Content uploaded by Tjaard Pijning
Author content
All content in this area was uploaded by Tjaard Pijning on Nov 29, 2022
Content may be subject to copyright.
Crystal Structure of 4,6-α-Glucanotransferase GtfC-ΔC from
Thermophilic Geobacillus 12AMOR1: Starch Transglycosylation in
Non-Permuted GH70 Enzymes
Tjaard Pijning,*Evelien M. te Poele, Tijn C. de Leeuw, Albert Guskov, and Lubbert Dijkhuizen
Cite This: https://doi.org/10.1021/acs.jafc.2c06394
Read Online
ACCESS Metrics & More Article Recommendations *
sı Supporting Information
ABSTRACT: GtfC-type 4,6-α-glucanotransferase (α-GT) enzymes from Glycoside Hydrolase Family 70 (GH70) are of interest for
the modification of starch into low-glycemic index food ingredients. Compared to the related GH70 GtfB-type α-GTs, found
exclusively in lactic acid bacteria (LAB), GtfCs occur in non-LAB, share low sequence identity, lack circular permutation of the
catalytic domain, and feature a single-segment auxiliary domain IV and auxiliary C-terminal domains. Despite these dierences, the
first crystal structure of a GtfC, GbGtfC-ΔC from Geobacillus 12AMOR1, and the first one representing a non-permuted GH70
enzyme, reveals high structural similarity in the core domains with most GtfBs, featuring a similar tunneled active site. We propose
that GtfC (and related GtfD) enzymes evolved from starch-degrading α-amylases from GH13 by acquiring α-1,6 transglycosylation
capabilities, before the events that resulted in circular permutation of the catalytic domain observed in other GH70 enzymes
(glucansucrases, GtfB-type α-GTs). AlphaFold modeling and sequence alignments suggest that the GbGtfC structure represents the
GtfC subfamily, although it has a so far unique alternating α-1,4/α-1,6 product specificity, likely determined by residues near
acceptor binding subsites +1/+2.
KEYWORDS: GtfC, α-glucanotransferase, Glycoside Hydrolase Family 70, Geobacillus, α-1,4/α-1,6 alternan
■INTRODUCTION
Starch is a major energy-providing ingredient in many of our
foods; it is digested by starch-degrading human enzymes in the
gastrointestinal tract. The action of these enzymes, such as α-
amylases and glucosidases, may result in an undesirably rapid
release of glucose in the blood, increasing the risk of
cardiovascular diseases in the long term.
1
To lower such
risks, the food industry is aiming to produce starch-based
products with altered molecular structure, endowing prebiotic
properties.
1−6
The 4,6-α-glucanotransferase (4,6-α-GT) en-
zymes from Glycoside Hydrolase Family 70 (GH70) provide a
promising strategy to modify starch in this way as they
introduce α-1,6 glycosidic linkages, resulting in a slower
degradation.
7−10
The first characterized GH70 4,6-α-GTs were
found in lactic acid bacteria (LAB)
11−17
and designated the
GH70 GtfB subfamily. More recently, however, enzymes with
4,6-α-GT reaction specificity were also characterized in non-
LAB species, sharing low sequence similarity with GtfB
enzymes (<30%); these were designated the GtfC subfamily.
18
Of the 30 putative GtfC enzymes found in public databases by
2018,
2
four have been biochemically characterized;
18−22
among them are enzymes from thermophilic bacteria,
increasing the potential of these enzymes in an industrial
setting, as they were able to convert starch into linear
isomalto-/maltooligosaccharides at high temperatures (60−68
°C).
22
For example, adding the Geobacillus 12AMOR1 GtfC
(GbGtfC) enzyme during bread baking showed antistaling
eects. In addition to GtfCs, a few 4,6-α-GT enzymes with
even lower sequence similarity were identified in (plant-
associated) bacteria. The characterized enzymes in this group
synthesized reuteran-like branched α-glucans instead of linear
products,
23,24
thus defining another GH70 4,6-α-GT subfamily
(GtfD).
The transglycosylation reaction catalyzed by GH70 4,6-α-
GTs involves three catalytic residues (two Asp and one Glu)
and has been described by two half-reactions, each involving an
oxocarbenium-ion type transition state, stabilized by an Asp
residue. The first half-reaction is α-1,4 specific cleavage of the
substrate and results in a covalent enzyme-glycosyl inter-
mediate, which is transferred with α-1,6 specificity to an
acceptor substrate in the second half-reaction, leading to
isomalto-/maltooligosaccharide (IMMO/IMMP) products
containing α-1,6 linked units at the non-reducing end.
Previously, we structurally characterized 4,6-α-GT enzymes
of the GtfB subfamily
25,26
and proposed a reaction scheme
involving sliding of intermediate products through the binding
groove. We then hypothesized that the substrate and product
specificity of dierent GtfB-type 4,6-α-GTs is related to the
accessibility of the active site binding groove, which is defined
by two loops (A1 and B). A (phylogenetic) survey suggested
that about 80% of enzymes in the GtfB subfamily feature a
tunneled binding groove,
26
while the remaining ones are
Received: September 14, 2022
Revised: November 14, 2022
Accepted: November 16, 2022
Articlepubs.acs.org/JAFC
© XXXX The Authors. Published by
American Chemical Society A
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
Downloaded via 94.214.140.255 on November 29, 2022 at 07:49:07 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
(much) more open, allowing for the processing of branched
substrates and products. The so far characterized GtfC-type
4,6-α-GTs generated linear products, although it has to be
noted that the tested substrates were also largely linear. To
date, no GtfC protein 3D structures have been reported; given
the low sequence identity with GtfB-type 4,6-α-GTs (<30%)
the question is whether GtfCs feature a tunnel or not and if a
similar diversity with regard to active site openness exists.
Interestingly, the GtfC from Geobacillus 12AMOR1 (GbGtfC)
was found to have a unique product specificity.
22
With a
limited hydrolytic activity, GbGtfC releases mainly maltose
instead of glucose from amylose V or maltoheptaose substrate,
synthesizing a main product containing alternating α-1,4/α-1,6
linkages instead of consecutive α-1,6 linkages. This suggests
that GbGtfC exclusively transfers maltosyl units instead of
glucosyl units, but the structural details that confer this
property remain to be uncovered.
Importantly, the GtfC-type 4,6-α-GTs dier from their
GtfB-type relatives (and GH70 glucansucrases) regarding
domain organization. First, GtfCs lack the circular permutation
of the (β/α)8-barrel in the catalytic domain A, as is the case in
GH13 α-amylases belonging to the same clan GH-H.
18,27,28
Despite this absence of permutation, all seven conserved
sequence regions I−VII found in GH-H enzymes were
predicted to be present.
18
Second, GtfCs were predicted to
lack domain V and to have a single-segment domain IV. This
domain IV was proposed to have been inserted into domain B
of an ancestor α-amylase of the GH13_5 subfamily, which
mainly originate from bacteria and also act on starch-like
substrates, but lack this domain IV.
2,18,25,28
Finally, some GtfC-
type enzymes were predicted to feature additional C-terminal
domains of the bacterial Ig (type 2) fold.
2,18
Phylogenetic
analysis and predicted domain organization lead to the
hypothesis that the GtfC subfamily represents an intermediate
in a linear evolutionary pathway between GH13_5 α-amylases
and GtfB-type 4,6-α-GTs.
2,18
Yet, since no GtfC 3D structures
have been reported, it is still unknown whether GtfCs resemble
more the α-amylases or the GtfBs structurally.
Here, we report the first crystal structure of a GtfC-type
enzyme, the 4,6-α-GT from Geobacillus 12AMOR1 (GbGtfC),
revealing the 3D structure of the core domains A, B, C, and the
single-segment domain IV. Despite the absence of circular
permutation, GbGtfC features a tunneled active site
architecture that closely resembles the majority of GtfB-type
4,6-α-GTs. The obtained structure of the GbGtfC-ΔC enzyme
(at 2.25 Å resolution), together with docking experiments
depicting donor and acceptor reactions, allowed us to pinpoint
the residues in the active site that likely contribute to its unique
“alternating” specificity. AlphaFold modeling confirmed that
GbGtfC features two C-terminal domains of the Ig (type 2)
fold that are absent in the crystallized construct. Finally, we
show that the GbGtfC 3D structure represents the GtfC α-GT
subfamily as currently known, suggesting that the structural
changes necessary to acquire the α-1,6 starch-transglycosylat-
ing specificity of GH70 α-GTs from starch-degrading GH13 α-
amylases took place before domain permutation events.
■MATERIALS AND METHODS
Expression and Purification. The cloning and expression of the
GbGtfC-ΔC construct, containing residues 33−738 of Geobacillus
12AMOR1 GtfC and a 20-residue N-terminal His-tag, have been
described before.
22
Briefly, the pET15b vector carrying the gtf C
construct was overexpressed in E. coli BL21 (DE3) cultures grown at
37 °C; harvested cells were resuspended and broken by sonication;
cell-free extract (CFE) was stored at 4 °C. The GbGtfC-ΔC protein
in the CFE was captured by immobilized metal anity chromatog-
raphy (IMAC) on a Ni-Sepharose column (Sigma-Aldrich, St. Louis,
MO) using an elution buer containing 20 mM Tris-HCl, pH 8.0, 100
mM NaCl, and 350 mM imidazole. Fractions with the highest
absorbance at 280 nm were pooled and concentrated using a VivaSpin
4 (molecular weight cuto 10 kDa) at 4000g. The final purification
step was done via size exclusion chromatography on an A
kta Micro
system equipped with a Superdex 200 Increase 10/300 column
(Cytiva, Marlborough, MA) at 12 °C. The elution buer contained 20
mM MES-NaOH, pH 6.1, 100 mM NaCl, and 1 mM CaCl2. The
center fractions of the peak eluting at 13.3−14.8 mL were pooled
(Figure S1) and concentrated as described above to obtain the final
GbGtfC-ΔC protein sample suitable for crystallization. Protein
concentrations were determined by measuring the absorbance at
280 nm using a NanoDrop One spectrophotometer (Isogen Life
Science, De Meern, The Netherlands).
Crystallization and Data Collection. Crystals of GbGtfC-ΔC
were grown at 20 °C using a 10.0 mg/mL protein solution, 20 mM
MES-NaOH, pH 6.1, 100 mM NaCl, and 1 mM CaCl2. The reservoir
solution contained 1.07−1.14 M (NH4)2SO4, 0.1 M MES-NaOH, pH
6.5, and 0.4 M Na3citrate, and hanging drops were prepared by mixing
1.5 μL of protein solution and 1.5 μL of reservoir solution. Prior to
data collection, crystals were briefly transferred to 1.25 M (NH4)2SO4,
0.05 M MES-NaOH, pH 6.5, 0.2 M Na3citrate, and 30% (v/v)
glycerol and flash-cooled in liquid nitrogen. X-ray diraction data
were collected at beamline I03 of the Diamond Light Source (UK)
and processed using XDS;
29
statistics are given in Table 1.
Structure Determination and Refinement. The crystal
structure of GbGtfC-ΔC was determined by the molecular
replacement method using PHASER;
30
a template model was
generated by the one-to-one protocol of Phyre
31
based on the
highest scoring structure from a Phyre search, the crystal structure of
Table 1. Crystallographic Data Collection and Refinement
Statistics
PDB entry 7ZC0
resolution (Å) 131.3−2.25 (2.31−2.25)
space group I4122
cell dimensions a, b, c (Å) 262.6, 262.6, 72.2
unique observations
a
59408 (4359)
redundancy
a
1.9 (1.9)
completeness (%)
a
99.3 (95.0)
mean I/σ(I)
a
20.5 (4.2)
Wilson B-factor (Å2) 33.7
Rpim
a
0.030 (0.245)
CC1/2
a
0.999 (0.795)
R/Rfree 0.252/0.292
number of non-hydrogen atoms
protein 5667
glycerol 24 (4 ×6)
Ca2+/water 1/257
average B-factors
protein (Å2) 55.0
glycerol (Å2) 43.5
Ca2+/water (Å2) 21.3/39.0
root mean square deviations
bond lengths (Å) 0.006
bond angles (deg) 1.38
Ramachandran
favored (%) 92.0
allowed (%) 7.0
outliers (%) 1.0
a
Values in parentheses represent the highest resolution shell.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
B
the α-amylase from Halothermothrix orenii (PDB: 3BC9).
32
The
asymmetric unit of the I4122 cell contains one protein molecule.
Refinement and model building was carried out using Refmac
33
and
COOT;
34
groups for TLS refinement were determined using Phenix
35
and were edited manually to include domain IV as a separate TLS
group. The B-factor distribution showed a large range of values, with
relatively high values for domains C and IV (Figure S2). Some
stretches of residues in domain IV lacked good electron density,
especially residues 271−282, which were later modeled guided by an
AlphaFold generated model.
36,37
The final refinement statistics and model quality are listed in Table
1. Structural figures were prepared with PyMOL (The PyMOL
Molecular Graphics System, Version 2.0 Schrodinger, LLC). DSSP
38
was used to define secondary structure. Atomic coordinates and
structure factors have been deposited at the Protein Data Bank with
entry 7ZC0. PDBeFold
39
was used to analyze structural similarities,
with a lowest acceptable match threshold of 70% or 40%.
AlphaFold Modeling of GbGtfC and Homologues. The full
sequence of GbGtfC (GenBank AKM18207.1, 903 amino acid
residues) was subjected to AlphaFold modeling.
36,37
The model with
the highest overall pLDDT (per-residue confidence) score was used
for comparison with the crystal structure.
Additionally, AlphaFold models were calculated for 4 other GtfC-
type enzymes (Table S1), from Heyndrickxia sporothermodurans (902
residues), Weizmannia coagulans DSM1 (954 residues), Exiguobacte-
rium sibiricum 255-15 (893 residues), and Exiguobacterium acetylicum
(892 residues).
Modeling Donor Substrate Binding. We used the native crystal
structure to map the substrate binding groove of GbGtfC-ΔC; an
initial model was obtained by superposition with maltoheptaose (G7)
bound to subsites +2 to −5 of Lr121 GtfB
25
and inspected in
PyMOL. We then adjusted the glycosidic torsion angles of glucosyl
units in further subsites, to fit the binding groove of GbGtfC-ΔC
without clashes. An extra glucosyl moiety was added at the reducing
end (subsite +3), yielding a final maltooctaose (G8) model. The
corresponding residues from four other GtfC enzymes (H.
sporothermodurans,E. sibiricum 255-15, E. acetylicum, and W. coagulans
DSM1), as well as a GtfB-type 4,6-α-GT from L. reuteri 121
(Q5SBM0), were selected for a sequence alignment with ESPript
3.0.
40
Molecular Docking. Mixed isomalto-maltooligosaccharides
(DP1−6) were setup using SWEET2
41
and AutoDock Tools (version
1.5.6)
42
and docked in the crystal structures of GbGtfC (this study)
and Lr121 GtfB
25
using Vina-Carb,
43
representing scenarios for the
donor reaction or for the acceptor reaction with a covalent glucosyl-
enzyme intermediate at the catalytic nucleophile D413. All docking
results were visually inspected in PyMOL, judged by hydrogen-bond
interactions with catalytic residues, and then grouped by visual
similarity. Details of the docking procedures and interpretation of the
results are given in the Supporting Information.
Phylogenetic Analysis. A BLASTp search with default
parameters was performed (January 18, 2022) with the sequence of
Geobacillus 12AMOR1 GtfC (Genbank AKM18207.1). Using the full
sequences of the resulting hits, multiple sequence alignments were
performed with MUSCLE
44
and inspected within JalView 2;
45
sequences lacking significant parts of the GH70 core (containing
the conserved sequence regions (motifs) I−VII) were deleted. This
initial alignment was extended by three extra sets of sequences
representing biochemically characterized bacterial enzymes: (a) eight
canonical α-amylases from GH13 subfamily 5 (GH13_5); (b) five
GH70 glucansucrase sequences; and (c) six GH70 GtfB sequences.
26
The sequences used for the final alignment are shown in Table S2.
Residues constituting three important loops in GH70 GtfB-type 4,6-
α-GTs were identified on the basis of previously determined
structures:
25,26
loop B in domain B and loops A1 and A2 in domain
A (note that, in non-permuted GH70 sequences, loop A1 is C-
terminal to loop A2). A phylogenetic tree was constructed in MEGA
X
46
using the Maximum Likelihood method; the tree with the highest
log likelihood was used. Initial tree(s) for the heuristic search were
obtained automatically by applying Neighbor-Join and BioNJ
algorithms to a matrix of pairwise distances estimated using a JTT
model and then selecting the topology with superior log likelihood
value. Branch lengths were measured in the number of substitutions
per site. All positions with less than 95% site coverage were
eliminated, i.e., fewer than 5% alignment gaps, missing data, and
ambiguous bases were allowed at any position (partial deletion
option). There was a total of 512 positions in the final data set. The
Figure 1. (a) Overall crystal structure of GbGtfC-ΔC, with the domains indicated. The active site is located at the interface of domains A and B,
with catalytic residues D413, D446, and E517 shown as sticks. The Ca2+ ion near the active site is shown as a green sphere, and the first (V26) and
last (K735) visible residues are indicated. (b) Superposition of the crystal structure of GbGtfC-ΔC with that of Lr121 GtfB (transparent gray;
PDB: 5JBD).
25
The Lr121 GtfB enzyme features a somewhat larger domain IV as well as longer loops in domains A, B, and C (e.g., the β2−β3 and
β4−β5 connections in domain B) but shares the same overall topology.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
C
bootstrap consensus tree was inferred from 1000 bootstrap
replicates.
47
For a comparison between acceptor subsite residues in GtfC- and
GtfB-type 4,6-α-GTs, the 63 putative GtfC sequences were aligned
with a subset of the 283 putative GtfB sequences from Pijning et al.;
26
this subset contained 233 sequences with long loops A1 and B
(totaling 37−40 residues), likely featuring a tunneled binding groove.
■RESULTS AND DISCUSSION
Crystal Structure of GbGtfC-ΔC. Overall Structure. We
determined the crystal structure of GbGtfC-ΔC at a resolution
of 2.25 Å from crystals containing one protomer in the
asymmetric unit (Figure 1a) consisting of residues V26-K735.
The crystal structure comprises domains A, B, C, and IV and is
the first one representing a non-permuted GH70 enzyme. The
catalytic domain A (residues 26−144 and 387−630) contains
the (β/α)8barrel also found in other GH70 enzymes, but, like
in GH13 α-amylases, it starts with strand β1 and is interrupted
after helix α3 by a long insertion, forming domain B, as well as
the auxiliary domain IV, which is absent in GH13 enzymes.
Despite being non-permuted, the overall topology of domain A
is very similar to that of other GH70 structures (e.g., Lr121
GtfB; Figure 1b). On the other hand, some dierences were
observed in the elements that connect the α-helices and β-
strands of the (β/α)8barrel (e.g., in the β2-α2, α3-β4, and α4-
β5 connection). Domain B (residues 145−222 and 333−386)
has the central twisted five-stranded antiparallel β-sheet also
observed in other GH70 structures but is more compact,
mainly due to shorter connections between the β-strands. For
example, the connection between strands β2 and β3 (residues
191−210) is about 30 residues shorter than it is in Lr121 GtfB
and lacks two α-helices, while the loop connecting strands β4
and β5 (residues 357−380) is about nine residues shorter. The
connection between strands β3 and β4 is “extended” by the
insertion of about 110 residues that constitute domain IV
(residues 223−332). Finally, domain C (residues 631−736)
displays a similar Greek key topology as in other GH70 and
GH13 structures, albeit some loops that connect the β-strands
are either shorter or longer.
Despite the low sequence similarity, the GbGtfC-ΔC core
structure closely resembles that of GtfB-type 4,6-α-GTs.
18,25,26
Yet, PDBeFold analysis of the core domains (A, B, and C) of
the GbGtfC-ΔC crystal structure revealed that the closest
structural homologues are α-amylases from Alicyclobacillus sp.
(PDB: 6GXV)
48
and Geobacillus stearothermophilus (PDB:
4UZU)
49
with Q-scores of 0.46/0.44 and root-mean-square
deviations (RMSD) of 1.95/1.88 Å, respectively. Both these α-
Figure 2. Detailed comparison of the GbGtfC-ΔC crystal structure (colored as in Figure 1; this study) with that of Lr121 GtfB-ΔNΔV
(transparent gray; PDB: 5JBD).
25
(a) Stereo figure of the superposition based on domain IV alone. In GbGtfC, the 110-residue domain IV
(orange) is a single-segment insertion in domain B (green). In Lr121 GtfB, domain IV consists of an N-terminal segment (IVn, light gray)
preceded by its domain V, and a C-terminal segment (IVc, dark gray) far apart in sequence, each superimposing partly with domain IV of GbGtfC-
ΔC. The loop (residues 271−282 of GbGtfC-ΔC) connecting the small β-sheet is indicated with an asterisk. (b) Stereo figure of the loop
architecture around the active sites; loop A1 (purple), and loop B (brown) cover donor subsites of the groove, while loop A2 (red) forms the base
of the groove. The corresponding loops in Lr121 GtfB (gray) largely follow the same course. The catalytic residues are shown as sticks (in GbGtfC
these are the nucleophile D413, general acid/base E446, transition state stabilizer D517).
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
D
amylases belong to subfamily GH13_5, confirming structurally
the previous observation that this is the α-amylase subfamily to
which GH70 enzymes are evolutionary closest.
25
Only after
including domain IV to the PDBeFold search, structural
homologues of GH70 enzymes were detected, the closest one
being the 4,6-α-GT GtfB-ΔNΔV from L. reuteri 121 (Lr121
GtfB; PDB: 5JBD)
25
with a lower Q-score (0.24) than the α-
amylases but also a somewhat lower RMSD value (1.72 Å).
The GbGtfC-ΔC 3D structure confirms the earlier notion
that at the domain level it represents an intermediate between
GH13 α-amylases and GH70 GtfB-type α-GTs; regarding the
structural details of the core domains, and especially the active
site region, it is clearly similar to the GH70 GtfB-type α-GTs
and more distant from the GH13 α-amylases.
Domain IV Structure. The GbGtfC-ΔC crystal structure
reveals for the first time an inserted, uninterrupted domain IV
of a GH70 enzyme (Figure 2a). Domain IV comprises 110
residues (223−332) and is much smaller than the correspond-
ing domains in GtfB-type enzymes (usually about 170−180
residues); it connects to domain B via two loops that lie
adjacent to each other. The crystallographic B-factors for
domain IV are on average higher than for other domains
(Figure S2), indicating that this domain may be flexible due to
a hinged connection with domain B. Some of its residues
hardly showed electron density (Figure S3), but since domain
IV superimposed almost perfectly with that of the AlphaFold
model (see Figure 2), we confidently decided to include all
residues of this domain in the crystal structure.
Notably, PDBeFold analysis of domain IV alone did not
reveal significant structural similarity to known 3D structures
(all Q-scores below 0.12); this domain thus can be considered
a previously unobserved fold. Still, a manual inspection showed
that parts of domain IV can be superimposed with that of other
GH70 structures (e.g., Lr121 GtfB) (Figure 2a), taking into
account that the N- and C-terminal halves of which they are
composed, are “switched” due to the permutation. Indeed, the
N-terminal part of domain IV of GbGtfC (residues 223−245)
superimposes reasonably well with the C-terminal part of
domain IV of Lr121 GtfB (residues 1586−1614), even though
both lack secondary structure elements. For the other segment
(GbGtfC residues 246−332), the superposition is more
dicult, as the corresponding Lr121 GtfB segment (residues
761−898) features longer α-helices and longer loops. In
GbGtfC domain IV, residues 271−282 form a loop at the “top”
of domain IV connecting a short parallel β-sheet; a similar
architecture is seen in the crystal structures of Lr121 GtfB
(PDB: 5JBD)
25
and Limosilactobacillus reuteri NCC2613
(Lr2613) GtfB (PDB: 7P38
26
) (albeit with longer con-
nections).
Active Site and Binding Groove. The GbGtfC crystal
structure is the first representative of the GH70 GtfC α-GT
subfamily. Overall, the architecture of its binding groove
closely resembles that of the 4,6-α-GT Lr121 GtfB, more than
that of α-amylases: while the latter features a fully open
binding groove, in GbGtfC, the presence of the two long loops
A1 (residues 532−552) and B (residues 338−352) near the
binding groove results in a tunnel-like architecture that
encompasses donor subsites −2 and −3 (Figure 2b), similar
to the situation in Lr121 GtfB.
25
Alignment of these loops
(Figure 3) reveals that their sequences dier significantly from
those in Lr121 GtfB and that a shorter loop B is
“compensated” by a longer loop A1. The third loop A2
(residues 86−96) lies beneath the binding groove and is highly
conserved; it has a similar architecture as in Lr121 GtfB. The
tunneled architecture of the binding groove of GbGtfC
resembles that of the majority of putative GtfB enzymes
26
and is in agreement with the fact that GbGtfC products are
linear.
22
As proposed earlier,
25
the presence of the tunnel may
contribute to processivity of the transglycosylation by keeping
intermediate products bound to the enzyme; a shift in the
Figure 3. Sequence alignment of selected regions of GH70 4,6-α-glucanotransferases: GtfC-type GTs from Geobacillus 12AMOR1 (GbGtfC; this
study), Heyndrickxia sporothermodurans (HsGtfC),
52
Exiguobacterium sibiricum 255−15 (EsGtfC),
18
Exiguobacterium acetylicum DSM1 (EaGtfC),
19
Weissella confusa (WcGtfC), and a representative GtfB-type GT from Limosilactobacillus reuteri 121 (Lr121 GtfB).
25
Blue headings comprise
sections from domain A; green headings are those in domain B. Before alignment, the Lr121 GtfB sequence was manually rearranged (indicated by
a*) to match the non-permuted domain organization of GtfC enzymes. The 370s loop was manually aligned based on structural superposition
between GbGtfC and Lr121 GtfB. Residue numbering is from GbGtfC. Below the alignment, the subsites with which the respective residues
potentially interact are shown, based on the model for donor substrate binding; the four positions near subsites +1 and +2 that vary are indicated
with yellow background. The bars below the alignment represent loops A1, A2, and B near the active site; their colors match those in Figure 2b.
The three catalytic residues are indicated (NU = nucleophile, A/B = acid/base, TS = transition state stabilizing residue).
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
E
direction of the donor side of the binding groove was proposed
to explain the observed range of products with consecutive α-
1,6 linkages. A dierent explanation was recently proposed by
Yang et al.
50
stating that intermediate products instead shift
toward the acceptor side of the binding groove, keeping intact
the hydrogen bond interaction between the 6-OH of the sugar
in subsite −1 and a conserved glutamine.
Product Specificity. Given the observed structural
similarity in the binding groove between GbGtfC and Lr121
GtfB, it is intriguing that Lr121 GtfB (and other GtfBs)
synthesize products with consecutive α-1,6 linkages, whereas
GbGtfC forms alternating α-1,4/α-1,6 linkages. In fact,
GbGtfC so far is the only biochemically characterized GtfC-
type 4,6-α-GT displaying this specificity; understanding this
unique property is important regarding its application in starch
modification.
22,51
We therefore compared the 3D structures of
GbGtfC (this study) and Lr121 GtfB
25
and used them to
perform molecular docking with donor and acceptor
substrates. The active site of GbGtfC seems more constrained
around subsites +1/+2 than in Lr121 GtfB (Figure S5a);
moreover, while the residues surrounding donor subsites are
largely conserved, GbGtfC diers at four positions near
acceptor binding subsites (Figure 3 and 4b). Residues H417
(motif II) and Y375 (370s loop) belong to the variable set of
residues that have been suggested to aect product specificity
in GtfB-type α-GTs.
26
Residue Y375 of GbGtfC is close to
subsite +2 and may provide an aromatic stacking platform or a
hydrogen bond; for the corresponding P968 of Lr121 GtfB,
this is not the case. The larger side chain of Y375 also results in
a more constrained acceptor binding space in GbGtfC. Next to
Y375 lies H417, near acceptor subsite +1. Mutation of the
corresponding N1019 in Lr121 GtfB to histidine significantly
changed the linkage ratio (α-1,4/α-1,6) of the products
synthesized from amylose.
25
The third and fourth non-
conserved positions, T346 and V348 from loop B, locate at
the opposite side of the subsite +1 sugar unit; they are replaced
by S918 and T920 in Lr121 GtfB. Together, while the four
positions are largely conserved in a subset of 233 GtfBs that
likely feature a tunnel (Figure S4), the 63 putative GtfCs have
a dierent and less conserved set. Notably, GbGtfC is unique
among GtfCs with Y375 replacing D or N or K and the T346/
V348 pair replacing mostly S/I or S/S. This suggests that
Y375, T346, and V348 of GbGtfC contribute to its unique
product specificity. Supporting evidence comes from a recent
study with H. sporothermodurans GtfC
52
postulating that
mutation of the corresponding S345/I347 to T/V resulted in
products with alternating α-1,6/α-1,4 linkages rather than
consecutive α-1,6 linkages.
It was proposed earlier that α-1,4/α-1,6 alternating end
products of GbGtfC can be explained by an α-1,6 trans-
glycosylation preference for maltosyl rather than glucosyl
moieties, supported by the accumulation of maltose and hardly
any glucose upon incubation of the enzyme with amylose V.
22
However, the synthesis of α-glucans by 4,6-α-GTs proceeds
through many cleavage and transfer steps. To understand how
the final product spectrum is obtained would require a
systematic analysis of every possible reaction for each possible
donor or acceptor substrate. Indeed, our docking experiments
suggested that the situation is more complicated than can be
explained by a single transglycosylation preference. Never-
theless, the docking experiments with GbGtfC and Lr121 GtfB
(methods and results described in the Supporting Information
and Figure S5) did allow us to derive some principles that
agree with the experimentally observed end products of either
enzyme.
22,25
First, a general and rather unexpected observation
was that, for both enzymes, donor and acceptor reactions do
not seem to be restricted to α-1,4-resp. α-1,6-specificity, but
also can occur with α-1,6- resp. α-1,4-specificity. Yet, α-1,6
Figure 4. (a) Model of a possible donor substrate, maltooctaose (G8), in the active site groove of GbGtfC with the enzyme represented as a
surface; domain A is colored blue and domain B is colored green. The three catalytic residues are shown as sticks. Part of the binding groove
features a tunnel spanning at least subsites −2 and −3. (b) The same G8 model, with surrounding residues shown as sticks; residues are colored
according to domain (blue = domain A, green = domain B) or loop name (red = loop A2, brown = loop B). Some of the corresponding residues in
Lr121 GtfB are shown with gray carbon atoms; in particular, Y375, H417, and T346 (GbGtfC) are close to acceptor subsites +1 and +2 and are
replaced by P968, N1019, and S918 from Lr121 GtfB, respectively.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
F
transglycosylations become dominant over α-1,4 transglycosy-
lations, because (intermediate) products of the latter can easily
“react back” because the glucosyl moiety in subsite +1 hardly
requires a change in conformation to act in a subsequent donor
reaction (Figure 5a). In contrast, α-1,6-transglycosylation
products do not react back as donors, as a large reorientation
would be needed for the subsite +1 glucosyl moiety to do so.
Second, we found that both enzymes are able to transfer
glucosyl as well as maltosyl moieties, but GbGtfC seems to be
less ecient in cleaving the non-reducing end (NR) terminal
α-1,4 linkage from maltosyl-ending intermediate products
(Figure 5b). For example, in a docking scenario with isopanose
in GbGtfC, the α-1,4 linkage did not assume a favorable
position for cleavage while the α-1,6 linkage did (Figure S5b).
The result is that, with GbGtfC, intermediate products with
NR maltosyl ends “survive”, and these are easily elongated by
α-1,6-transglycosylation, favoring the formation of alternating
glucan products. The experimentally observed maltose in the
reaction pool of GbGtfC
22
likely results from a more ecient
α-1,4-transglycosylation of glucose than in Lr121 GtfB. Finally,
the docking results suggest that the described dierences
between GbGtfC and Lr121 GtfB relate to interactions of
Figure 5. (a) Docking experiments comparing donor and acceptor reactions regarding the +1 sugar unit, shown here for GbGtfC (similar
observations were made for Lr121 GtfB). The left panel shows that, for a maltose α-1,4-reacting acceptor (yellow sticks), the conformation of the
+1 glucosyl does not dier much from that of a maltotetraose donor (cyan lines). In contrast, for α-1,6-specific donor and acceptor reactions, the
+1 sugar unit assumes very dierent orientations, as is shown for a maltose acceptor (yellow sticks) and 6′O-α-maltotriosyl-glucose donor (cyan
lines) (right panel). (b) Docking of isopanose in GbGtfC (yellow and light gray carbon atoms for ligand and surrounding residues, respectively)
and Lr121 GtfB (cyan and dark gray carbon atoms, respectively). In contrast to the situation in Lr121 GtfB, the trisaccharide assumes a
conformation unlikely to be α-1,4 cleaved by GbGtfC.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
G
donor/acceptor substrates in subsites +1 and +2 with the non-
conserved residues described above (Table S3), further
supporting the role of these residues in determining the
unique product specificity of GbGtfC.
AlphaFold Model of Full-Length GbGtfC and Other
GtfC Enzymes. The average per-residue confidence score
(pLDDT) of the highest ranked AlphaFold model of GbGtfC
was 92.6. The N-terminal 32 residues of GbGtfC correspond
to the signal peptide and expectedly showed significantly lower
pLDDT scores (Figure S6a); omitting these residues improved
the average pLDDT to 94.8 (Table S1), indicating a highly
reliable model. The AlphaFold model superposed well with the
crystal structure (RMSD = 0.79 Å for 591 Cαatoms), even for
most of the loop regions (Figure 6a); nevertheless, some
dierences were observed. First, domain IV has a slightly
dierent orientation relative to the core of the enzyme (Figure
6a), supporting the notion that this domain may be slightly
flexible around the hinge formed by the two loops connecting
it to domain B. On its own, the modeled domain IV
superimposes well with that in the crystal structure (Figure
6b) and includes the segments that showed poorly defined
electron density. The second most obvious dierences between
the modeled and experimental structure occur in the loop
regions near the active site (Figure 6c). The AlphaFold models
show slightly dierent conformations of loops A1 and B, with
shifts up to 3.6 Å with respect to the crystal structure, but the
general course of the loops is the same. In the active site
region, almost all side chains were modeled with the same
rotamer as that of the crystal structure; exceptions are H372,
Y375 and L378 (not shown).
The AlphaFold model of GbGtfC also includes the C-
terminal ∼165 residues that are absent in the crystallized
construct; as predicted previously,
22
they form two bacterial Ig-
like type 2 domains (Ig2), which connect to domain C via a
short loop (residues 734−738) (Figure 6a). Although the high
pLDDT scores for the Ig2 domains of GbGtfC (Figure S6a)
indicate reliable modeling of their fold, the relative orientation
of these domains is modeled with less confidence, especially
regarding the C-terminal Ig2 domain. Domain Ig2a (residues
739−823) and domain Ig2b (residues 824−903) share low
sequence identity (26.2%) but have the same immunoglobulin
fold; they can be superimposed giving an RMSD of 0.74 Å.
Both domains contain nine β-strands and form two opposing,
mostly antiparallel β-sheets (Figure 6d). However, the first two
β-strands (A and B) can be considered interrupted, and this
results in subsheets composed of A−B’, B−E−D, and A’−G−
F−C.
The BLASTp results indicate that on a residue level GbGtfC
is rather unique among GtfC subfamily enzymes: it is the only
enzyme from a Geobacillus species, and the closest homologues
in terms of sequence (from H. sporothermodurans) show 76.3%
sequence identity. Some of its residues near the binding groove
are dierent from most GtfC sequences (see above). This
raised the question how representative the GbGtfC 3D
structure is for the GtfC subfamily of 4,6-α-GTs. We therefore
constructed AlphaFold models of four other GtfC-type GTs
(Table S1), three of which were characterized as 4,6-α-GTs
synthesizing linear isomalto/maltooligosaccharides with con-
secutive α-1,6 linkages. The AlphaFold models showed
comparable pLDDT scores and very similar folds (Figure
S7a), reflected in low RMSD values of 0.54−0.72 Å upon Cα
superposition with GbGtfC. Notably, the high structural
conservation includes not only the core domains A, B, and C
but also domain IV. Near the active site region, loops A1, A2,
and B have somewhat lower pLDDT scores (not shown).
Although there are slight dierences in position with
dierences up to 3.8 Å (in the tip of loop A1), these loops
have the same architecture as in GbGtfC and form a tunnel at
the donor side of the binding groove (Figure S7b). We thus
suggest that, although Geobacillus 12AMOR1 GbGtfC has
some unique features near the active site, the 3D structure of
the core domains of this enzyme represents the whole GtfC
subfamily, at least for the 63 sequences found so far.
Like GbGtfC, the C-terminal domains of the GtfC from H.
sporothermodurans,E. acetylicum, and E. sibiricum 255-15
feature two Ig2 domains; for the latter, this was already
Figure 6. AlphaFold model of GbGtfC (gray) superposed on the GbGtfC-ΔC crystal structure (colored domains); the 29 N-terminal residues of
the AlphaFold model were omitted while the C-terminal Ig2 domains extend away from domain C. (a) Overall superposition with RMSD = 0.79 Å.
(b) Superposition based on domain IV (residues 223−332), with RMSD = 0.55 Å. (c) Loop regions near the active site. (d) Topology of the Ig2a
domain of the AlphaFold model with the β-strands labeled; the Ig2b domain (see a) has the same topology.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
H
predicted in an earlier study.
18
Ig2 domains occur in various
bacterial and phage surface proteins and have been proposed
to play a role in cell surface adhesion or carbohydrate
binding.
53
For GtfCs, this remains to be investigated, but since
these enzymes are extracellular and process carbohydrates,
such functions seem to be possible. On the other hand, the
predicted structure of the W. coagulans DSM1 GtfC features
three C-terminal SRC Homology 3 (SH3) domains (Figure
S7a) of about 60 residues each; SH3 domains are thought to
mediate protein−protein interactions.
54
Variations in the
length of the C-terminal parts of the GtfC sequences found
by the BLASTp search (see below) suggests that the type and
the number of copies of the C-terminal domains could be
related to the bacterial species and its specific natural
environment.
Phylogenetic Relations and Evolutionary Aspects. A
BLASTp search with the Geobacillus 12AMOR1 GtfC
(GbGtfC) sequence yielded a total of 102 putative non-
permuted bacterial sequences containing the four conserved
GH70 motifs in the order I−II−III−IV (Table S2). All
sequences originate from non-LAB species, but based on their
sequence alignment they could be divided in two groups. The
first group contains 63 hits, more than double the number of
sequences identified in 2018
2
and shows sequence identities of
52.9−76.3% with GbGtfC. The enzymes within this group
originate mainly from Gram-positive soil or marine bacteria
such as Weizmannia coagulans or Exiguobacterium species; for
example, the earlier characterized GtfCs from Exiguobacterium
sibiricum 255-15
18
and Weizmannia coagulans DSM1
20
belong
to this group. Most sequences have a length of around 900
residues and share high sequence similarity, suggesting that
they are GtfC-type α-glucanotransferases constituting a similar
domain organization with the three core domains (A, B, and
C), an inserted domain IV, and extra C-terminal domains. The
second group, containing the remaining 39 sequences, showed
lower overall sequence identities (40.4−49.9% to GbGtfC)
and originate mostly from Gram-negative bacteria such as
Azotobacter chroococcum (a plant-associated nitrogen-fixing
Figure 7. Unrooted phylogenetic tree calculated from the 121 GH70 and GH13 amino acid sequences. GH70 contains several subfamilies:
glucansucrases (GS)/branching sucrases (Brs), GtfB-, GtfC-, and GtfD-type α-GTs, indicated by dierent colors. The number preceding each
sequence corresponds to the numbering in Table S2. GbGtfC (this study) is highlighted; sequences for which an AlphaFold model has been
calculated are indicated with a yellow dot. Important evolutionary branch separation events are indicated (I, II, and III).
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
I
species) or Burkholderia (animal/plant pathogen). Including
the previously characterized enzymes from Azotobacter
chroococcum NCIMB 8003
24
and from the Gram-positive
Paenibacillus beijingensis DSM 24997,
23
this group represents
putative GtfD-type α-glucanotransferases. In general, the
sequences in this group are shorter at the C-terminal end,
suggesting that they do not feature Ig-like domains.
A more detailed analysis of the sequence alignment of the
GH70 motifs, loops A1, A2, B, and the 370s loop within the
GtfC group revealed that they are highly conserved regarding
residue type (selected enzymes from this group are shown in
Figure 3) as well as loop length (Table S2). The Geobacillus
12AMOR1 GtfC sequence is rather unique in these regions
(note that it is the only Geobacillus entry found). Nevertheless,
the alignment strongly suggests that all 63 putative GtfC-type
α-GTs found so far feature a tunneled binding groove, prefer
mostly linear starch substrates, and synthesize linear α-glucan
products; this is also supported by the AlphaFold models of
selected GtfCs (see above). Whether GbGtfC is the only GtfC
synthesizing products with alternating α-1,4/α-1,6 linkages
remains to be investigated. A detailed biochemical character-
ization of more GtfC-type GTs and their products is needed to
confirm this.
A phylogenetic tree generated from the extended alignment
of GH70 and GH13_5 enzymes (Figure 7) sheds more light
on the GH13/GH70 evolutionary pathways originally
conceived by Vujicic-Zagar et al.
55
and later extended/refined
by Gangoiti et al.
18
A clear distinction is seen between the
GH13_5 α-amylases that degrade but not transglycosylate
starch substrates and the GH70 enzymes that acquired α-1,6
transglycosylation capabilities. Importantly, for the GH70
sequences, three bifurcation points (I, II, and III) are apparent
(Figure 7). Point I signifies the distinction between non-
permuted and permuted GH70 enzymes. On one hand, in
non-LAB species, the enzymes remained non-permuted, and
later evolved dierently in Gram-positive (GtfC) or (mostly)
Gram-negative (GtfD) enzymes (point II): while the GtfC-
type enzymes acquired extra C-terminal domains and kept the
tunnel-like architecture, the GtfD-type enzymes seem to have
evolved to feature shorter loops A1 (Table S2) likely related to
their reaction specificity involving more branched substrates
and products.
23,24
On the other hand, in LAB species,
permutation did take place (via gene duplication) (Figure
7); a later bifurcation (point III) signifies that part of the
enzymes changed their substrate specificity from starch (GtfB)
to sucrose (glucansucrases, branching sucrases) by further
adapting their active site architecture.
25
Notably, despite the absence of permutation and despite a
dierent domain composition, the GtfC- and GtfD-clades are
phylogenetically closer to other GH70 enzymes (GtfB-type α-
GTs, glucansucrases, and branching sucrases) than they are to
the GH13_5 α-amylases. The GbGtfC-ΔC crystal structure (as
well as the AlphaFold models of other GtfC enzymes) clearly
confirms this, showing the high structural similarity with GtfB
Figure 8. Evolutionary pathway depicting the domain organization and permutation in GH13 and GH70 enzymes, partly based on earlier
findings.
18,55
The core domains A, B, and C are present throughout; N- and C-termini are indicated with Nt and Ct, respectively. GH13 α-
amylases, appearing in all kingdoms of life, acquired transglycosylation specificity by changing structural elements around the active site, while the
additional insertion of domain IV into domain B resulted in a (still non-permuted) GH70 ancestor α-GT. From such an ancestor (likely
corresponding to point I in Figure 6), two “branches” evolved. The first branch evolved in non-LAB species: GH70 GtfC- and GtfD-type 4,6-α-
glucanotransferase enzymes (4,6-α-GT) remained non-permuted, featuring the same single-segment domain IV, and acquiring additional C-
terminal Ig2- or SH3-type domains. In the second branch, evolving in LAB species, the GH70 GtfB-type, glucansucrase (GS) and branching sucrase
(BrS) enzymes became circularly permuted, with domain IV consisting of two segments far apart in sequence. The enzymes in this branch acquired
dierent auxiliary domains at their N- and/or C-termini, again far apart in sequence. Notably, as the GbGtfC crystal structure shows, the structure
of the core domains in the non-LAB and LAB branches is remarkably similar, especially in the active site region. In all (sub)families, domain A
contains the four homology motifs I−IV; circular permutation in GH70 enzymes changes their order such that motif I is placed C-terminal of
motifs II−III−IV; thus, the order changes from I−II−III−IV to II−III−IV−I.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
J
enzymes, but diering from the GH13_5 α-amylases, which
have a more open active site groove and lack certain structural
elements in the core domains (e.g., a two-helix/loop insertion
between β-strands 7 and 8, as well as the long loops A1 and B).
Thus, the high structural similarity between GtfC and GtfB-
type α-GTs shows that the gene duplication step occurring in
LAB did not lead to large structural changes in the core
domains, consistent with their shared substrate and reaction
specificity (α-1,4 cleavage followed by α-1,6-transglycosylation
of starch-like compounds). This also suggests that the changes
that were necessary to acquire α-1,6 transglycosylation
specificity, as well as the insertion of domain IV, took place
before the division between LAB and non-LAB (bifurcation
point I), likely in bacterial α-amylase enzymes and leading to
an ancestor α-GT enzyme (Figure 8). The role of domain IV
in GH70 enzymes and why it was inserted is unclear; while
there are examples of starch-targeting GH13 α-amylases with a
carbohydrate binding domain (CBM) inserted in domain B,
56
in GbGtfC (and other GH70 enzymes), domain IV structurally
does not resemble a CBM domain and did not reveal
carbohydrate binding sites. Finally, the phylogenetic tree
shows that within the GtfC clade, the Geobacillus 12AMOR1
GtfC is in a rather unique position, perhaps related to the
observed dierences in residues surrounding the binding
groove as described above.
■ASSOCIATED CONTENT
*
sı Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.jafc.2c06394.
Figures of elution profile of three size exclusion
chromatography runs, crystal structure of GbGtfC-ΔC,
stereo figures of the GbGtfC crystal structure and
electron density, sequence logos of four non-conserved
residues, selected docking results for donor and acceptor
reactions in GbGtfC and Lr121 GtfB, AlphaFold model
of GbGtfC, and comparison of AlphaFold models of
GbGtfC and four other GtfC-type GTs, discussion of
molecular docking, and tables of AlphaFold models of
selected GtfC-type GTs with their relative sequence
identity, list of sequences used for the alignment of 121
GH70/GH13 enzymes using the GbGtfC sequence as
reference, correlation between reactivity and subsite +1/
+2 glucosyl interactions, and references (PDF)
■AUTHOR INFORMATION
Corresponding Author
Tjaard Pijning −Biomolecular X-ray Crystallography,
Groningen Biomolecular Sciences and Biotechnology Institute
(GBB), University of Groningen, 9747 AG Groningen, The
Netherlands; orcid.org/0000-0003-4107-3663;
Phone: +31503634385; Email: t.pijning@rug.nl;
Fax: +31503634800
Authors
Evelien M. te Poele −Microbial Physiology, Groningen
Biomolecular Sciences and Biotechnology Institute (GBB),
University of Groningen, 9747 AG Groningen, The
Netherlands; CarbExplore Research B.V., 9747 AA
Groningen, The Netherlands
Tijn C. de Leeuw −CarbExplore Research B.V., 9747 AA
Groningen, The Netherlands; orcid.org/0000-0002-7452-
2819
Albert Guskov −Biomolecular X-ray Crystallography,
Groningen Biomolecular Sciences and Biotechnology Institute
(GBB), University of Groningen, 9747 AG Groningen, The
Netherlands; orcid.org/0000-0003-2340-2216
Lubbert Dijkhuizen −Microbial Physiology, Groningen
Biomolecular Sciences and Biotechnology Institute (GBB),
University of Groningen, 9747 AG Groningen, The
Netherlands; CarbExplore Research B.V., 9747 AA
Groningen, The Netherlands
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.jafc.2c06394
Funding
This work was financially supported by Royal AVEBE (to
CarbExplore Research BV) and the University of Groningen
(to T.P.).
Notes
The authors declare the following competing financial
interest(s): E.M.t.P., T.C.d.L., and L.D. are employed by
CarbExplore Research BV, which has received financial
support from Royal AVEBE.
■ACKNOWLEDGMENTS
The beamline sta at beamline I03 of the Diamond Light
Source is acknowledged for assistance during X-ray diraction
data collection. The authors thank Egor Marin for assistance
with AlphaFold modeling and the Center for Information
Technology of the University of Groningen for their support
and for providing access to the Peregrine high performance
computing cluster.
■REFERENCES
(1) Gangoiti, J.; Corwin, S. F.; Lamothe, L. M.; Vafiadi, C.;
Hamaker, B. R.; Dijkhuizen, L. Synthesis of novel α-glucans with
potential health benefits through controlled glucose release in the
human gastrointestinal tract. Crit. Rev. Food Sci. Nutr. 2020,60, 123−
146.
(2) Gangoiti, J.; Pijning, T.; Dijkhuizen, L. Biotechnological
potential of novel glycoside hydrolase family 70 enzymes synthesizing
α-glucans from starch and sucrose. Biotechnol. Adv. 2018,36, 196−
207.
(3) Te Poele, E. M.; Corwin, S. G.; Hamaker, B. R.; Lamothe, L. M.;
Vafiadi, C.; Dijkhuizen, L. Development of slowly digestible starch
derived α-glucans with 4,6-α-glucanotransferase and branching
sucrase enzymes. J. Agric. Food Chem. 2020,68, 6664−6671.
(4) Gu, F.; Borewicz, K.; Richter, B.; van der Zaal, P. H.; Smidt, H.;
Buwalda, P. L.; Schols, H. A. In Vitro fermentation behavior of
isomalto/malto-polysaccharides using human fecal inoculum indicates
prebiotic potential. Mol. Nutr. Food Res. 2018,62, No. 1800232.
(5) Jurásková, D.; Ribeiro, S. C.; Silva, C. C. G. Exopolysaccharides
produced by lactic acid bacteria: from biosynthesis to health-
promoting properties. Foods 2022,11, 156.
(6) Miao, M.; Jiang, B.; Jin, Z.; BeMiller, J. N. Microbial starch-
converting enzymes: Recent insights and perspectives. Compr. Rev.
Food Sci. Food Saf. 2018,17, 1238−1260.
(7) Leemhuis, H.; Dobruchowska, J. M.; Ebbelaar, M.; Faber, F.;
Buwalda, P. L.; van der Maarel, M. J.; Kamerling, J. P.; Dijkhuizen, L.
Isomalto/malto-polysaccharide, a novel soluble dietary fiber made via
enzymatic conversion of starch. J. Agric. Food Chem. 2014,62, 12034−
12044.
(8) Meng, X.; Gangoiti, J.; Bai, Y.; Pijning, T.; Van Leeuwen, S. S.;
Dijkhuizen, L. Structure-function relationships of family GH70
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
K
glucansucrase and 4,6-α-glucanotransferase enzymes, and their
evolutionary relationships with family GH13 enzymes. Cell. Mol.
Life Sci. 2016,73, 2681−2706.
(9) Li, X.; Fei, T.; Wang, Y.; Zhao, Y.; Pan, Y.; Li, D. Wheat starch
with low retrogradation properties produced by modification of the
GtfB enzyme 4,6-α-glucanotransferase from Streptococcus thermophi-
lus.J. Agric. Food Chem. 2018,66, 3891−3898.
(10) van der Zaal, P. H.; Schols, H. A.; Bitter, J. H.; Buwalda, P. L.
Isomalto/malto-polysaccharide structure in relation to the structural
properties of starch substrates. Carbohydr. Polym. 2018,185, 179−
186.
(11) Kralj, S.; Grijpstra, P.; van Leeuwen, S. S.; Leemhuis, H.;
Dobruchowska, J. M.; van der Kaaij, R. M.; Malik, A.; Oetari, A.;
Kamerling, J. P.; Dijkhuizen, L. 4,6-α-Glucanotransferase, a novel
enzyme that structurally and functionally provides an evolutionary
link between Glycoside Hydrolase enzyme families 13 and 70. Appl.
Environ. Microbiol. 2011,77, 8154−8163.
(12) Leemhuis, H.; Dijkman, W. P.; Dobruchowska, J. M.; Pijning,
T.; Grijpstra, P.; Kralj, S.; Kamerling, J. P.; Dijkhuizen, L. 4,6-α-
Glucanotransferase activity occurs more widespread in Lactobacillus
strains and constitutes a separate GH70 subfamily. Appl. Microbiol.
Biotechnol. 2013,97, 181−193.
(13) Gangoiti, J.; van Leeuwen, S. S.; Gerwig, G. J.; Duboux, S.;
Vafiadi, C.; Pijning, T.; Dijkhuizen, L. 4,3-α-Glucanotransferase, a
novel reaction specificity in Glycoside Hydrolase family 70 and clan
GH-H. Sci. Rep. 2017,7, 39761.
(14) Gangoiti, J.; van Leeuwen, S. S.; Meng, X.; Duboux, S.; Vafiadi,
C.; Pijning, T.; Dijkhuizen, L. Mining novel starch-converting
Glycoside Hydrolase 70 enzymes from the NestleCulture Collection
genome database: The Lactobacillus reuteri NCC 2613 GtfB. Sci. Rep.
2017,7, 9947.
(15) Meng, X.; Gangoiti, J.; de Kok, N.; van Leeuwen, S. S.; Pijning,
T.; Dijkhuizen, L. Biochemical characterization of two GH70 family
4,6-α-glucanotransferases with distinct product specificity from
Lactobacillus aviarius subsp. aviarius DSM 20655. Food Chem. 2018,
253, 236−246.
(16) Ispirli, H.; Simsek, O
.; Skory, C.; Sagdıc, O.; Dertli, E.
Characterization of a 4,6-α-glucanotransferase from Lactobacillus
reuteri E81 and production of malto-oligosaccharides with immune-
modulatory roles. Int. J. Biol. Macromol. 2019,124, 1213−1219.
(17) Yang, W.; Sheng, L.; Chen, S.; Wang, L.; Su, L.; Wu, J.
Characterization of a new 4,6-α-glucanotransferase from Limosilacto-
bacillus fermentum NCC 3057 with ability of synthesizing low
molecular mass isomalto-/maltopolysaccharide. Food Bioscience
2022,46, 101514.
(18) Gangoiti, J.; Pijning, T.; Dijkhuizen, L. The Exiguobacterium
sibiricum 255−15 GtfC enzyme represents a novel Glycoside
Hydrolase 70 subfamily of 4,6-α-glucanotransferase enzymes. Appl.
Environ. Microbiol. 2016,82, 756−766.
(19) Kralj, S. Compositions and methods comprising the use of
Exiguobacterium acetylicum and Bacillus coagulans α-glucanotransferase
enzymes. US11072783B2, 2017.
(20) Xiang, G.; Buwalda, P. L.; van der Maarel, M. J. E. C.;
Leemhuis, H. The thermostable 4,6-α-glucanotransferase of Bacillus
coagulans DSM 1 synthesizes isomaltooligosaccharides. Amylase 2021,
5, 13−22.
(21) Wissuwa, J.; Stokke, R.; Fedøy, A. E.; Lian, K.; Smalås, A. O.;
Steen, I. H. Isolation and complete genome sequence of the
thermophilic Geobacillus sp. 12AMOR1 from an Arctic deep-sea
hydrothermal vent site. Stand. Genomic Sci. 2016,11, 16.
(22) Te Poele, E. M.; van der Hoek, S. E.; Chatziioannou, A. C.;
Gerwig, G. J.; Duisterwinkel, W. J.; Oudhuis, L.A.A.C.M.; Gangoiti, J.;
Dijkhuizen, L.; Leemhuis, H. GtfC enzyme of Geobacillus sp.
12AMOR1 represents a novel thermostable type of GH70 4,6-α-
glucanotransferase that synthesizes a linear alternating (α1→6)/(α1
→4) α-glucan and delays bread staling. J. Agric. Food Chem. 2021,69,
9859−9868.
(23) Gangoiti, J.; Lamothe, L.; van Leeuwen, S. S.; Vafiadi, C.;
Dijkhuizen, L. Characterization of the Paenibacillus beijingensis DSM
24997 GtfD and its glucan polymer products representing a new
Glycoside Hydrolase 70 subfamily of 4,6-α-glucanotransferase
enzymes. PLoS One 2017,12, No. e0172622.
(24) Gangoiti, J.; van Leeuwen, S. S.; Vafiadi, C.; Dijkhuizen, L. The
Gram-negative bacterium Azotobacter chroococcum NCIMB 8003
employs a new Glycoside Hydrolase family 70 4,6-α-glucanotransfer-
ase enzyme (GtfD) to synthesize a reuteran like polymer from
maltodextrins and starch. Biochim. Biophys. Acta 2016,1860, 1224−
1236.
(25) Bai, Y.; Gangoiti, J.; Dijkstra, B. W.; Dijkhuizen, L.; Pijning, T.
Crystal structure of 4,6-α-glucanotransferase supports diet-driven
evolution of GH70 enzymes from α-amylases in oral bacteria.
Structure 2017,25, 231−242.
(26) Pijning, T.; Gangoiti, J.; Te Poele, E. M.; Börner, T.;
Dijkhuizen, L. Insights into broad-specificity starch modification
from the crystal structure of Limosilactobacillus reuteri NCC 2613 4,6-
α-glucanotransferase GtfB. J. Agric. Food Chem. 2021,69, 13235−
13245.
(27) MacGregor, E. A.; Jespersen, H. M.; Svensson, B. A circularly
permuted α-amylase-type α/β-barrel structure in glucan-synthesizing
glucosyltransferases. FEBS Lett. 1996,378, 263−266.
(28) Janecek, S.; Svensson, B.; MacGregor, E. A. α-Amylase: an
enzyme specificity found in various families of glycoside hydrolases.
Cell. Mol. Life Sci. 2014,71, 1149−1170.
(29) Kabsch, W. XDS. Acta Crystallogr. D Biol. Crystallogr. 2010,66,
125−132.
(30) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M.
D.; Storoni, L. C.; Read, R. J. Phaser crystallographic software. J. Appl.
Crystallogr. 2007,40, 658−674.
(31) Kelley, L. A.; Mezulis, S.; Yates, C. M.; Wass, M. N.; Sternberg,
M. J. The Phyre2 web portal for protein modeling, prediction and
analysis. Nat. Protoc. 2015,10, 845−858.
(32) Tan, T. C.; Mijts, B. N.; Swaminathan, K.; Patel, B. K.; Divne,
C. Crystal structure of the polyextremophilic α-amylase AmyB from
Halothermothrix orenii: details of a productive enzyme-substrate
complex and an N domain with a role in binding raw starch. J. Mol.
Biol. 2008,378, 852−870.
(33) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Refinement of
macromolecular structures by the maximum-likelihood method. Acta
Crystallogr. D Biol. Crystallogr. 1997,53, 240−255.
(34) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features
and development of Coot. Acta Crystallogr. D. Struct. Biol. 2010,66,
486−501.
(35) Liebschner, D.; Afonine, P. V.; Baker, M. L.; Bunkóczi, G.;
Chen, V. B.; Croll, T. I.; Hintze, B.; Hung, L. W.; Jain, S.; McCoy, A.
J.; Moriarty, N. W.; Oeffner, R. D.; Poon, B. K.; Prisant, M. G.; Read,
R. J.; Richardson, J. S.; Richardson, D. C.; Sammito, M. D.; Sobolev,
O. V.; Stockwell, D. H.; Terwilliger, T. C.; Urzhumtsev, A. G.; Videau,
L. L.; Williams, C. J.; Adams, P. D. Macromolecular structure
determination using X-rays, neutrons and electrons: recent develop-
ments in Phenix. Acta Crystallogr. D. Struct. Biol. 2019,75, 861−877.
(36) Jumper, J.; Evans, R.; Pritzel, A.; Green, T.; Figurnov, M.;
Ronneberger, O.; Tunyasuvunakool, K.; Bates, R.; Zídek, A.;
Potapenko, A.; Bridgland, A.; Meyer, C.; Kohl, S. A. A.; Ballard, A.
J.; Cowie, A.; Romera-Paredes, B.; Nikolov, S.; Jain, R.; Adler, J.;
Back, T.; Petersen, S.; Reiman, D.; Clancy, E.; Zielinski, M.;
Steinegger, M.; Pacholska, M.; Berghammer, T.; Bodenstein, S.;
Silver, D.; Vinyals, O.; Senior, A. W.; Kavukcuoglu, K.; Kohli, P.;
Hassabis, D. Highly accurate protein structure prediction with
AlphaFold. Nature 2021,596, 583−589.
(37) Varadi, M.; Anyango, S.; Deshpande, M.; Nair, S.; Natassia, C.;
Yordanova, G.; Yuan, D.; Stroe, O.; Wood, G.; Laydon, A.; Zídek, A.;
Green, T.; Tunyasuvunakool, K.; Petersen, S.; Jumper, J.; Clancy, E.;
Green, R.; Vora, A.; Lutfi, M.; Figurnov, M.; Cowie, A.; Hobbs, N.;
Kohli, P.; Kleywegt, G.; Birney, E.; Hassabis, D.; Velankar, S.
AlphaFold protein structure database: massively expanding the
structural coverage of protein-sequence space with high-accuracy
models. Nucleic Acids Res. 2022,50, D439−D444.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
L
(38) Touw, W. G.; Baakman, C.; Black, J.; te Beek, T. A.; Krieger, E.;
Joosten, R. P.; Vriend, G. A series of PDB-related databanks for
everyday needs. Nucleic Acids Res. 2015,43, D364−8.
(39) Krissinel, E.; Henrick, K. Secondary-structure matching (SSM),
a new tool for fast protein structure alignment in three dimensions.
Acta Crystallogr. D Biol. Crystallogr. 2004,60, 2256−2268.
(40) Robert, X.; Gouet, P. Deciphering key features in protein
structures with the new ENDscript server. Nucleic Acids Res. 2014,42,
W320−4.
(41) Bohne, A.; Lang, E.; von der Lieth, C. W. SWEET - WWW-
based rapid 3D construction of oligo- and polysaccharides.
Bioinformatics 1999,15, 767−768.
(42) Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew,
R. K.; Goodsell, D. S.; Olson, A. J. AutoDock4 and AutoDockTools4:
Automated docking with selective receptor flexibility. J. Comput.
Chem. 2009,30, 2785−2791.
(43) Nivedha, A. K.; Thieker, D. F.; Makeneni, S.; Hu, H.; Woods,
R. J. Vina-Carb: Improving glycosidic angles during carbohydrate
docking. J. Chem. Theory Comput. 2016,12, 892−901.
(44) Edgar, R. C. MUSCLE: multiple sequence alignment with high
accuracy and high throughput. Nucleic Acids Res. 2004,32, 1792−
1797.
(45) Waterhouse, A. M.; Procter, J. B.; Martin, D. M.; Clamp, M.;
Barton, G. J. Jalview Version 2 - a multiple sequence alignment editor
and analysis workbench. Bioinformatics 2009,25, 1189−1191.
(46) Kumar, S.; Stecher, G.; Li, M.; Knyaz, C.; Tamura, K. MEGA
X: Molecular evolutionary genetics analysis across computing
platforms. Mol. Biol. Evol. 2018,35, 1547−1549.
(47) Jones, D. T.; Taylor, W. R.; Thornton, J. M. The rapid
generation of mutation data matrices from protein sequences. Comput.
Appl. Biosci. 1992,8, 275−282.
(48) Agirre, J.; Moroz, O.; Meier, S.; Brask, J.; Munch, A.; Hoff, T.;
Andersen, C.; Wilson, K. S.; Davies, G. J. The structure of the AliC
GH13 α-amylase from Alicyclobacillus sp. reveals the accommodation
of starch branching points in the α-amylase family. Acta Crystallogr. D.
Struct. Biol. 2019,75, 1−7.
(49) Offen, W. A.; Viksoe-Nielsen, A.; Borchert, T. V.; Wilson, K. S.;
Davies, G. J. Three-dimensional structure of a variant ‘Termamyl-like’
Geobacillus stearothermophilus α-amylase at 1.9 Å resolution. Acta
Crystallogr. F. Struct. Biol. Commun. 2015,71, 66−70.
(50) Yang, W.; Su, L.; Wang, L.; Wu, J.; Chen, S. α-
Glucanotransferase from the glycoside hydrolase family synthesizes
α(1−6)-linked products from starch: Features and synthesis pathways
of the products. Trends Food Sci. Technol. 2022,128, 160−172.
(51) Li, X.; Wang, Y.; Mu, S.; Ji, X.; Zeng, C.; Yang, D.; Dai, L.;
Duan, C.; Li, D. Structure, retrogradation and digestibility of waxy
corn starch modified by a GtfC enzyme from Geobacillus sp.
12AMOR1. Food Bioscience 2022,46, 101527.
(52) Yang, W.; Sheng, L.; Su, L.; Chen, S.; Wu, J. Directed mutation
of two key amino acid residues alters the product structure of the new
4,6-α-glucanotransferase from Bacillus sporothermodurans.J. Agric.
Food Chem. 2021,69, 14680−14688.
(53) Kelly, G.; Prasannan, S.; Daniell, S.; Fleming, K.; Frankel, G.;
Dougan, G.; Connerton, I.; Matthews, S. Structure of the cell-
adhesion fragment of intimin from enteropathogenic Escherichia coli.
Nat. Struct. Biol. 1999,6, 313−318.
(54) Kurochkina, N.; Guha, U. SH3 domains: modules of protein-
protein interactions. Biophys. Rev. 2013,5, 29−39.
(55) Vujicic-Zagar, A.; Pijning, T.; Kralj, S.; Lopez, C. A.; Eeuwema,
W.; Dijkhuizen, L.; Dijkstra, B. W. Crystal structure of a 117 kDa
glucansucrase fragment provides insight into evolution and product
specificity of GH70 enzymes. Proc. Natl. Acad. Sci. U.S.A. 2010,107,
21406−21411.
(56) Arnal, G.; Cockburn, D. W.; Brumer, H.; Koropatkin, N. M.
Structural basis for the flexible recognition of α-glucan substrates by
Bacteroides thetaiotaomicron SusG. Protein Sci. 2018,27, 1093−1101.
Journal of Agricultural and Food Chemistry pubs.acs.org/JAFC Article
https://doi.org/10.1021/acs.jafc.2c06394
J. Agric. Food Chem. XXXX, XXX, XXX−XXX
M
