Glycobiology vol. 25 no. 1 pp. 3–7, 2015
doi:10.1093/glycob/cwu085
Advance Access publication on August 18, 2014
COMMUNICATION
Theoretical study of enzymatic catalysis explains why
the trapped covalent intermediate in the E303C mutant
of glycosyltransferase GTB was not detected in the
wild-type enzyme
Adela Bobovská2,3, Igor Tvaroška2,4, and Juraj Kóňa1,2
2
Slovak Academy of Sciences, Institute of Chemistry, Center for Glycomics,
Bratislava 845 38, Slovakia; 3Faculty of Natural Sciences, Department of
Physical and Theoretical Chemistry, Comenius University, Bratislava 4 842 15,
Slovakia; and 4Department of Chemistry, Faculty of Natural Sciences,
Constantine The Philosopher University, Nitra SK-949 74, Slovakia
Received on December 11, 2013; revised on July 25, 2014; accepted on
August 12, 2014
Hybrid quantum mechanics/molecular mechanics calculations were used to study the catalytic mechanism of the
retaining human α-(1,3)-galactosyltransferase (GTBWT)
and its E303C mutant (GTBE303C). Both backside (via covalent glycosyl-enzyme intermediate, CGEI) and frontside
SNi-like mechanisms (via oxocarbenium-ion intermediate,
OCII) were investigated. The calculations suggest that both
mechanisms are feasible in the enzymatic catalysis. The nucleophilic attack of the acceptor substrate to the anomeric
carbon of OCII is the rate-determining step with an overall
reaction barrier (ΔE ‡ = 19.5 kcal mol−1) in agreement with
an experimental rate constant (kcat = 5.1 s−1). A calculated
α-secondary kinetic isotope effect (α-KIE) of 1.27 (GTBWT)
and 1.26 (GTBE303C) predicts dissociative character of the
transition state in agreement with experimentally measured
α-KIE of other retaining glycosyltransferases. Remarkably,
stable CGEI in GTBE303C compared with its counterpart in
GTBWT may explain why the CGEI has been detected by
mass spectrometry only in GTBE303C (Soya N, Fang Y,
Palcic MM, Klassen JS. 2011. Trapping and characterization of covalent intermediates of mutant retaining glycosyltransferases. Glycobiology, 21: 547–552).
Keywords: glycosyl-enzyme intermediate / kinetic isotope
effect / mechanism of glycosyltransferases / molecular
modeling / structure of transition state
Retaining glycosyltransferases (GTs) belong to a large GT
family of enzymes which catalyze the transfer of saccharides
1
To whom correspondence should be addressed: Tel: +421-2-59410203; Fax:
+421-2-59410222; e-mail: [email protected]
from activated donor substrates to various acceptor substrates
with retention of the stereo configuration of the anomeric carbon
of the transferred sugar unit (Taniguchi et al. 2002; Zhang et al.
2010). These enzymes play crucial roles in a variety of biological processes (Ohtsubo and Marth 2006) and are implicated in
many diseases and infections (Freeze and Aebi 2005; Vigerust
and Shepherd 2007). A retaining α-(1,3)-galactosyltransferase
(GTB, EC 2.4.1.37, the CAZy GT family 6, GT-A fold) is
the final GT in blood group B biosynthesis (Blanken and
Vandeneijnden 1985). The enzyme catalyzes the transfer of galactose (Gal) from uridine 5′-diphosphogalactose (UDP-Gal) to
H antigen acceptor [α-L-Fuc-(1,2)-β-D-Gal-O-R, where R is
glycolipid or glycoprotein, Supplementary data, Figure S1].
The structure and biochemical properties of GTB have been extensively studied (Marcus et al. 2003; Alfaro et al. 2008;
Sindhuwinata et al. 2010; Soya et al. 2011).
Based on X-ray structures of GTB (Patenaude et al. 2002;
Alfaro et al. 2008) and mass spectrometry analysis (Soya et al.
2011), a double displacement mechanism (double SN2) via a
covalent glycosyl-enzyme intermediate (CGEI) has been proposed with Glu303 residue as the catalytic nucleophile in a first
reaction step (Figure 1). Indeed, a single-point mutation of the
Glu303 with Ala resulted in 30,000-fold decrease of the enzymatic activity (Patenaude et al. 2002). However, CGEI has
never been trapped for the wild-type GTB, and its existence has
only been confirmed for the E303C mutant (Soya et al. 2011).
The kinetic analysis with bovine GTB (α1,3GalT), from the
same GT family 6 as GTB, demonstrated relatively high activity of the E317A mutant (0.1% activity of wild type), indicating
that Glu317 (the counterpart of Glu303 in GTB) plays a role of
the catalytic nucleophile less likely (Molina et al. 2007; Zhang
et al. 2010). However, following chemical rescue studies of the
E317A mutant of α1,3GalT with azide support a nucleophilic
role of Glu317 (Monegal and Planas 2006).
Large values of 1.20 (Lee et al. 2011) and 1.23 (Kim et al.
1988) for the α-secondary kinetic isotope effect (α-KIE) measured for other retaining GTs indicate a highly dissociative
oxocarbenium-ion-like transition state (OCI-TS). This may also
support an alternative stepwise SNi-like mechanism via an
oxocarbenium-ion intermediate (OCII) or a concerted SNi-like
mechanism via OCI-TS. The stepwise SNi-like mechanism has
recently been supported by computational modeling studies for
several retaining GTs: trehalose-6-phosphate synthase OtsA
© The Author 2014. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected]
3
A Bobovská et al.
Fig. 1. Proposed catalytic mechanisms for GTB.
(Ardèvol and Rovira 2011), α-1,2-mannosyltransferase (Kre2p/
Mnt1p) (Bobovská et al. 2014) and polypeptide N-acetylgalactosaminyltransferase 2 ( ppGalNAcT2) (Gómez et al. 2014;
Lira-Navarrete et al. 2014; T. Trnka et al. in preparation) and the
concerted mechanism was suggested for the lipopolysaccharylα-1,4-galactosyltransferase C (LgtC) (Tvaroška 2004; Gómez,
Polyak, et al. 2012). Because OtsA, Kre2p/Mnt1p, ppGalNAcT2
and LgtC belong to retaining GTs, which lack a suitably positioned catalytic nucleophile within the active site, the proposed
mechanisms could not be generalized for the retaining GTs
from the GT family 6. Subsequently, the enzymatic mechanism
of bovine GT α1,3GalT was studied with various computational
approaches (Gómez, Lluch, et al. 2012; Gómez et al. 2013;
Rojas-Cervellera et al. 2013). Using static hybrid quantum
mechanics/molecular mechanics (QM/MM) calculations on multiple enzyme:substrates complexes extracted from a molecular
dynamics simulations (Gómez, Lluch, et al. 2012; Gómez et al.
2013) concluded that both SNi-like (via OCII) and SN2 mechanism via CGEI are feasible for α1,3GalT. However, a systematic validation of localized transition states (TSs) for α1,3GalT
on the potential energy surface (PES) by frequency calculations
was not always possible due to flatness of the PES (Gómez
et al. 2013) and their understanding will require further validation. Rojas-Cervellera et al. (2013) used hybrid QM/MM
metadynamics simulations. They did not localize OCII on free
energy surface and catalytic process followed double SN2
mechanism via a stable CGEI. Based on experimental and theoretical studies the interpretations of the mechanisms of the
retaining GT from the GT family 6 remain unclear. Here, we
modeled two mechanisms, the frontside SNi-like via OCII and
the double displacement backside SN2 via CGEI for the human
GTBWT and its GTBE303C mutant. We focused on a comparison
of GTB and α1,3GalT mechanisms, and attempted to explain
4
why CGEI was detected by mass spectrometry analysis for the
GTBE303C mutant and not for the GTBWT.
Here we used static hybrid QM/MM calculations (DFT-M062X/OPLS2005) (Kaminski et al. 2001; Zhao and Truhlar 2008)
using the QSite program of the Schrödinger package (Qsite,
2011). The QM/MM methodology (an additive scheme) with
hydrogen caps and electrostatic treatment at the QM/MM interface was employed (Senn and Thiel 2007) (see computational
details in Supplementary data). We have used the static QM/
MM approach since it allows us to use an accurate QM method
and a larger QM part of the QM/MM system, the factors important for the correct prediction of reactivity of glycosidic
bonds (Bras et al. 2008; Kóňa and Tvaroška 2009; Passos et al.
2011; Gómez, Polyak, et al. 2012). The enzyme models were
based on a high-resolution X-ray crystal structure of a ternary
complex of a cis-AABB GTB mutant with donor and acceptor
substrates (PDB ID: 2RJ7) (Alfaro et al. 2008) and contain an
explicit water cap around the active site. The appropriateness of
the PDB structure, as a starting reactive form of Michaelis
complex, was validated by 100 ns molecular dynamic (MD)
simulations (the results from MD simulations are discussed in
Supplementary data). The QM part of the model consisted of
175 atoms and includes all structural fragments which directly
interacted with Mn2+ ion co-factor and are in the vicinity of the
anomeric reaction center (Supplementary data, Figure S1).
Here, we want to emphasize that the size of the QM part of the
GTB model was critical for the prediction of mechanistic features of the catalytic mechanism. We found that the inclusion of
all ligands directly interacting with oxygens of diphosphate
group of UDP-Gal into the QM part of the enzyme model was
necessary to localize all stationary points correctly for the stepwise SNi-like mechanism (namely an OCIIs 2A and 2B, and
transition state TS12 for the formation of OCII from Michaelis
complex 1). On the other hand, inclusion of a smaller part of
the acceptor substrate and its protein environment did not alter
structural and energetic parameters of the transition states and
intermediate (see Supplementary data, Tables SV–SIX and
Figure S9). Here, we performed the systematic validation of
transition state structures by various computational protocols
[frequency, intrinsic reaction coordinate (IRC) and kinetic isotopic effect calculations] to avoid misleading mechanistic interpretations of the reaction mechanism. The IRC calculations
allowed to sliding downhill from the TS along the TS eigenvector, calculating the gradient and taking a small step in the
negative gradient direction to reach energy minima corresponding to the TS. (All computational details including schemes
and figures are enclosed in the Supplementary data.)
The protonation states for all ionizable amino acid residues
of GTBWT and GTBE303C in the presence of bound Mn2+ ion
co-factor, donor and acceptor substrates for in vivo pH = 7.5
were assigned using the Propka v.2 program (Bas et al. 2008).
The predicted pKa = 5.6 of Glu303 indicates ionized glutamate
form (ca. 98.8%) at pH = 7.5. However, its slightly increased
pKa value implies that it can be easily protonated/deprotonated
at in vivo pH (catalytic Asp or Glu nucleophiles have pKa
values usually suppressed <4.5 to control a stable ionized form.
By this strategy an enzyme generates a nucleophile. Aspartyl
proteases are a typical example (Smith et al. 1996; Xie et al.
1997). Thus, Glu303 of GTB could not be the ideal candidate
Mechanism of glycosyltransferase GTB
for a catalytic nucleophile. A decreased nucleophilic character
of analogous catalytic nucleophile Glu317 in α1,3GalT was
also predicted by QM/MM calculations where its reactivity was
evaluated in the presence and absence of the acceptor substrate
(Gómez et al. 2013). The situation for Cys303 in GTBE303C is
different. Its predicted pKa = 9.1 indicates preferred a neutral
thiol form, although a minor amount (ca. 2.5%) of Cys303 is
allowed to be in the nucleophilic thiolate state at pH = 7.5.
Again, Cys303 does not seem to be a suitable candidate for the
catalytic nucleophile. Because Cys303 is buried in the hydrophobic environment consisted of substrates and GTB and is
with no direct contact with ionizable active-site amino acids, a
mode of its deprotonation to a reactive thiolate form is not
clear. Because both Glu303 and Cys303 are allowed to remain
in ionized states, albeit in different amounts, we included them
into reaction schemes in their ionized forms—the only forms
relevant for a role as a catalytic nucleophile.
To localize saddle points we mapped potential energy
surface (PES) along three reaction coordinates (distances C1D–
O1D, C1D–O3A and C1D–O1E; Supplementary data, Figures S4
and S5). After TS geometry optimization and IRC calculations,
the following reaction pathways were found. The path 1 →
TS12 → 2A → 2B → TS24 → 4 represents a stepwise SNi-like
mechanism via OCII, while the path 1 → TS12 → 2A →
TS23 → 3 → TS32 → 2B → TS24 → 4 represents double displacement process via CGEI (Supplementary data, Figures S2
and S3). As it is depicted in Figure 2 (Supplementary data,
Figure S2), the calculated stepwise SNi-like reaction path consists of two reaction steps: dissociation of the C1D–O1D glycosidic bond in UDP-Gal donor substrate (via TS12) and the
rate-determining formation of a new C1D–O3A glycosidic bond
between the transferred Gal and the acceptor sugar residue
Fig. 2. Calculated energy profiles (ΔEQM/MM) for GTBWT (black) and
GTBE303C (gray).
(via TS24). The IRC calculations confirmed that TS12 corresponds to 1 and OCII 2A, and TS24 to OCII 2B and intermediate 4. The step 2A→2B does not contain any chemical reaction
or conformational interconversion of the Gal ring. It represents
a translational movement of the oxocarbenium ion from UDP
to the 3-OH group of the terminal sugar ring of the acceptor
substrate. The QM/MM calculations predicted this step as a barrierless process. The correctness of all stationary point structures was validated by vibrational frequency calculations
(Supplementary data, Table SI). They confirmed that only one
glycosidic bond (C1D–O1D) is broken in TS12 and only one
(C1D–O3A) is formed in TS24. Indeed, the interatomic distances of breaking/forming bonds in these transition states correspond to dissociative reaction processes [d(C1D–O1D) = 2.17 Å
versus d(C1D–O3A) = 2.83 Å and d(C1D–O1E) = 2.85 Å in
TS12 (GTBWT); d(C1D–O3A) = 2.04 Å versus d(C1D–O1D) =
3.13 Å and d(C1D–O1E) = 2.66 Å in TS24 (GTBWT),
Supplementary data, Table SII]. The dissociative character of
the TS24 for the rate-determining step was also confirmed by
the α-KIE calculations. The calculated large values of 1.27
(GTBWT) and 1.26 (GTBE303C) are in reasonable agreement
with experimentally measured α-KIE of other retaining GTs
(Kim et al. 1988; Lee et al. 2011; Chan et al. 2012). An overall
reaction barrier (ΔE ‡) of 19.5 kcal mol−1 calculated for SNi-like
mechanism is in agreement with the measured kcat value (5.1 s−1)
for GTB (estimated ΔG ‡exp is ca. 17 kcal mol−1) (Marcus et al.
2003) and theoretically predicted barriers for the retaining GTs
(Ardèvol and Rovira 2011; Gómez, Polyak, et al. 2012; Gómez,
Lluch, et al. 2012; Gómez et al. 2013; Rojas-Cervellera et al.
2013; Bobovská et al. 2014; Gómez et al. 2014; Lira-Navarrete
et al. 2014).
For the alternative mechanism via CGEI, four reaction steps
were predicted (Supplementary data, Figure S3). The first and
last steps are identical with the stepwise SNi-like mechanism
described in the previous text. Thus, the formation of CGEI is
not concerted SN2 reaction started from 1, but stepwise
SN1-like process started from OCII (2A → TS23 → 3). The
anomeric carbon C1 in TS23 is attacked by the carboxylate of
Glu303 (or thiolate of Cys303 in the case of the GTBE303C)
from the β-face with d(C1D–O1E) = 1.88 Å (GTBWT) and d
(C1D–S1C) = 2.73 Å (GTBE303C). The IRC calculations of
TS23 resulted to OCII 2A and CGEI 3. It means that the formation of CGEI is the stepwise dissociative process via OCII and
the accepted double displacement SN2 mechanism proposed for
α1,3GalT (Gómez et al. 2013; Rojas-Cervellera et al. 2013) is
not appropriate for the retaining GTB. To localize a saddle
point for the formation of enzyme products directly from
CGEI, we scanned PES along the C1D–O3A and C1D–O1E (or
C1D–S1C in the case of GTBE303C) reaction coordinates.
Surprisingly, the CGEI 3 collapses to OCII 2B via TS32. Then,
the reaction follows the formation of the final catalytic product
4 via TS24. This reaction step is identical with one found for
the SNi-like mechanism. The attempts to optimize a transition
state structure structurally analogous to TS24 and belonging
solely to the concerted SN2 reaction failed. Overall, double
displacement mechanism via CGEI for both GTBWT and
GTBE303C is a stepwise process consisted of four reaction steps
with the rate-determining nucleophilic attack of the acceptor to
the anomeric carbon of OCII (Figure 2).
5
A Bobovská et al.
During the reaction, the conformation of the transferred
Gal ring changes from the stable 4C1 to the partially planar 4E
form in both GTBWT and GTBE303C [1(4C1) → TS12(4E) →
2A(4E) → TS23(4E/ 4H3) → 3(4E) → TS32(4E/ 4H3) → 2B(4E) →
TS24(4E) → 4(4E). See Supplementary data, Table SIV for
detailed puckering parameters].
Although the formation of CGEI is feasible for an enzymatic
reaction [ΔE ‡(1→TS23) = 16.8 and 12.0 kcal mol−1 for
GTBWT and GTBE303C], it is slower, compared with that for the
formation of OCII [ΔE ‡(1→TS12) = 11.0 and 11.3 kcal mol−1
for GTBWT and GTBE303C]. However, the rate-determining step
in both mechanisms is the formation of the product 4; therefore,
both mechanisms cannot be distinguished.
To understand why CGEI has been detected by mass spectrometry only in GTBE303C but not in GTBWT (Soya et al.
2011), we focused on a comparison of the energy profiles of
the double displacement mechanism. For both GTBWT and
GTBE303C they are similar except for the formation of CGEI
and its conversion to the product (Figure 2, Table I). While
CGEI is highly unstable in GTBWT (ΔE = 14.3 kcal mol−1) and
represents a shallow minimum on the PES, it is remarkably
stable in GTBE303C (ΔE = −14.7 kcal mol−1) occupying a deep
energy well with high activation energies for the forward and
reverse reaction steps (ΔE ‡E303C = 27.3 kcal mol−1 for 3 →
TS32 → 2B and 26.7 kcal mol−1 for 3 → TS23 → 2A steps). A
substantial difference in thermodynamic and kinetic nature
between CGEI in GTBWT and GTBE303C may explain the mass
spectrometry experiment with the E303C mutant (Soya et al.
2011). A similar conclusion was stated for α1,3GalT where
the formation of CGEI with Glu317 was endoergic
(ΔE = 11.4 ± 4.2 kcal mol−1) with a reverse barrier of only
3.7 ± 1.3 kcal mol−1 (Gómez et al. 2013).
Using QM/MM calculations with advanced DFT M06-2X
method, we found that both SNi-like mechanism via OCII and
SN1-like via CGEI are feasible for GTB and its E303C mutant
and these reactions have the stepwise nature. Similar conclusions about the preference of the mechanism were found for
α1,3GalT, albeit mechanistically, the formation of CGEI in
α1,3GalT was predicted as a concerted SN2 reaction (Gómez
et al. 2013). Glu303 may participate in catalysis as the nucleophile, however, this may not be its principal function. It seems
that the role of Glu303 is (i) in stabilization of reactive OCII
and protection from unwanted hydrolysis, (ii) controlling a
stereo configuration at the anomeric carbon in the transferred
Gal moiety and (iii) appropriate acceptor binding as it was
Table I. Relative energies (ΔEQM/MM) calculated for optimized stationary
points (kcal mol−1)
1
TS12
2A
2B
TS24
4
TS23
3
TS32
6
GTBWT
GTBE303C
0.0
11.0
10.8
12.0
19.5
6.7
16.8
14.3
15.8
0.0
11.3
11.2
12.5
16.2
−1.6
12.0
−14.7
12.6
suggested in previous studies on retaining α1,3GalT (Boix
et al. 2002; Zhang et al. 2003; Gómez, Llunch, et al. 2012;
Gómez et al. 2013).
Supplementary data
Supplementary data for this article are available online at http://
glycob.oxfordjournals.org/.
Acknowledgments
We thank Monika M. Palcic from Carlsberg Laboratory in
Valby for comments concerning the mechanism of GTs and
careful reading of the manuscript, and Stephen V. Evans from
University of Victoria in Canada for providing us with a crystal
structure of the E303C mutant of GTB.
Conflict of interest statement
None declared.
Funding
This work was supported by the Scientific Grant Agency of
MESR and SAS (the project VEGA-02/0101/11), Slovak
Research and Development Agency (the project APVV-048412), the SAS-NSC (Taiwan) Joint Research Programme (the
project SAS-NSC JRP 2012/8) and the Research & Development
Operational Programmes funded by the ERDF (CEGreenI,
Contract No. 26240120001 and CEGreenII, Contract No.
26240120025).
Abbreviations
α-KIE, α-secondary kinetic isotope effect; CGEI, covalent
glycosyl-enzyme intermediate; DFT, density functional theory;
Gal, galactose; GTB, human α-(1,3)-galactosyltransferase;
GTBWT, wildtype enzyme of human α-(1,3)-galactosyltransferase; GTBE303C, E303C mutant of human α-(1,3)-galactosyltransferase; GTs, glycosyltransferases; IRC, intrinsic reaction
coordinate; Kre2p/Mnt1p, α-1,2-mannosyltransferase; LgtC,
lipopolysaccharyl-α-1,4-galactosyltransferase C; MD, molecular dynamics; M06-2X, Minnesota M06-class functional with
double the amount of nonlocal exchange; OCII, oxocarbeniumion intermediate; OCI-TS, oxocarbenium-ion-like transition
state; PES, potential energy surface; ppGalNAcT2, polypeptide N-acetylgalactosaminyltransferase 2; QM/MM, quantum
mechanics/molecular mechanics; SN2, substitution nucleophilic bimolecular; SNi, substitution nucleophilic internal; SN1,
substitution nucleophilic unimolecular; TSs, transition states;
UDP-Gal, uridine 5′-diphosphogalactose
References
Alfaro JA, Zheng RB, Persson M, Letts JA, Polakowski R, Bai Y, Borisova SN,
Seto NO, Lowary TL, Palcic MM, et al. 2008. ABO(H) blood group A and
B glycosyltransferases recognize substrate via specific conformational
changes. J Biol Chem. 283:10097–10108.
Ardèvol A, Rovira C. 2011. The molecular mechanism of enzymatic glycosyl
transfer with retention of configuration: Evidence for a short-lived
oxocarbenium-like species. Angew Chem Int Ed. 50:1–6.
Mechanism of glycosyltransferase GTB
Bas DC, Rogers DM, Jensen JH. 2008. Very fast prediction and rationalization
of pK(a) values for protein-ligand complexes. Proteins Struct Funct Bioinf.
73:765–783.
Blanken WM, Vandeneijnden DH. 1985. Biosynthesis of terminal “Gal-Alpha1-]3gal-beta-1-]4glcnac-R oligosaccharide sequences on glycoconjugates –
Purification and acceptor specificity of a UDP-Gal-N-acetyllactosaminide
alpha-]3-galactosyltransferase from calf thymus. J Biol Chem. 260:
2927–2934.
Bobovská A, Tvaroška I, Kóňa J. 2014. A theoretical study on the catalytic
mechanism of the retaining alpha-1,2-mannosyltransferase Kre2p/Mnt1p:
The impact of different metal ions on catalysis. Org Biomol Chem.
12:4201–4210.
Boix E, Zhang YN, Swaminathan GJ, Brew K, Acharya KR. 2002. Structural
basis of ordered binding of donor and acceptor substrates to the retaining
glycosyltransferase, alpha-1,3-galactosyltransferase. J Biol Chem.
277:28310–28318.
Bras NF, Moura-Tamames SA, Fernandes PA, Ramos MJ. 2008. Mechanistic
studies on the formation of glycosidase-substrate and glycosidase-inhibitor
covalent intermediates. J Comput Chem. 29:2565–2574.
Chan J, Tang A, Bennet J. 2012. A stepwise solvent-promoted SNi reaction of
alpha-D-glucopyranosyl fluoride: Mechanistic implications for retaining glycosyltransferases. J Am Chem Soc. 134:1212–1220.
Freeze HH, Aebi M. 2005. Altered glycan structures: The molecular basis of
congenital disorders of glycosylation. Curr Opin Struct Biol. 15:490–498.
Gómez H, Lluch JM, Masgrau L. 2012. Essential role of glutamate 317 in
galactosyl transfer by alpha3GalT: A computational study. Carbohydr Res.
356:204–208.
Gómez H, Lluch JM, Masgrau L. 2013. Substrate-assisted and nucleophilically
assisted catalysis in bovine alpha 1,3-galactosyltransferase. Mechanistic
implications for retaining glycosyltransferases. J Am Chem Soc. 135:
7053–7063.
Gómez H, Polyak I, Thiel W, Lluch JM, Masgrau L. 2012. Retaining glycosyltransferase mechanism studied by QM/MM methods: Lipopolysaccharylalpha-1,4-galactosyltransferase C transfers alpha-galactose via an oxocarbenium ion-like transition state. J Am Chem Soc. 134:4743–4752.
Gómez H, Rojas R, Patel D, Tabak LA, Lluch JM, Masgrau L. 2014. A computational and experimental study of O-glycosylation. Catalysis by human
UDP-GalNAc polypeptide: GalNAc transferase-T2. Org Biomol Chem.
12:2645–2655.
Kaminski GA, Friesner RA, Tirado-Rives J, Jorgensen WL. 2001. Evaluation
and reparametrization of the OPLS-AA force field for proteins via comparison with accurate quantum chemical calculations on peptides. J Phys Chem
B. 105:6474–6487.
Kim SC, Singh AN, Raushel FM. 1988. The mechanism of glycogensynthetase as determined by deuterium-isotope effects and positional isotope
exchange experiments. J Biol Chem. 263:10151–10154.
Kóňa J, Tvaroška I. 2009. Comparative DFT study on the alpha-glycosidic
bond in reactive species of galactosyl diphosphates. Chem Papers.
63:598–607.
Lee SS, Hong SY, Errey JC, Izumi A, Davies GJ, Davis BG. 2011. Mechanistic
evidence for a front-side, SNi-type reaction in a retaining glycosyltransferase. Nat Chem Biol. 7:631–638.
Lira-Navarrete E, Iglesias-Fernández J, Zandberg WF, Compaňón I, Kong Y,
Corzana F, Pinto BM, Clausen H, Peregrina JM, Vocadlo DJ, et al. 2014.
Substrate-guided front-face reaction revealed by combined structural
snapshots and metadynamics for the polypeptide N-acetylgalactosaminyltransferase 2. Angew Chem Int Ed. 53:doi:10.1002/anie.201402781.
Marcus SL, Polakowski R, Seto NO, Leinala E, Borisova S, Blancher A,
Roubinet F, Evans SV, Palcic MM. 2003. A single point mutation reverses
the donor specificity of human blood group B-synthesizing galactosyltransferase. J Biol Chem. 278:12403–12405.
Molina P, Knegtel RMA, Macher BA. 2007. Site-directed mutagenesis of glutamate 317 of bovine alpha-1,3Galactosyltransferase and its effect on enzyme
activity: Implications for reaction mechanism. Biochim Biophys Acta Gen
Subj. 1770:1266–1273.
Monegal A, Planas A. 2006. Chemical rescue of alpha 3-galactosyltransferase.
Implications in the mechanism of retaining glycosyltransferases. J Am Chem
Soc. 128:16030–16031.
Ohtsubo K, Marth JD. 2006. Glycosylation in cellular mechanisms of health
and disease. Cell. 126:855–867.
Passos O, Fernandes PA, Ramos MJ. 2011. Theoretical insights into the catalytic mechanism of beta-hexosaminidase. Theor Chem Acc. 129:119–129.
Patenaude SI, Seto NOL, Borisova SN, Szpacenko A, Marcus SL, Palcic MM,
Evans SV. 2002. The structural basis for specificity in human ABO(H) blood
group biosynthesis. Nat Struct Biol. 9:685–690.
Qsite, version 5.7, Schrödinger, LLC, New York, NY, 2011.
Rojas-Cervellera V, Ardevol A, Boero M, Planas A, Rovira C. 2013. Formation
of a covalent glycosyl-enzyme species in a reatining glycosyltransferase.
Chem Eur J. 19:14018–14023.
Senn HM, Thiel W. 2007. QM/MM methods for biological systems. Top Curr
Chem. 268:173–290.
Sindhuwinata N, Munoz E, Munoz FJ, Palcic MM, Peters H, Peters T. 2010.
Binding of an acceptor substrate analog enhances the enzymatic activity of
human blood group B galactosyltransferase. Glycobiology. 20:718–723.
Smith R, Brereton IM, Chai RY, Kent SBH. 1996. Ionization states of the catalytic residues in HIV-1 protease. Nat Struct Biol. 3:946–950.
Soya N, Fang Y, Palcic MM, Klassen JS. 2011. Trapping and characterization of covalent intermediates of mutant retaining glycosyltransferases.
Glycobiology. 21:547–552.
Taniguchi N, Honke K, Fukuda M. 2002. Handbook of Glycosyltransferases
and Related Genes. 2nd ed. New York, NY: Springer.
Tvaroška I. 2004. Molecular modeling insights into the catalytic mechanism of
the retaining galactosyltransferase LgtC. Carbohydr Res. 339:1007–1014.
Vigerust DJ, Shepherd VL. 2007. Virus glycosylation: Role in virulence and
immune interactions. Trends Microbiol. 15:211–218.
Xie D, Gulnik S, Collins L, Gustchina E, Suvorov L, Erickson JW. 1997.
Dissection of the pH dependence of inhibitor binding energetics for an aspartic protease: Direct measurement of the protonation states of the catalytic
aspartic acid residues. Biochemistry. 36:16166–16172.
Zhang R, Yip VLY, Withers SG. 2010. Mechanisms of enzymatic glycosyl
transfer. In: Mander L, Liu H-W, editors. Comprehensive Natural Products
II: Chemistry and Biology. Kidlington: Elsevier. p. 385–422.
Zhang YN, Swaminathan GJ, Deshpande A, Boix E, Natesh R, Xie ZH,
Acharya KR, Brew K. 2003. Roles of individual enzyme-substrate
interactions by alpha-1,3-galactosyltransferase in catalysis and specificity.
Biochemistry. 42:13512–13521.
Zhao Y, Truhlar DG. 2008. The M06 suite of density functionals for main group
thermochemistry, thermochemical kinetics, noncovalent interactions, excited
states, and transition elements: Two new functionals and systematic testing of four
M06-class functionals and 12 other functionals. Theor Chem Acc. 120:215–241.
7