The reaction coordinate of a bacterial GH47 α-mannosidase: A combined quantum mechanical and structural approach

Article abstract of DOI:10.1002/anie.201205338

Mannosides in the southern hemisphere: Conformational analysis of enzymatic mannoside hydrolysis informs strategies for enzyme inhibition and inspires solutions to mannoside synthesis. Atomic resolution structures along the reaction coordinate of an inverting α-mannosidase show how the enzyme distorts the substrate and transition state. QM/MM calculations reveal how the free energy landscape of isolated α-D-mannose is molded on enzyme to only allow one conformationally accessible reaction coordinate. Copyright

Full text of DOI:10.1002/anie.201205338

Angewandte
Chemie
DOI: 10.1002/anie.201205338
Computational Chemistry
The Reaction Coordinate of a Bacterial GH47 a-Mannosidase: A
Combined Quantum Mechanical and Structural Approach**
Andrew J. Thompson, Jerome Dabin, Javier Iglesias-Fernꢀndez, Albert Ardꢁvol, Zoran Dinev,
Spencer J. Williams, Omprakash Bande, Aloysius Siriwardena, Carl Moreland, Ting-Chou Hu,
David K. Smith, Harry J. Gilbert, Carme Rovira,* and Gideon J. Davies*
The reaction coordinates of diverse mannosidases are of
fundamental mechanistic interest and are increasingly rele-
vant to inform the chemical syntheses of mannosides.[1]
Furthermore, the importance of a-mannosidase catalysis in
crucial biochemical events, in both the healthy cell and in the
context of disease, imparts special relevance to the dissection
of enzyme action and the specific inhibition of these enzymes.
One particularly important group are the so-called mannosi-
dase I Golgi and endoplasmic reticulum (ER) a-mannosi-
dases, and their homologues, grouped in the sequence-based
CAZY[2] family GH47. These enzymes are involved in the
biosynthetic “trimming” and/or remodeling of mannose-
containing N-glycans, or in their degradation. All character-
ized GH47 enzymes are inverting a-1,2 mannosidases; the
biosynthetic ER enzymes act on Man9GlcNAc2 glycans
(generating Man8GlcNAc2[3]), while the Golgi enzymes trim
the oligosaccharides ultimately into Man5GlcNAc2;[4] and
a third subgroup of ER-located enzymes play roles in the
degradation of misfolded proteins as part of the protein
folding quality control apparatus.[3a,5]
[*] A. J. Thompson, Dr. J. Dabin, Dr. T.-C. Hu, D. K. Smith,
Prof. G. J. Davies
Department of Chemistry, University of York
Heslington, York, YO10 5DD (UK)
E-mail: gideon.davies@york.ac.uk
J. Iglesias-Fernꢀndez, Dr. A. Ardꢁvol, Prof. C. Rovira
Computer Simulation & Modeling Laboratory
Parc Cientꢂfic de Barcelona
Baldiri Reixac 4, 08028 Barcelona (Spain)
and
Institut de Quꢂmica Teꢃrica i Computacional (IQTCUB) (Spain)
E-mail: crovira@pcb.ub.es
J. Iglesias-Fernꢀndez, Prof. C. Rovira
Present address: Departament de Quꢂmica Orgꢄnica
Universitat de Barcelona
Insight into the (a/a)7 GH47 fold and the presence of an
essential active site Ca2+ ion was first provided through the
determination of the structure of the yeast Saccharomyces
cerevisiae class I a-1,2 mannosidase.[6] Subsequently, through
studies of the human class I enzyme, a ring-flipped 1C4
conformation was observed for the inhibitor deoxymannojir-
imycin bound at the active center and mirrored in the
conformation of the bound bicyclic inhibitor, kifunensine.[7]
The Michaelis complex of the human enzyme, formed with an
S-linked disaccharide substrate mimic (1), was observed in
Martꢂ i Franquꢁs 1, 08028 Barcelona (Spain)
Dr. A. Ardꢁvol
Present address: Department of Chemistry and Applied
Biosciences, ETH Zurich, USI Campus
Via Giuseppe Buffi 13, CH-6900 Lugano (Switzerland)
Prof. C. Rovira
Instituciꢅ Catalana de Recerca i Estudis AvanÅats (ICREA)
Passeig Lluꢂs Companys 23, 08020 Barcelona (Spain)
Dr. Z. Dinev, Assoc. Prof. S. J. Williams
School of Chemistry and Bio21 Molecular Science and
Biotechnology Institute, University of Melbourne
Parkville (Australia)
3
a S1 conformation,[3b] an observation later corroborated by
docking calculations.[8] Together these data supported a con-
formational coordinate for GH47 enzymes, potentially
through a 3H4 conformation at, or close to, the transition
state. An understanding of the conformation of the transition
state is important in efforts to develop inhibitors based on the
concept of transition state mimicry; this is of great interest in
developing reagents to study the intersection of pathways
involving N-glycan maturation and remodeling as well as
degradation and disposal.
Dr. O. Bande, Dr. A. Siriwardena
Universitꢆ de Picardie Jules Vernes, Laboratoire des Glucides
(CNRS-FRE 3617), 33, Rue Saint Leu, 80039 Amiens (France)
C. Moreland, Prof. H. J. Gilbert
Institute for Cell and Molecular Biosciences
The Medical School, Newcastle University
Framlington Place, Newcastle upon Tyne, NE2 4HH (UK)
[**] We thank the UK Biotechnology and Biological Sciences Research
Council (BBSRC), the Spanish Ministry of Economy and Compet-
itiveness (MINECO, CTQ2011-25871), and the Generalitat de
Catalunya (2009SGR-1309). A.S. thanks the Indo-French Centre for
the Promotion of Advance Research (IFCPAR/CEFIPRA) for funding
and for a postdoctoral fellowship (to O.B.). A.A. acknowledges
a long-term fellowship from the European Molecular Biology
Organization (EMBO). We acknowledge the computer support,
technical expertise, and assistance provided by the Barcelona
Supercomputing Center: Centro Nacional de Supercomputaciꢅn
(BSC-CNS).
Herein, we present the structural analysis of Michaelis,
transition state, and product mimicking complexes with
a bacterial GH47 a-mannosidase, from Caulobacter strain
K31, which we show shares the specificity of the human GH47
enzymes for a-1,2 glycans. Caulobacter GH47 crystals diffract
to sub-ꢀngstrçm resolution (Table 1), thus allowing exquisite
atomic resolution analysis of ligand distortion: the Michaelis
complex of the S-linked substrate mimic 1 reveals the
equivalent distortion seen previously for the human ortho-
logue,[3b] whereas novel complexes with the inhibitors manno-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201205338.
Angew. Chem. Int. Ed. 2012, 51, 1 – 7
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
Table 1: 3D structure data for CkGH47.[a]
on a range of aryl a-d-mannosides, but kinetics could be
determined through fluoride-release using a-d-mannosyl
fluoride (Supporting Information, Figure S2). This substrate
yielded
k
cat = 1.9 sꢀ1 and KM = 5.0 mm with kcat/KM =
0.38 sꢀ1 mmꢀ1, thus allowing determination of a pH optimum
of approximately 7.0. Screening of a range of mannose-
disaccharides at the pH optimum showed CkGH47 to be an a-
1,2-mannosidase with kcat = 111 ꢁ 3 minꢀ1 and KM = 120 ꢁ
7 mm [kcat/KM = 0.93 minꢀ1 mmꢀ1 (15.4 sꢀ1 mmꢀ1)] against a-
1,2-mannobiose, with no observable activity against all other
linkages, which is consistent both with the activity of the
eukaryotic enzymes and also with a complex subsequently
obtained with 1, as discussed below. Working at the pH
optimum, CkGH47 a-mannosidase was inhibited by 1, 2, and
3 with KD values of 755 nm, 47 nm, and 99 nm respectively, as
determined by isothermal titration calorimetry (Figure S3).
Crystals of the Caulobacter GH47 enzyme diffract to
atomic (ꢂ 1 ꢀ) resolution. Apo-enzyme structures, as well as
those with 1–3 defining the reaction coordinate, were solved
to resolutions of 0.85, 0.85, 1.10, and 1.10 ꢀ, respectively
(Figure 1). As for previously solved mammalian orthologues,
the native structure of CkGH47 is an (a/a)7 helical barrel with
the catalytic center featuring the essential Ca2+ ion in a closed
cavity. Uniquely however, this bacterial orthologue also
coordinates a second Ca2+ located approximately 6.5 ꢀ
from the first at the base of the active site pocket, the
relevance of which, if any, is unclear. The reaction coordinate
was studied through complexes of 1–3. The non-hydrolyzable
disaccharide 1, binds in the ꢀ1 and + 1 subsites with the ꢀ1
Native
1
2
3
resolution [ꢇ]
Rmerge
I/sigma
comp. [%]
Rcryst/Rfree
rmsdbonds
rmsdangles
PDB Code
50–0.85
0.04 (0.39)
17.9 (2.6)
91.9 (63.4)
0.097/0.11
0.009
50–0.85
0.04 (0.57)
16.9 (1.7)
97.2 (80.8)
0.097/0.11
0.010
50–1.10
0.07 (0.28)
18.7 (5.4)
99.2 (94.2)
0.089/0.11
0.010
50–1.10
0.04 (0.21)
32.2 (6.7)
89.6 (49.5)
0.083/0.10
0.010
1.39
4AYO
1.43
4AYP
1.39
4AYQ
1.39
4AYR
[a] Values in parentheses denote highest resolution shell.
imidazole 2 and noeuromycin 3 allow the definition of
a complete conformational coordinate in which the manno-
side passes from a Michaelis complex in the 3S1 conformation
via a 3H4 transition state to product bound in a 1C4 con-
formation. Crucially, these atomic resolution analyses inform
a quantum mechanical description of the conformational
sphere of a-d-mannopyranose, presenting for the first time
the southern hemisphere, which is of direct relevance to
GH47 a-mannosidase catalysis. Quantum mechanical calcu-
lations show that the free energy landscape (FEL) of the
isolated a-d-mannopyranose molecule is strongly perturbed
on-enzyme and is restricted to a single conformational
pathway for the enzyme-bound ligand that is entirely
3
subsite mannoside distorted to an 3,OB / S1 conformation (1;
Figure 1). The active center Ca2+ ion bridges O2 and O3 of the
mannoside, thus facilitating ring distortion (to yield an O2-
C2-C3-O3 torsion angle of 478) with the putative nucleophilic
water aligned for in-line attack (water-C1-S1 angle of 1558).
This water interacts both with the active center Ca2+ and the
putative catalytic base, Glu365 (distances 2.4 and 2.7 ꢀ,
respectively). As with the human enzyme complex formed
with 1, the putative catalytic acid does not interact directly
with the interglycosidic atom with a Glu121 OE2–S1 distance
3
3
1
consistent with the experimentally supported S1! H4! C4
catalytic conformational progression.
To obtain a GH47 model that is easily produced and
diffracts to near atomic resolution a variety of enzymes were
studied. The gene encoding the bacterial GH47 from
Caulobacter strain K31 (CkGH47) was synthesized in
a codon-optimized form for expression in Escherichia coli
(see the Supporting Information). The enzyme was inactive
Figure 1. Observed electron density for 1–3 bound to Caulobacter Strain K31 a-1,2-mannosidase at 0.85–1.1 ꢇ resolution. The conformations
observed are the 3,OB/3S1 conformation for substrate mimic 1, the 3H4 conformation for the putative transition state, mannoimidazole 2, and 1C4
for noeuromycin 3, the observed conformation of which mimics that seen in the product.
2
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 7
These are not the final page numbers!

Angewandte
Chemie
of 4.5 ꢀ, which suggests an indirect protonation event,
perhaps through a water molecule. Given the atomic reso-
lution of the diffraction data, not only are many of the
“riding” hydrogen atoms observable but also the hydrogen
atoms of the O4 and O6 hydroxy groups of the ꢀ1 subsite
sugar, allowing an unprecedented level of detail of the
substrate interactions of this class of enzyme.
Complexes were also obtained with mannoimidazole 2
and noeuromycin 3. Mannoimidazole 2 is a putative transition
state mimic and binds in a 3H4 conformation, which provides
evidence for the transition state conformation. In solution,
noeuromycin 3 exists in a- and b-configured hemiaminal
forms (1:2 d-manno/d-gluco configurations),[9] 3 binds to
CkGH47 as the less favored d-manno-configured hemiaminal
in a ring-flipped 1C4 conformation, which mimics the product.
The major difference between the 3,OB/3S1 conformation of
3
the substrate mimic 1, the H4 conformation of the putative
transition state mimic 2, and the 1C4 conformation of the
product mimic 3 is the position of the (pseudo)anomeric
carbon. The change in the position of this atom is consistent
with the electrophilic migration of C1 from the glycosidic
oxygen to the nucleophilic water along the substitution
reaction coordinate.
Previous computational studies on b-d-glucopyranose, b-
d-mannopyranose, and a-l-fucopyranose have revealed that
the isolated sugar can access only a small number of possible
conformational coordinates, as reflected in lower free energy
states. However, individual enzyme families select only one of
these possible routes.[10] For a-d-mannopyranose (Figure 2a)
3
5
conformations B2,5, OS2, 3,OB, S1, B1,4, S1, and 2,5B (center of
the diagram), as well as 4C1, are among the most stable.
Moreover, these conformations were found to be preactivated
for catalysis in terms of energy, anomeric charge, and
structural parameters (see Figure S8 for the preactivation
index). The observed conformations for pseudo-Michaelis
3
complexes in a-mannosidases (OS2 and S1 in retaining and
inverting enzymes, respectively) are among the most stable
and thus represent preactivated conformations. GH125
inverting a-mannosidases show no substrate distortion in
their pseudo-Michaelis complex and whereas the conforma-
tional pathway for GH125 is unclear, it is certainly consistent
Figure 2. a) Conformational free energy landscape (FEL) of isolated
4
a-d-mannopyranose (Mercator projection including both northern and
southern hemispheres); contoured at 1 kcalmolꢀ1. b) Conformational
FEL of the a-mannosyl residue at the ꢀ1 enzyme subsite of Caulo-
bacter a-1,2-mannosidase, contoured at 1 kcalmolꢀ1. Symbols plot the
observed conformations of 1, 2, and 3.
with an initial C1 conformation as this is one of the most
preactivated conformations identified here (see the Support-
ing Information).[11]
To quantify how GH47 a-1,2-mannosidases restrict the
substrate conformational landscape, metadynamic calcula-
tions were repeated using the substrate a-1,2-mannobiose,
starting from the crystallographically-determined Michaelis
complex (the sulfur atom of the S-linked disaccharide 1 was
computationally replaced by oxygen). As shown in Figure 2b,
the enzyme acts to fundamentally restrict the energetically
accessible conformational landscape of the ꢀ1 a-mannosyl
0.5 kcalmolꢀ1 more stable and wider. This indicates a con-
formational equilibrium between these two structures in the
Michaelis complex; however, nucleophilic substitution can
only occur through 3,OB/3S1 conformations, as only in these
ꢀ
conformations is the C1 H bond in a pseudo-equatorial
4
residue. Of particular note, the undistorted C1 conformer is
orientation. Bonding differences associated with the sulfur/
oxygen linkage and aglycon interactions with the + 1 subsite
likely favor the 3,OB/3S1 distorted conformation observed for
1.
no longer an energy minimum. On enzyme, the 3,OB/3S1
conformations are the only stable distorted conformations,
3
1
defining a clear 3,OB/3S1! H4! C4 conformational pathway
for the reaction coordinate, Scheme 1. Both energy minima
(corresponding to 3,OB/3S1 and 1C4 conformations) are almost
of the same stability, although the former is approximately
The strategy of using an S-linked substrate mimic with
wild-type enzyme has been widely used in the conformational
study of the Michaelis complexes of glycoside hydrolases
Angew. Chem. Int. Ed. 2012, 51, 1 – 7
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
only for the transition state conformation, but also for the
conformational coordinate of catalysis. This may offer a direct
signature allowing assignment of the conformational itinerary
of d-manno-active enzymes in future structural analyzes. The
combination of atomic resolution structural analysis with
QM/MM methods allows unparalleled insight into the con-
formational pathway of a southern hemisphere a-mannosi-
dase. The narrowing (indeed, funneling) of the accessible
conformational free energy surface when the enzyme struc-
ture is imposed on the surface for free a-d-mannose indicates
a single reaction coordinate conformational progression for
the GH47 family; one consistent with known 3D structures of
this class of enzyme.
Experimental Section
Crystals of CkGH47 were grown as described in the Supporting
Information with 3D structures determined by molecular replace-
ment using data collected at beamlines I03, I04, and I041 of the
Diamond Light Source (UK). Further details on structure solution
and refinement are included in the PDB headers and the Supporting
Information. Mannosidase activity was determined both using
disaccharide substrates and a-mannosyl fluoride, as described in the
Supporting Information. a-mannosyl fluoride (ManF) was synthe-
sized by fluorination of per-O-acetylated mannose with HF/pyridine
followed by deprotection with catalytic sodium methoxide in
methanol, as described previously.[13]
Scheme 1. Reaction coordinate for inverting GH47 a-mannosidases[3b]
3
via a H4 transition state. Partial atom numbering is shown for
reference and Ca2+ is shown as a sphere.
Thermodynamic studies of inhibitor binding, which were deter-
mined at the optimal pH for catalysis (pH 7.0), were performed using
a MicroCal iTC200 calorimeter. Assays were carried out at 258C with
compounds 1, 2, and 3 (all at 500 mm) titrated into the ITC cell
containing 50 mm CkGH47. Dissociation constants (KD) were calcu-
lated for each assay using the Origin 7 software package (MicroCal).
Quantum mechanical calculations were performed using Density
Functional Theory-based molecular dynamics (MD), according to the
Car-Parrinello method.[14] The Kohn–Sham orbitals were expanded in
a plane wave (PW) basis set with a kinetic energy cutoff of 80 Ry. The
Perdew, Burke, and Ernzerhoff generalized gradient-corrected
approximation (PBE)[15] was selected in view of its good performance
in previous work on isolated sugars,[16] glycosidases,[17] and glycosyl-
transferases.[18] The metadynamics algorithm[19] was used to explore
the conformational free energy landscape of a-d-mannopyranose,
taking as collective variables two of the puckering coordinates of
Cremer and Pople[20] (q,f), in the spirit of the pioneering work by
Dowd, French, and Reilly.[21] QM/MM calculations of a-1,2-manno-
sidase in complex with a-1,2-d-mannobiose were performed using the
method developed by Laio et al.,[22] which combines Car-Parrinello
MD, with force-field MD. The a-1,2-d-mannobiose and the calcium
cation were treated quantum-mechanically, whereas the AMBER
force field[23] was used for the rest of the protein and the solvent. The
electrostatic interactions between the QM and MM regions were
handled by a fully Hamiltonian coupling scheme, wherein the short-
range electrostatic interactions between the QM and the MM regions
are explicitly taken into account for all atoms[22] (see the Supporting
Information for further details).
(GHs). The QM/MM optimized structure of the Michaelis
complex with the natural substrate is in excellent agreement
with the experimental X-ray structure of the Michaelis
complex with the S-linked disaccharide substrate mimic
1 (see the Supporting Information), and displays the reactive
conformation expected for an inverting GH, as well as a water
molecule ideally oriented for in-line nucleophilic attack on
the anomeric carbon.
Mannoimidazole 2 has special features that make it an
illuminating probe of mannosidase reaction coordinates,
including the presence of an sp2-hybridized (pseudo)anomeric
carbon, and the ability to easily adopt the flattened boat and
half-chair conformations expected for oxocarbenium-ion-like
transition states. For GH2 retaining b-mannosidases, 2 and its
derivatives have been observed in a B2,5 conformation, with
linear free energy relationships showing an almost perfect
correlation between the free energy of the binding of the
inhibitor (logKI) and the free energy of the binding of the
transition state (logKM/kcat).[12] This suggests that this con-
formation mimics that of the transition state, and supports
a 1S5 to B2,5 to OS2 progression for the first half of the reaction
coordinate. Similarly, 2 was bound in a B2,5 conformation to
a GH92 inverting a-mannosidase, which is indicative of
1
a OS2!B2,5! S5 conformational progression; this observation
Received: July 6, 2012
Published online: && &&, &&&&
is consistent with the reverse of that seen for the glycosylation
half-reaction of GH2.
Keywords: computational chemistry · conformation analysis ·
In summary, species 2 was found to bind in a 3H4
conformation on GH47, exactly as previously proposed for
the GH47 transition state. This is supported by both computa-
tional and structural approaches. The bound conformation
determined for this one compound alone is diagnostic not
.
enzyme mechanisms · glycosidase inhibitor · mannose
[1] D. Crich, Acc. Chem. Res. 2010, 43, 1144 – 1153.
4
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 7
These are not the final page numbers!

Angewandte
Chemie
[2] B. L. Cantarel, P. M. Coutinho, C. Rancurel, T. Bernard, V.
Lombard, B. Henrissat, Nucleic Acids Res. 2009, 37, D233 – 238.
[3] a) Y. Wu, M. T. Swulius, K. W. Moremen, R. N. Sifers, Proc. Natl.
Acad. Sci. USA 2003, 100, 8229 – 8234; b) K. Karaveg, A.
Siriwardena, W. Tempel, Z. J. Liu, J. Glushka, B. C. Wang,
K. W. Moremen, J. Biol. Chem. 2005, 280, 16197 – 16207.
[4] a) Y. D. Lobsanov, F. Vallee, A. Imberty, T. Yoshida, P. Yip, A.
Herscovics, P. L. Howell, J. Biol. Chem. 2002, 277, 5620 – 5630;
b) W. Tempel, K. Karaveg, Z.-J. Liu, J. Rose, B.-C. Wang, K. W.
Moremen, J. Biol. Chem. 2004, 279, 29774 – 29786.
[5] a) S. W. Mast, K. W. Moremen in Glycobiology, Vol. 415 (Ed.:
M. Fukuda), Elsevier, London, 2006, pp. 31 – 46; b) K. W. More-
men, M. Molinari, Curr. Opin. Struct. Biol. 2006, 16, 592 – 599.
[6] F. Vallꢁe, F. Lipari, P. Yip, B. Sleno, A. Herscovics, P. L. Howell,
EMBO J. 2000, 19, 581 – 588.
[7] F. Vallee, K. Karaveg, A. Herscovics, K. W. Moremen, P. L.
Howell, J. Biol. Chem. 2000, 275, 41287 – 41298.
[8] a) C. Mulakala, W. Nerinckx, P. J. Reilly, Carbohydr. Res. 2006,
341, 2233 – 2245; b) C. Mulakala, W. Nerinckx, P. J. Reilly,
Carbohydr. Res. 2007, 342, 163 – 169; c) D. Cantffl, W. Nerinckx,
P. J. Reilly, Carbohydr. Res. 2008, 343, 2235 – 2242.
[9] H. Liu, X. Liang, H. Sohoel, A. Bulow, M. Bols, J. Am. Chem.
Soc. 2001, 123, 5116 – 5117.
[10] For a review, see: G. J. Davies, A. Planas, C. Rovira, Acc. Chem.
Res. 2012, 45, 308 – 316.
[12] L. N. Tailford, W. A. Offen, N. L. Smith, C. Dumon, C. Morland,
J. Gratien, M.-P. Heck, R. V. Stick, Y. Blꢁriot, A. Vasella, H. J.
Gilbert, G. J. Davies, Nat. Chem. Biol. 2008, 4, 306 – 312.
[13] M. Hayashi, S.-I. Hashimoto, R. Noyori, Chem. Lett. 1984, 1747 –
1750.
[14] R. Car, M. Parrinello, Phys. Rev. Lett. 1985, 55, 2471 – 2474.
[15] J. P. Perdew, K. Burke, M. Ernzerhof, Phys. Rev. Lett. 1996, 77,
3865 – 3868.
[16] a) X. Biarnꢁs, A. Ard›vol, A. Planas, C. Rovira, A. Laio, M.
Parrinello, J. Am. Chem. Soc. 2007, 129, 10686 – 10693; b) A.
Ard›vol, X. Biarnꢁs, A. Planas, C. Rovira, J. Am. Chem. Soc.
2010, 132, 16058 – 16065.
[17] a) X. Biarnꢁs, J. Nieto, A. Planas, C. Rovira, J. Biol. Chem. 2006,
281, 1432 – 1441; b) X. Biarnꢁs, A. Ard›vol, J. Iglesias-Fernan-
dez, A. Planas, C. Rovira, J. Am. Chem. Soc. 2011, 133, 20301 –
20309; c) L. Petersen, A. Ard›vol, C. Rovira, P. J. Reilly, J. Am.
Chem. Soc. 2010, 132, 8291 – 8300; d) I. J. Barker, L. Petersen,
P. J. Reilly, J. Phys. Chem. B 2010, 114, 15389 – 15393.
[18] A. Ard›vol, C. Rovira, Angew. Chem. 2011, 123, 11089 – 11093;
Angew. Chem. Int. Ed. 2011, 50, 10897 – 10901.
[19] A. Laio, M. Parrinello, Proc. Natl. Acad. Sci. USA 2002, 99,
12562 – 12566.
[20] D. Cremer, J. A. Pople, J. Am. Chem. Soc. 1975, 97, 1354 – 1358.
[21] M. K. Dowd, A. D. French, P. J. Reilly, Carbohydr. Res. 1994,
264, 1 – 19.
[22] A. Laio, J. VandeVondele, U. Rothlisberger, J. Chem. Phys. 2002,
116, 6941 – 6947.
[23] D. A. Pearlman, D. A. Case, J. W. Caldwell, W. S. Ross, T. E.
Cheatham, S. Debolt, D. Ferguson, G. Seibel, P. Kollman,
Comput. Phys. Commun. 1995, 91, 1 – 41.
[11] K. J. Gregg, W. F. Zandberg, J. H. Hehemann, G. E. Whitworth,
L. H. Deng, D. J. Vocadlo, A. B. Boraston, J. Biol. Chem. 2011,
286, 15586 – 15596.
Angew. Chem. Int. Ed. 2012, 51, 1 – 7
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!

.
Angewandte
Communications
Communications
Computational Chemistry
A. J. Thompson, J. Dabin,
J. Iglesias-Fernꢁndez, A. Ardꢂvol,
Z. Dinev, S. J. Williams, O. Bande,
A. Siriwardena, C. Moreland, T.-C. Hu,
D. K. Smith, H. J. Gilbert, C. Rovira,*
G. J. Davies*
&&&& — &&&&
Mannosides in the southern hemisphere:
Conformational analysis of enzymatic
mannoside hydrolysis informs strategies
for enzyme inhibition and inspires solu-
tions to mannoside synthesis. Atomic
resolution structures along the reaction
coordinate of an inverting a-mannosi-
dase show how the enzyme distorts the
substrate and transition state. QM/MM
calculations reveal how the free energy
landscape of isolated a-d-mannose is
molded on enzyme to only allow one
conformationally accessible reaction
coordinate.
The Reaction Coordinate of a Bacterial
GH47 a-Mannosidase: A Combined
Quantum Mechanical and Structural
Approach
6
www.angewandte.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 1 – 7
These are not the final page numbers!

Angewandte
Chemie
Angew. Chem. Int. Ed. 2012, 51, 1 – 7
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7
These are not the final page numbers!