.nlm.nih.gov). The ClustalW
tool on the EBI server (http://www.ebi.ac.uk/clustalw) was used to
generate multiple sequence alignments, which were displayed using Gene doc 2.6.02 (12). The sequences of the respective genes
were deposited to the GenBank at NCBI under accession nos.
AY937387 and AY937388.
Cloning and expression in E. coli The complete genes were
PCR-amplified under standard conditions using the Gene Amp
PCR System 9700 (PE Applied Biosystems) with the Expand
High-Fidelity PCR System (Roche Diagnostics, Mannheim, Germany), for insertion into the expression vector pET-22b(+)
(Novagen, Madison, WI, USA) incorporating the C-terminal hexahistidine tag. The gene cda13 from the Laceyella strain was amplified using the primers 1-F and 1-R introducing the restriction
sites NdeI and XhoI, whereas cda13 from the Anoxybacillus strain
was amplified using 2-F and 2-R with the restriction sites NcoI and
NotI (Table 1). The PCR products were purified with a QIAEXII
Gel Extraction kit (Qiagen) after gel separation. Both the PCR
product and the vector were digested with appropriate restriction
enzymes (New England Biolabs, Beverly, MA, USA), and the
vector was treated with bacterial alkaline phosphatase before being ligated to the insert using T4 DNA ligase (Invitrogen Life
Technologies, Frederick, MD, USA). The resulting plasmids were
transformed into E. coli Nova Blue cells (Novagen). Inserts from
positive clones were completely sequenced using the T7 forward
and T7 reverse primers. Those containing correct gene inserts
were transformed into the E. coli expression strain BL21(DE3)
(Novagen). The genes were expressed in cells subjected to shake
flask cultivation at 37°C as described previously (13) or produced
in a 2.5 l bioreactor by substrate-limited fed-batch cultivation according to Ramchuran et al. (14). After 2–3 h of induction, the
cells were harvested by centrifugation at 5000×g and 4°C for
5 min.
Purification
Upon passing through a French pressure cell at
1.3 ×108 Pa, cells were disrupted and cell debris and unbroken cells
were removed by centrifugation for 90 min at 100,000×g at 4°C.
The His-tagged recombinant proteins were purified by immobilized metal ion affinity chromatography (IMAC) as described previously (15). Bound proteins were eluted using a stepwise gradient
of elution buffer (100–500 mM imidazole, 0.75 M NaCl, 20 mM
Tris–HCl, pH 7.5). AfCda13 was purified in two additional steps
including gel filtration on Superdex 200 HighLoad 16/60 (50 mM
Tris–HCl, pH 7.4, containing 150 mM NaCl) and anion exchange
chromatography on UNO Q1 using a NaCl gradient from 0 to 2 M
at pH 5.3.
Total protein analysis
The total protein content in the purified protein sample was estimated by the bicinchoninic acid-copper method (Sigma) or according to Bradford with BSA as a standard. SDS–PAGE according to Laemmli (16) was carried out to
analyze the purity of the enzymes.
Activity analysis Hydrolyzing activity was measured by a
modified 3,5-dinitrosalicylic acid (DNS) assay of reducing sugars
382
J. BIOSCI. BIOENG.,
TURNER ET AL.
(17). To 180 µl of prewarmed substrate in buffer (containing up to
2% starch, pullulan, α- and β-cyclodextrin, amylose, amylopectin,
and glycogen), 5–100 µg of protein was added, and the mixture
was incubated at an appropriate temperature for 15 min. The linearity of time and protein dependence was ensured. The enzymatic
reaction was stopped by adding 300 µl of DNS solution and boiling the samples for 10 min. Tubes were cooled on ice for 15 min
before an aliquote of 200 µl was transferred to a microtitre plate.
Absorbance was read at 550 nm; 1 U corresponds to the release of
1 µmol of reducing sugar equivalents (expressed as glucose) per
minute. The apparent pH optima were measured between pH 4 and
9 at 55°C (sodium acetate, pH 4.0–6.0; potassium phosphate, pH
6.0–8.0; Tris–HCl, pH 8.0–9.0). The temperature dependence of
activity was measured between 25°C and 70°C in potassium phosphate buffer adjusted to the respective pH optimum.
Thin-layer chromatography Product formation following
the enzymatic digestion (at various incubation times and different
enzyme loadings) of α- and β-CDs, starch, amylose, and pullulan,
was examined by thin-layer chromatography (TLC) using an aluminum sheet, 20 ×20 cm silica gel 60 F254 (Merck Eurolabs, Darmstadt, Germany). The hydrolysis pattern of β-CD, starch, and
pullulan was visualized after the separation of products in a running solution consisting of isopropanol, acetone, and H2O at a ratio
of 2: 2:1. The plate was sprayed with Preval Sprayer (85% acetone, 1.7% aniline, and 1.7% diphenylamine) twice and sugar spots
were visualized at 140°C. Alternatively, hydrolysis products were
separated in butanol: ethanol : water (5 :3 : 2) solution, and visualized by dipping in 5% H2SO4 and baking for 15 min at 120°C.
Irreversible thermal inactivation and calcium ion dependence
Thermal inactivation studies were carried out by incubating enzyme samples (0.04 mg ml–1 LsCda13 and 0.01 mg ml–1 AfCda13)
in 50 mM potassium phosphate buffer, pH 6.4, at 60°C. Aliquots
were removed at intervals (0–30 min) and cooled on ice, and residual activity was measured by the activity assay described above.
Additional experiments were performed for 120 min in the same
buffer but at different temperatures (57°C, 60°C and 67°C) to
monitor the effect of calcium on activity and stability. In this case
the enzymes were dialyzed against 0.5 M EDTA in 50 mM piperazine buffer (pH 6.0) overnight, followed by the exchange of buffer
to 50 mM piperazine, pH 6.0, without EDTA prior to thermal inactivation analysis in the absence or presence of calcium chloride.
Gel filtration
Native molecular weight was estimated by gel
filtration using Superdex 200 Highload 16/60 in 50 mM Tris–HCl,
pH 7.4, and 150 mM NaCl at a flow rate of 1 ml min–1 using ribonuclease A (13.7 kDa), chymoptrypsin (25 kDa), ovalbumin (43
kDa), BSA (66 kDa), aldolase (158 kDa) and catalase (232 kDa) as
standards. Purified protein (1–2 mg) was applied. The experiment
was repeated on a Bioprep SE 1000/17 column (Bio-Rad, Hercules, CA) using the same buffer at a flow rate of 0.5 ml min–1.
Separation of subunits of AfCda13
Gel filtration was also
used in determining the oligomeric state after equilibration with
KCl. Protein samples were incubated in 2 M KCl for 32 h to ensure
a complete equilibration and then applied to a Superdex 200 column equilibrated in 50 mM Tris–HCl, pH 7.4, containing 2.0 M
KCl, 150 mM NaCl and 5 mM DTE and eluted with the same
buffer at a flow rate of 1 ml min–1. The results were confirmed on a
Bioprep SE 1000/17 column. Enzyme activity in the presence of
KCl was measured both for the monomeric and dimeric fractions
using α-cyclodextrin and starch by the method described above.
The reassociation of the dimer was confirmed after dialyzing the
monomeric enzyme fraction for 16 h against 50 mM Tris–HCl with
150 mM NaCl and 5 mM DTE, and activity was again determined
as described above.
RESULTS
Primary structure and domain organization Sequence
analysis revealed an open reading frame (ORF) consisting
of 1749 nucleotides (Lscda13) encoding a protein of 582
amino acids in the Laceyella strain. The Anoxybacillus
strain hosted an ORF of 1761 nucleotides (Afcda13) encoding 586 amino acids. Neither of the genes contained a signal
peptide. In both cases the isolated gene encodes the first described glycoside hydrolase from the respective species, and
the two protein sequences encoded by the corresponding
genes were 44% identical to each other. As will be shown
below, the encoded enzymes were classified as CDases, and
they are hence named AfCda13 and LsCda13 hereafter, in
accordance with the classification proposed by Henrissat et
al. (18).
The domain boundaries of the novel enzymes were identified after alignment to homologous proteins identified by a
BLAST search using the deduced amino acid sequence of
the corresponding gene as the query sequence. Four of the
ten highest probability matches found for each sequence
were common for both enzymes. Of the 16 resulting sequences from both searches, 14 originated from microorganisms classified under the same taxonomic order (Bacillales) as the two query sequences. The remaining two were
encoded by genes originating from organisms under the
order Thermoanaerobacteriales, and showed 47% sequence
identity to LsCda13, which was also the second highest
probability match found in the search. LsCda13 was most
similar to TVAII, a 3D-structure-determined NPase from
Thermoactinomyces vulgaris (65% identity), and was shown
to encompass four domains: N, A, B, and C (19, 20). AfCda13
exhibited the highest overall sequence identity (74%) with a
maltogenic amylase from Geobacillus stearothermophilus
(21), and a slightly lower (69%) identity with the 3D-structure-determined NPase from the same microorganism, with
a similar domain composition as TVAII (22). Both enzymes
could hence be predicted to exhibit the domain structure
typical of those of MAases, NPases and CDases, which consist of an N-terminal domain, domain N (proposed to be
involved in determining substrate specificity [1, 22–24]), a
catalytic module composed of two domains (domain A
[(β/α)8-barrel] and domain B protruding from this), and finally domain C (Fig. 1).
Catalytic module
The catalytic module (domains A
and B) is the most conserved. The AfCda13 module was
80% and 77% identical to the MAase and NPase of G.
stearothermophilus, respectively, and the LsCda13 module
showed 73% identity to the catalytic module of TVAII. The
identity (55%) between the catalytic modules of Af and
LsCda13 is lower, but is still the second highest found to
FIG. 1. Schematic overview of domain organization of LsCda13
and AfCda13. The N-terminal domain (N) contains approximately 120
residues, and is followed by the catalytic module (approximately 470
residues), composed of domains A and B (50 residues), and finally the
80-residue C-domain (C).
VOL. 100, 2005
TWO CDases DIFFERING IN OLIGOMERIC STATE AND ACTIVITY PROFILE
date for LsCda13. The four regions typical of GH13 were
conserved in the two novel enzymes, as were the three catalytic residues located in three of the conserved regions, namely, II (D325 LsCda13/D328 AfCda13), III (E354/E357) and
383
IV (D421/D424), of the catalytic module. Additional regions with a high sequence conservation have been defined
for CD-degrading enzymes (8), and were also found in the
Af and LsCda sequences (Fig. 2). These regions are also
(Continued on next page)
384
TURNER ET AL.
J. BIOSCI. BIOENG.,
FIG. 2. Multiple sequence alignment of homologous CDases, MAases and NPases identified by BLAST. The domains are indicated by thick
lines on top of the alignment (grey line, N-domain; black line, A-domain; broken line, B-domain; wave line, C-domain). The Ca2+-binding loop is
indicated by a thin black line. Conserved regions are boxed, and those corresponding to regions I–IV of GH13 are indicated by Roman numbers. A
consensus sequence is shown in bold below the aligned sequences. Completely conserved residues are shown in upper cases, residues conserved in
more than 80% of the sequences are shown in lower cases, positions in which a specific group of residues is conserved are indicated by numbers
(1 = N, D; 3= T, S; 4 = K, R; 5= Y, F, W; 6 = I, L, V, M). The catalytic residues are indicated by a triangle on top of the alignment. An abbreviated
name of the organism, enzyme function annotation, and accession number are found in front of each sequence. Bsp, Bacillus sphaericus; Bs, Bacillus sp.; Oih, Oceanobacillus iheyensis; Af, Anoxybacillus flavithermus; Bst, Geobacillus stearothermophilus; Ban, Bacillus anthracis; Bce, Bacillus cereus; Bha, Bacillus halodurans; Bac, Bacillus acidopullullyticus; Ls, Laceyella sacchari; Tvu, Thermoactinomyces vulgaris; Tet, Thermoanaerobacter ethanolicus; Tte, Thermoactinomyces tengcongensis. CD, Cyclodextrinase; MA, maltogenic amylase; NP, neopullulanase; AA,
α-amylase.
located in the A-, and B-domains (encompassing residues
127–501 in LsCda13 and 131–504 in AfCda13). The first of
the additional conserved regions contains a sequence (Y, Q,
I, F, P, E/D, R, F, A/P/N, N, G, D/N) of the A-domain corresponding to a unique Ca2+-binding loop found in both
TVAII and Geobacillus NPase (22, 24). This Ca2+-binding
site is possibly conserved among these related CD-degrading enzymes, as the residues proposed to be Ca2+-binding
ligands (and most of their neighboring residues in the primary structure) are conserved in both Af and LsCda13, and
in the other aligned sequences.
The Asp residue conserved in many enzymes harboring
the more well-known conserved primary calcium ion-binding site located between domains A and B, e.g., in the Taka
α-amylase from Aspergillus oryzae (25), is in the CD-degrading enzymes replaced by a lysine, and both TVAII and
NPase from G. stearothermophilus lack this site (the only
calcium ion detected in the structure of these enzymes was
located in the loop described above). For Ls and AfCda13,
like other CD-degrading enzymes, the lysine residue is conserved at this position (K295 LsCda13/ K297 AfCda13).
Another conserved region (W, L, R/E, G, D, Q, F, D, A/S,
W, M, N, Y, P/L) present in both Af and LsCda13 is located
between the classical regions III and IV, and is proposed to
belong to the CD-binding site (8).
Domain N
The N-domain of each enzyme (residues
1-113 in LsCda13 and 1-118 in AfCda13) had a lower degree of sequence conservation than the catalytic modules.
This domain from LsCda13 was 58% identical to the N-domain of TVAII, which was the most similar domain found.
The N-domain of AfCda13 showed 68% identity to the corresponding domains of both the MAase and the NPase from
G. stearothermophilus, while the identity to the N-domain
of LsCda13 was only 30.5%. Most of the conserved residues in the N-domain are aromatic residues, which are
located at 11 positions in the domain of the two novel enzymes, and in all their aligned homologues (Fig. 2). The
conserved tyrosine (located at position 45 in both Af and
LsCda13) has been proposed to be of importance for CD
binding and possibly also transglycosylating activity of
NPase from G. stearothermophilus (22).
Domain C
The sequences of the C-domains (residues
504–582 in LsCda13 and 507–586 in AfCda13) were the
least conserved. The C-domain of LsCda13 showed 40%
identity to that of TVAII, which was the highest identity
observed for this domain. The second highest match was
found for a completely different enzyme, namely, trehalose6P-hydrolase of Yersinia pestis, in which the C-terminus
VOL. 100, 2005
TWO CDases DIFFERING IN OLIGOMERIC STATE AND ACTIVITY PROFILE
FIG. 3. SDS–PAGE analysis of LsCda13 (A) and AfCda13 (B) purified by IMAC. Lane 1, Molecular weight marker; lane 2, LsCda13
(A) and AfCda13 (B).
showed 34% identity with the LsCda13 domain. Between
the two newly isolated enzymes, the C-domain identity was
only 19.8%. The C-domain of AfCda13 displayed a higher
identity to known sequences, and was 55% identical to the
MAase of G. stearothermophilus, and 38% identical to the
NPase from the same microorganism. The function of the
C-domain has been thought to be the stabilization of the
catalytic module, by shielding hydrophobic residues from
the solvent (1). When interpreting the alignment, very few
residues were found to be conserved in this domain, and
completely conserved residues were only found at four positions in the domain (Fig. 2).
Expression and purification
AfCda13 and LsCda13
were expressed in the E. coli strain BL21(DE3) (Novagen)
and the presence of the active form of both proteins was
confirmed using a microtiter plate-based DNS-activity screen
against a single substrate. The His-tag was utilized in a onestep purification of the two recombinant proteins by IMAC,
resulting in an enzyme purity of 80–90% for both enzymes
(Fig. 3). AfCda13 was then further purified, as described in
Materials and Methods, to ensure a purity of at least 90%
during the characterization steps.
Activity profile
Hydrolyzing activity was screened by
the microtiter plate-based 3,5-dinitrosalicylic acid assay of
reducing sugars using the substrates starch, pullulan, glycogen, amylose, amylopectin and α- and β-cyclodextrins. Specific activity was then determined for all substrates that
yielded positive results in the activity screen. The kinetic
data showed cyclodextrins to be the best substrates for both
enzymes, but interestingly there appeared to be a preference
of AfCda13 for α-CD and LsCda13 for β-CD (Table 2). Furthermore, the specific activity of AfCda13 for α-cyclodextrin was about tenfold higher than that of LsCda13, for
which the hydrolyzed starch levels were comparable to
those of cyclodextrin hydrolysis. Amylopectin and glycogen were not hydrolyzed at significantly high rates, suggesting that neither of the enzymes attack α-1,6 glycosidic linkages. Using β-CD as the substrate at a fixed temperature
(55°C), the apparent pH optimum was determined, and the
optimum pHs for LsCda13 and AfCda13 were found to be
6 and 6.5, respectively. The apparent optimum temperatures
for activity were 57°C and 57.5°C for LsCda13 and AfCda13,
at their respective optimum pH, using β-CD for LsCda13
and both α- (the most preferred substrate) and β-CDs for
AfCda13 (Fig. 4).
Hydrolysis products
Product formation following the
enzymatic digestion of α- and β-CDs, starch, amylose and
pullulan at different time intervals was examined by TLC.
LsCda13 degraded α- and β-CDs to mainly maltose and
glucose. AfCda13 also hydrolyzed α- and β-CDs to maltose
and glucose, but (albeit the higher initial specific activity)
some maltotriose, maltotetraose and maltoheptaose (for
β-CD) were still present after 24 h of incubation (Fig. 5).
This could be due to transglycosylation, as was proposed in
another study (26). Pullulan was degraded by both enzymes
to only one end product, which has been confirmed to be
panose (data not shown), a branched trisaccharide. Starch
and amylose were degraded by both enzymes to mainly
maltose and some glucose.
Oligomeric state
The Mr values estimated from SDS–
PAGE were 70 and 71 kDa for LsCda13 and AfCda13, respectively. These values are well in accordance with the
theoretically calculated molecular weights of 69.1 kDa
(LsCda13) and 72.8 kDa (AfCda13), including the tags. Gel
filtration experiments indicate that AfCda13 exists mainly
as a dimer at the optimum pH (native Mr of 148 ± 10 kDa),
whereas LsCda13 yielded a native M r of 74 ± 10 kDa, corresponding to a monomer in solution (Fig. 6). Confirmatory
gel filtration on a Bioprep SE 1000/17 column indicated the
same values for the native molecular weights.
To analyze the impact of oligomeric structure on substrate specificity, AfCda13 was incubated with KCl to shift
the monomer-dimer equilibrium. The incubation for 32 h in
KCl resulted in a conversion of approximately 10% of the
dimeric protein into a monomer. After the separation of
monomeric and dimeric proteins via gel filtration, the activ-
TABLE 2. Specific activities and kinetic parameters of LsCda13 and AfCda13 determined at 55°C for 15 min and at optimal pHa
Specific activityb (U mg–1)
kcat (s–1)
Km (mM/g l–1)
LsCda13
AfCda13
LsCda13
AfCda13
LsCda13
AfCda13
α-CD (1%)
40 ± 2
448 ± 28
46.1
545
7.9/7.7
0.28/0.27
β-CD (1%)
64 ± 2
259 ± 15
73.7
315
12/14
0.32/0.36
Soluble starch (2%)
32 ± 2
14 ± 3
36.8
16.9
5
11
NDc
Pullulan (1%)
37 ± 0.5
3.0 ± 0.1
42.6
3.6
NDc
Amylose (1%)
20 ± 4
11 ± 7
23
13.4
NDc
NDc
–1
Km is given both in mM and gl . The molecular mass of the monomer is used in Kcat determination to obtain values per active site.
a
pH 6 for LsCda13 and pH 6.5 for AfCda13.
b
Activities are in units per milligram of pure protein. Each value is the mean of at least three determinations ± SD.
c
ND, Not determined.
Substrate
385
386
J. BIOSCI. BIOENG.,
TURNER ET AL.
FIG. 4. The temperature dependence of LsCda13 (A) and AfCda13 (B) on activity was measured using the substrate that resulted in the highest
apparent specific activities of both enzymes. The activity of LsCda13 (0.04 mg ml–1) was measured in 50 mM potassium phosphate buffer, pH 6.0,
for 15 min using β-CD as the substrate. The activity of AfCda13 (0.01 mg ml–1) was determined in 50 mM sodium phosphate buffer, pH 6.5, after a
15 min reaction with α-CD.
FIG. 5. Thin-layer chromatography of hydrolysis products of β-CD, pullulan, and starch produced by LsCda13 (A) and AfCda13 (B), incubated at 55°C and 54°C, respectively. Different enzyme loadings were used to obtain suitable amounts of products formed. (A) Lane 1, Standards
(glucose [G1], maltose [G2], maltotetraose [G4], maltopentaose [G5], maltohexaose [G6], and maltoheptaose [G7]); lane 2, β-CD, 1 h; lane 3,
β-CD, 2 h; lane 4, β-CD, 4 h; lane 5, β-CD, 8 h; lane 6, β-CD, 16 h; lane 7, pullulan, 1 h; lane 8, pullulan, 2 h; lane 9, pullulan, 4 h; lane 10, pullulan, 8 h; lane 11, pullulan, 16 h; lane 12, starch, 1 h; lane 13, starch, 2 h; lane 14, starch, 4 h; lane 15, starch, 8 h; lane 16, starch, 16 h. (B) Lane 1,
Standards (glucose [G1], maltose [G2], maltotetraose [G4], maltopentaose [G5], maltohexaose [G6] and maltoheptaose [G7]); lane 2, β-CD, 0 h;
lane 3, β-CD, 1 h; lane 4, β-CD, 2 h; lane 5, β-CD, 4 h; lane 6, β-CD, 8 h; lane 7, β-CD, 24 h; lane 8, pullulan, 0 h; lane 9, pullulan, 1 h; lane 10,
pullulan, 2 h; lane 11, pullulan, 4 h; lane 12, pullulan, 8 h; lane 13, pullulan, 24 h; lane 14, starch, 0 h; lane 15, starch, 1 h; lane 16, starch, 2 h; lane
17, starch, 4 h; lane 18, starch, 8 h; lane 19, starch, 24 h.
ity of both fractions was measured in the presence of KCl.
The measurements of dimer activity without KCl revealed
that KCl did not affect the activity of the dimer. As compared with the dimeric protein, the monomeric enzyme retained only 23% of the specific activity towards α-cyclodextrin (Fig. 7B). The subsequent removal of KCl increased
the activity up to 80% of that monitored before KCl was introduced, but the amount of protein was too low to be detected after gel filtration. However, the increase in activity
after the removal of KCl is most likely a consequence of the
restoration of the dimer, since KCl was shown not to affect
the activity of the dimeric protein. Furthermore, the activity on soluble starch was not significantly affected by the
monomerization of AfCda13 (Fig. 7).
Thermal inactivation and calcium dependence
The
data for irreversible thermal inactivations at 60°C, prior to
either EDTA or Ca2+ treatment, were fitted to first-order inactivation kinetics yielding calculated half-lives of 4 and 20
min for LsCda13 and AfCda13, respectively. The effect of
calcium on thermostability was examined for both enzymes.
At 57°C (the optimum temperature), the activity of LsCda13
was completely lost after 20 min, but addition of 1 mM calcium effectively stabilized the enzyme to retain almost
100% of activity even after an incubation time of 120 min
(Fig. 8). Increasing the temperature to 60°C (120 min incubation) resulted in a 50% activity decrease, and a further increase in temperature to 67°C resulted in the complete loss
of activity after 120 min despite the presence of calcium
(data not shown). The thermal inactivation of AfCda13 was
not significantly affected by the addition of calcium (data
not shown). EDTA treatment also did not affect the activity
of AfCda13, suggesting that the enzyme either does not require calcium or contains calcium in a tightly bound form.
DISCUSSION
In this paper, the first known glycoside hydrolases from
both a Laceyella strain and an Anoxybacillus strain are described. These enzymes have specificity (CD > pullulan ≈
starch) prompting their classification among the CDases. In
most cases, these three substrates (CD, pullulan and starch)
were shown to be degraded to different degrees by enzymes
VOL. 100, 2005
TWO CDases DIFFERING IN OLIGOMERIC STATE AND ACTIVITY PROFILE
FIG. 6. Elution profiles of LsCda13 (A) and AfCda13 (B) after gel
filtration on Superdex 200 in 50 mM Tris–HCl, pH 7.4, containing
150 mM NaCl. Activity (circles) and protein concentration (squares)
are displayed. The calibration curve including the standards and the
eluted proteins are shown in panel C. The standards used were ribonuclease A (13.7 kDa), chymoptrypsin (25 kDa), ovalbumin (43 kDa),
BSA (66 kDa), aldolase (158 kDa) and catalase (232 kDa).
387
classified as CDases, MAases and NPases. Similarly to the
enzymes studied here, CDs appear to be the preferred substrates for many enzymes within all the three groups (8), although all enzymes of these three groups have not been
tested for all three substrates.
Another factor that further motivates classification of
AfCda13 and LsCda13 among the CDases (compared with
NPases or MAases), is the finding that both genes lack a putative signal peptide, indicating an intracellular location. In
many microbial species more than one type of GH13 enzyme is present, and can, depending on function, be either
intracellularly or extracellularly located. It is reasonable to
believe that enzymes whose main function in vivo is the
degradation of polymeric substrates should be located extracellularly. Cyclodextrinases, on the other hand, are intracellularly located, and a specific uptake mechanism for CDs
has been suggested for the bacterium Klebsiella oxytoca (8,
27). This, together with results on CD degradation in other
bacteria, suggests the presence of a specific bacterial carbohydrate metabolizing system (8). When this is taken into account, the physiological role of these CDases would thus
likely be in the metabolism of small oligosaccharides (to
mainly maltose and glucose) to allow their use in energy utilization in their native bacterium.
From sequence alignment and protein prediction we can
deduce that the two novel enzymes isolated have very similar overall structures (Fig. 1). As stated above, both prefer
cyclomaltodextrins as their substrates and exhibit largely
similar patterns of CD, starch and pullulan degradation, but
the activity of AfCda13 on CDs is markedly higher. Both
proteins contain the N-terminal domain, which has been reported to be important for high cyclomaltodextrinase activity and which is not present in amylases lacking CD specificity. This domain has also been proposed to be involved in
dimerization through its interaction with the active site region of the other monomer. As a result, a deep and narrow
cleft is formed, which can accommodate a CD molecule,
and it has hence been suggested that dimerization controls
substrate specificity depending on substrate size (23, 28).
Among CDases (and related MAases and NPases) the oligomeric state is most commonly the dimeric form (8). This is
also the case for AfCda13, as experimental results show this
enzyme to exist mainly as a dimer at the optimum pH. A
portion of this enzyme could be transferred to the monomeric state by incubating it with a high concentration of
KCl. Moreover, as the monomeric fraction exhibited a significantly lower activity on α-CD (23% of its original activity), it clearly suggests that dimerization is important for a
high CD activity. LsCda13 on the other hand, despite the
high similarity in sequence to the dimeric enzyme TVAII, is
a monomer in solution, and also has a much lower specific
activity for both α- and β-CD. Only few examples have so
far been reported with monomeric enzymes degrading CDs.
One of the few is the CDase from Thermoanaerobacter
ethanolicus, an enzyme that is also unique in that it has been
reported to be cell-bound (29). The initial activities reported
for this enzyme on α- and β-CDs are 71 and 62 U mg–1, respectively. This activity is rather similar to that found for
the monomeric LsCda13, but significantly lower than that of
dimeric AfCda13 (Table 2), which based on the kinetic data
388
TURNER ET AL.
J. BIOSCI. BIOENG.,
FIG. 7. Relative activity towards α-cyclodextrin (squares) and starch (circles) of dimeric (A) and monomeric (B) fractions of AfCda13 after
equilibration with KCl and gel filtration. The assay was performed at 57.5°C using 50 mM potassium phosphate buffer, pH 6.5, 5 mM DTE, the
corresponding substrate (0–0.5%) and protein (10 µg).
FIG. 8. Thermal inactivation of LsCda13 in absence (squares), and
presence (circles) of 1 mM Ca2+. The enzyme (5 µg) was incubated in
40 µl of a mixture of 50 mM piperazine, pH 6.0, and 1 mM CaCl2 at
57°C for up to 120 min. The residual activity was measured at 56.5°C
in 50 mM piperazine, pH 6.0, 1 mM CaCl2 and 0.5% β-cyclodextrin
using 30 µl of the incubated enzyme mixture. Relative values are given,
and 100% activity corresponds to 30 U mg–1 without calcium and 60
U mg–1 in the presence of calcium.
also shows a higher affinity for the CDs.
The reason for the difference in oligomeric state between
Af and LsCda13 is not clarified, and to solve this question,
structural information on the interaction of specific residues
would be needed. It is however interesting to note that the
additional conserved region (between the IIIrd and IVth
classical regions) proposed to belong to the CD-binding site
(8) is present in both Af and LsCda13, and so is Y45 in the
N-domain, also proposed to be important for CD interaction
(22). Based on the data from the sequence similarity search,
the lower degree of sequence conservation of the N-domain
of the respective enzyme indicates that interaction differences involving this domain determine the oligomeric state.
It has previously been noted by Kamitori et al. (24), that the
two enzymes TVAI and TVAII, despite similar overall domain structure, have very different degrees of intermolecular interaction between the monomers, in which TVAII
(which is expected to be dimeric), forms a number of interactions via the N-domain, while TVAI is not expected to be
dimeric and shows much weaker interactions.
The majority of the well-conserved residues in the N-domain are aromatic residues. These residues from other glycoside hydrolases are known to be important in binding interactions with carbohydrate substrates (30, 31), suggesting
that the involvement in CD binding is the main function of
the domain among the CDases. If dimerization does not
occur, the module is however not located very close to the
active site and its function in this case is unclear.
The sequences of the C-domains are not well conserved
among the CDases, and when the isolated module was used
for similarity search, the number of significant matches was
limited to only about ten sequences for each enzyme. Keeping in mind that the suggested function of this domain is to
shield hydrophobic residues of the A-domain from the solvent (1), this could explain the lower degree of sequence
conservation, as this would be dependent on the exposed
residues of the A-domain of each enzyme.
The thermostability of LsCda13 was shown to be dependent on the presence of calcium, indicating the presence of
at least one calcium-ion-binding site. Both TVAII (24), and
the related Geobacillus NPase (22), contain a unique Ca2+binding loop, which is located in a stretch of primary sequence that is very well conserved among the CD-degrading enzymes including LsCda13 and AfCda13. The thermostability of AfCda13 was however not affected by the addi-
VOL. 100, 2005
TWO CDases DIFFERING IN OLIGOMERIC STATE AND ACTIVITY PROFILE
tion of calcium, and a harsh EDTA treatment also did not
affect enzyme activity. Currently, we do not know whether
this is an indication of calcium independence or if a putative
calcium-ion-binding site has a higher affinity in this enzyme, and therefore is saturated throughout the purification
and characterization steps. A similar phenomenon was reported for TVAII, which has been shown to possess one calcium-ion-binding site (24).
In conclusion, despite the similarity in sequence and composition of structural elements between the two enzymes reported in this paper, the remaining differences between the
proteins, such as the relatively lower level of sequence identity between the N- and C-terminal domains, result in different subunit interactions. The dimerization is an important
factor that accounts for the higher CD activity of AfCda13,
but our results also confirm that monomeric enzymes (e.g.,
LsCda13) are capable of cyclodextrin hydrolysis.
10.
11.
12.
13.
14.
ACKNOWLEDGMENTS
We thank Joachim Ronnermark and Anna Schrum for help in
analytical work.
This work was supported by the European Union (contract no.
QLK3-CT-2000-01068), and the Swedish Research Council (VR).
Magnus Bergvall Foundation, Bengt Lundqvists Minne and the
Royal Physiographic Society in Lund are thanked for additional
support.
15.
16.
REFERENCES
1. MacGregor, E. A., Janecek, S., and Svensson, B.: Relationships of sequence and structure to specificity in the α-amylase family of enzymes. Biochim. Biophys. Acta, 1546, 1–20
(2001).
2. Coutinho, P. M. and Henrissat, B.: Carbohydrate-active enzymes: an integrated database approach, p. 3–12. In Gilbert,
H. J., Davies, G. J., Henrissat, B., and Svensson, B. (ed.), The
3rd Carbohydrate Bioengineering Meeting. The Royal Society of Chemistry, Newcastle upon Tyne (1999).
3. van der Maarel, M. J. E. C., van der Veen, B., Uitdehaag,
J. C. M., Leemhuis, H., and Dijkhuizen, L.: Properties and
applications of starch-converting enzymes of the α-amylase
family. J. Biotechnol., 94, 137–155 (2002).
4. Guzmán-Maldonado, H., and Paredes-López, O.: Amylolytic enzymes and products derived from starch: a review.
Crit. Rev. Food Sci. Nutr., 35, 373–403 (1995).
5. Kuriki, T. and Imanaka, T.: The concept of the α-amylase
family: structural similarity and common catalytic mechanism. J. Biosci. Bioeng., 87, 557–565 (1999).
6. Yamamoto, T.: Enzyme chemistry and molecular biology of
amylases, p. 81–100. In Yamamoto, T., Kitahata, S., Ciba, S.,
Hiromi, K., Matsuura, Y., Minamiura, N., Nitta, Y., Shinke,
R., and Yamane, K. (ed.), Enzyme chemistry and molecular
biology of amylases and related enzymes. CRC Press, Boca
Raton (1995).
7. Rose, T. M., Schultz, E. R., Henikoff, J. G., Pietrokovski,
S., McCallum, C. M., and Henikoff, S.: Consensus-degenerate hybrid oligonucleotide primers for amplification of distantly related sequences. Nucleic Acids Res., 26, 1628–1635
(1998).
8. Park, K.-H., Kim, T.-J., Cheong, T.-K., Kim, J.-W., Oh,
B.-H., and Svensson, B.: Structure, specificity and function
of cyclomaltodextrinase, a multispecific enzyme of the α-amylase family. Biochim. Biophys. Acta, 1478, 165–185 (2000).
9. Thompson, J. D., Higgins, D. G., and Gibson, T. J.:
17.
18.
19.
20.
21.
22.
23.
24.
25.
389
CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic
Acids Res., 22, 4673–4680 (1994).
Vieira, J. and Messing, J.: New pUC-derived cloning vectors
with different selectable markers and DNA replication origins.
Gene, 100, 189–194 (1991).
Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D.,
Seidman, J. G., Smith, J. A., and Struhl, K.: Current protocols in molecular biology, p. 15.14.11. John Wiley & Sons,
New York (2001).
Nicholas, K. B., Nicholas, H. B. J., and Deerfield, D. W. I.:
GeneDoc: analysis and visualization of genetic variation. Embnew. News, 4, 14 (1997).
Turner, P., Holst, O., and Nordberg Karlsson, E.: Optimized expression of soluble cyclomaltodextrinase of thermophilic origin in Escherichia coli by using a soluble fusion-tag
and by tuning of inducer concentration. Protein Expr. Purif.,
39, 54–60 (2005).
Ramchuran, S. O., Nordberg Karlsson, E., Velut, S., Maré
de, L., Hagander, P., and Holst, O.: Production of heterologous thermostable glycoside hydrolases and the presence of
host-cell proteases in substrate limited fed-batch cultures of
Escherichia coli BL21(DE3). Appl. Microbiol. Biotechnol.,
60, 408–416 (2002).
Abou Hachem, M., Nordberg Karlsson, E., BartonekRoxå, E., Raghothama, S., Simpson, P. J., Gilbert, H. J.,
Williamson, M. P., and Holst, O.: Carbohydrate-binding
modules from a thermostable Rhodothermus marinus xylanase:
cloning, expression and binding studies. Biochem. J., 345,
53–60 (2000).
Laemmli, U.: Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature, 227, 680–685
(1970).
Miller, G. L.: Use of dinitrosalicylic acid reagent for determination of reducing sugar. Anal. Chem., 31, 426–428 (1959).
Henrissat, B., Teeri, T. T., and Warren, R. A. J.: A scheme
for designating enzymes that hydrolyse the polysaccharides in
the cell walls of plants. FEBS Lett., 425, 352–354 (1998).
Tonozuka, T., Mogi, S., Shimura, Y., Ibuka, A., Sakai, H.,
Matsuzawa, H., Sakano, Y., and Ohta, T.: Comparison of
primary structures and substrate specificities of two pullulanhydrolyzing α-amylases, TVA I and TVA II, from Thermoactinomyces vulgaris R-47. Biochim. Biophys. Acta, 1252,
35–42 (1995).
Kamitori, S., Kondo, S., Okuyama, K., Yokota, T., Shimura,
Y., Tonozuka, T., and Sakano, Y.: Crystal structure of Thermoactinomyces vulgaris R-47 α-amylase II (TVAII) hydrolyzing cyclodextrins and pullulan at 2.6 Å resolution. J. Mol.
Biol., 287, 907–921 (1999).
Cha, H.-J., Yoon, H.-G., Kim, Y.-W., Lee, H.-S., Kim, J.-W.,
Kweon, K.-S., Oh, B.-H., and Park, K.-H.: Molecular and
enzymatic characterization of a maltogenic amylase that hydrolyzes and transglycosylates acarbose. Eur. J. Biochem., 253,
251–262 (1998).
Hondoh, H., Kuriki, T., and Matsuura, Y.: Three-dimensional structure and substrate binding of Bacillus stearothermophilus neopullulanase. J. Mol. Biol., 326, 177–188 (2003).
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.,
and other 4 authors: Crystal structure of a maltogenic amylase provides insights into a catalytic versatility. J. Biol. Chem.,
274, 26279–26286 (1999).
Kamitori, S., Abe, A., Ohtaki, A., Kaji, A., Tonozuka, T.,
and Sakano, Y.: Crystal structure and structural comparison
of Thermoactinomyces vulgaris R-47 α-amylase 1 (TVAI) at
1.6 Å resolution and α-amylase 2 (TVAII) at 2.3 Å resolution.
J. Mol. Biol., 318, 443–453 (2002).
Tada, S., Iimura, Y., Gomi, K., Takahashi, K., Hara, S.,
390
26.
27.
28.
29.
TURNER ET AL.
and Yoshizawa, K.: Cloning and nucleotide sequence of the
genomic Taka-amylase A gene of Aspergillus oryzae. Agric.
Biol. Chem., 53, 593–599 (1989).
Turner, P., Nilsson, C., Svensson, D., Holst, O., Gorton, L.,
and Nordberg Karlsson, E.: Monomeric and dimeric cyclomaltodextrinases reveal different modes of substrate degradation. Biologia, Bratislava, 60 (Suppl. 16) (2005). (in press)
Fiedler, G., Pajatsch, M., and Böck, A.: Genetics of a novel
starch utilisation pathway present in Klebsiella oxytoca. J.
Mol. Biol., 256, 279–291 (1996).
Park, K. H.: The multisubstrate specificity and the quaternary
structure of cyclodextrin-/pullulan-degrading enzymes. J. Appl.
Glycosci., 48, 293–299 (2001).
Podkovyrov, S. M., Burdette, D., and Zeikus, J. G.: Analysis of the catalytic center of cyclomaltodextrinase from Thermoanaerobacter ethanolicus 39E. FEBS Lett., 317, 259–262
J. BIOSCI. BIOENG.,
(1993).
30. Gilbert, H. J., Bolan, D. N., Szabo, L., Xie, H., Williamson,
M. P., Simpson, P. J., Jamal, S., Boraston, A. B., Kilburn,
D. G., and Warren, R. A. J.: An update on carbohydrate binding modules, p. 89–98. In Teeri, T. T., Svensson, B., Gilbert,
H. J., and Feizi, T. (ed.), The 4th Carbohydrate Bioengineering Meeting. The Royal Society of Chemistry, Stockholm
(2002).
31. Boraston, A. B., McLean, B. W., Kormos, J. M., Alam, M.,
Gilkes, N. R., Haynes, C. A., Tomme, C. A., Kilburn, D. G.,
and Warren, R. A. J.: Carbohydrate-binding modules: diversity of structure and function, p. 202–211. In Gilbert, H. J.,
Davies, G. J., Henrissat, B., and Svensson, B. (ed.), The 3rd
Carbohydrate Bioengineering Meeting. The Royal Society of
Chemistry, Newcastle upon Tyne (1999).