Download Full-Text PDF

molecules
Review
Neutron Crystallography for the Study of Hydrogen
Bonds in Macromolecules
Esko Oksanen 1,2 , Julian C.-H. Chen 3 and Suzanne Zoë Fisher 1,4, *
1
2
3
4
*
Science Directorate, European Spallation Source ERIC, Tunavägen 24, 22100 Lund, Sweden;
[email protected]
Department of Biochemistry and Structural Biology, Lund University, Sölvegatan 39, 22362 Lund, Sweden
Bioscience Division, Los Alamos National Laboratory, Los Alamos, NM 87545, USA; [email protected]
Department of Biology, Lund University, Sölvegatan 35, 22362 Lund, Sweden
Correspondence: [email protected]; Tel.: +46-72-179-2250
Academic Editor: Steve Scheiner
Received: 12 March 2017; Accepted: 1 April 2017; Published: 7 April 2017
Abstract: The hydrogen bond (H bond) is one of the most important interactions that form the
foundation of secondary and tertiary protein structure. Beyond holding protein structures together,
H bonds are also intimately involved in solvent coordination, ligand binding, and enzyme catalysis.
The H bond by definition involves the light atom, H, and it is very difficult to study directly, especially
with X-ray crystallographic techniques, due to the poor scattering power of H atoms. Neutron protein
crystallography provides a powerful, complementary tool that can give unambiguous information to
structural biologists on solvent organization and coordination, the electrostatics of ligand binding, the
protonation states of amino acid side chains and catalytic water species. The method is complementary
to X-ray crystallography and the dynamic data obtainable with NMR spectroscopy. Also, as it
gives explicit H atom positions, it can be very valuable to computational chemistry where exact
knowledge of protonation and solvent orientation can make a large difference in modeling. This article
gives general information about neutron crystallography and shows specific examples of how the
method has contributed to structural biology, structure-based drug design; and the understanding of
fundamental questions of reaction mechanisms.
Keywords: hydrogen bond; neutron crystallography; structural enzymology; nuclear density maps;
solvent networks
1. Hydrogen Bonds in Protein Structures
A hydrogen bond (H bond) is established through the sharing of an H atom that is covalently
attached to an electronegative donor (D), with a different electronegative acceptor atom (A), and can be
written as D-H . . . A. H bonds are one of the major drivers that participate in the correct folding and
organization of secondary and tertiary structural elements in proteins. In addition to their fundamental
role in structure, H bonds are also intimately involved in protein:ligand binding interactions and also
form the basis for most types of enzyme-catalyzed reactions [1]. Proteins contain many different kinds
of H bond donors and acceptors, from the backbone amide and carbonyl groups, solvent, to the side
chains of most amino acids residues. H bonds are directional, which is the driving force behind the
specificity of molecular recognition processes in ligand binding [2].
Studies in the 90s aimed at investigating the H bond frequency between H bond donors/acceptors
and have been carried out on a representative set of protein crystal structures available at the time.
It was observed that most donors and acceptors engage in H bonds, with only a small percentage
remaining as unsatisfied atoms. Many amino acid side chains can and do make multiple H bonds, e.g.,
Molecules 2017, 22, 596; doi:10.3390/molecules22040596
www.mdpi.com/journal/molecules
Molecules 2017, 22, 596
2 of 26
Lys, Arg, Asp, while others do not. For protein structure refinement and prediction calculations this is
an important consideration to include [3].
From a theoretical chemistry point of view, the H bond interaction energy can essentially be
described as being composed of a combination of covalent and weak interactions, e.g., van der Waals
forces [2,4]. The energy associated with an H bond is 2–10 kcal/mol. While this may seem insignificant
it is important to remember that these bonds are flexible and dynamic in that they are constantly
broken and reformed under a range of temperatures and physiological conditions [2]. There is quite
a large variation in H bond properties as observed in protein structures, especially in bond angles
when comparing secondary structure elements to side chain characteristics. The D . . . A distances
are typically between 1.7 and 2.4 Å, and the angle at the H atom can be anywhere from 130◦ to 170◦ ,
compared to ~150◦ at the acceptor [5]. As H bonds are difficult to directly observe, they are generally
inferred when the donor-acceptor distance is less than ~3.5 Å, and the angles at the donor/acceptor
are >90◦ [1]. Simple H atom sharing between two electronegative atoms represents the most common
type of H bond, however, it is now accepted that there are additional types of weaker H bonds; e.g.,
C-H groups as donors, N-H . . . S, and C-H . . . π interactions with aromatic amino acid residues as
acceptors [1,6]. A study conducted using NMR spectroscopy, DFT calculation, and data mining of
high resolution protein crystal structures revealed that C-H . . . π type interactions are in fact quite
common in proteins. These interactions are H bond-like in nature and occur between the methyl
groups of Ile, Leu, and Val and the delocalized π electrons in aromatic residues. The results showed
an average of four of these for every one hundred residues and are a functionally important part of
protein structures [7]. H bonds in macromolecules are challenging to observe and classify, accordingly
the criteria for identifying them in proteins are quite broad, as explained above. Another type of H
bond that is commonly found in protein structures is the bifurcated H bond. In this case the slightly
positively charged H atom is delocalized between two acceptor atoms. These are most often seen
between Ser/Thr side chains with Asp/Glu as acceptors [8]. This indicates that H bonds cannot be
identified only based on geometric criteria and that some additional structural or chemical knowledge
is needed to fully understand all the H bonds in a protein. In recent years, bifurcated H bonds have
also been directly observed in neutron protein crystal structures between Tyr and solvent, and Lys and
Thr [9,10].
H atoms are abundant in macromolecules, comprising roughly half of the elements that proteins
are composed of (C, H, N, O, S). However, despite their abundance and structural-functional
importance, including but not limited to H bonding, electrostatic interactions, solvent coordination
and solvent activation, H atoms and their interactions are very difficult to directly observe. This is
especially true for X-ray crystallographic studies where information regarding H bonds are usually
derived based on their heavy atom neighbor positions and the distances between them. The direct
determination of H atom positions in protein crystal structures is restricted mainly due to their low
X-ray scattering power, their inherent mobile nature, and limited crystallographic resolution [11].
2. Neutron Crystallography and Hydrogen Atoms
Neutron protein crystallography (NPX) is well-suited to the determination of the
three-dimensional structures of proteins due to their sensitivity to the light atom H and lack of radiation
damage at room temperature. NPX is conceptually very similar to standard X-ray crystallography,
and crystallographic resolution is understood in the same way. The main difference between the
methods lies in how neutrons are scattered from atoms. X-rays are scattered from the clouds of
electrons around atomic nuclei, and the magnitude depends on the atomic Z number leading to heavy
atoms scattering X-rays better than light atoms (Table 1). In practice this most often means that H
atoms are invisible, except for ultra-high resolution structures that are exceptionally well-ordered, i.e.,
display low crystallographic B factors. Neutrons are scattered from atomic nuclei and the magnitude
is independent of the Z number. Table 1 shows neutron scattering lengths for the most common atom
Molecules 2017, 22, 596
3 of 26
types found in proteins. Coherent scattering lengths of neutrons are very similar in magnitude for C,
N, 1 H (and 2 H or D), and O, making it as easy to observe C as H (or D) [11,12].
Table 1. Neutron and X-ray scattering lengths for atom types commonly found in proteins § .
Isotope
Z
Neutron
Scattering Length
(10−13 cm)
Incoherent
Cross Section
σi (10−28 m2 )
X-ray Scattering
Length (10−13 cm),
sin θ = 0
X-ray Scattering
Length (10−13 cm),
(sin θ)/λ = 0.5 Å−1
1H
1
1
6
7
8
16
−3.74
6.67
6.65
9.37
5.80
2.80
80.27
2.05
0.00
0.50
0.00
0.00
2.80
2.80
16.90
19.70
22.50
45.00
0.20
0.20
4.80
5.30
6.20
1.90
2H
(D)
12 C
14 N
16 O
32 S
§
Taken from [11,13].
In practice this means that even at medium (2.0–2.5 Å crystallographic resolution) it is routine to
readily observe the three dimensional positions of H atoms. By extension this means that neutrons are
also used to study H bonds [14]. The isotope of 1 H, deuterium (2 H or D), has superior neutron scattering
characteristics, making it optimal to replace H with D whenever possible (Table 1). The presence of
1 H presents several practical problems for crystallographic data collection. The large incoherent cross
section (Table 1) contributes to background as the incoherently scattered neutrons do not contribute to
measured Bragg reflections. In addition, the coherent scattering length is negative and leads to troughs
instead of peaks and can produce cancellation of positive peaks of adjacent atoms, complicating and
confusing analysis of nuclear density maps when 1 H is present. These effects can be minimized or
even eliminated by D exchange (partial deuteration) or expression of the target protein under fully
deuterated conditions (perdeuteration). Deuteration in general leads to higher quality maps and
enables less ambiguous data analysis and structure interpretation [11,13].
Out of all H atoms in a protein, ~25% are titratable and can be exchanged by vapor H/D exchange
or by soaking crystals in D2 O-containing solutions. These include H found in the bulk solvent, ordered
solvent, polar and charged amino acid side chains, and the protein backbone. Neutron structures
derived from H/D exchanged proteins represent the vast majority of deposited structures in the
PDB [11]. The remaining 75% of H atoms are non-exchangeable, typically found attached to carbons
and have to be incorporated as D during protein expression. This is routinely done by growing E. coli
cultures for protein expression in deuterated minimal media, such as M9, composed of stock solutions
and components dissolved in D2 O [11]. For the highest level D incorporation (~99%) it is best to use a
perdeuterated carbon source, e.g., perdeuterated glycerol or glucose. This is very costly however and
~85% incorporation can be achieved by using deuterated M9 minimal media but supplementing with
an unlabeled carbon source [15].
Cultures can be adapted in progressively increasing amounts of D2 O or grown to high densities
and switched to deuterated media just prior to induction of protein expression. All proteins are then
extracted and purified as usual, under hydrogenous conditions, with lost D atoms back-exchanged
at a later point. Depending on the strategy chosen the resulting protein will have varying degrees
and distribution of D incorporation. While perdeuteration is preferable from a theoretical point of
view, there are practical drawbacks that make it difficult to achieve in practice and the low numbers
of neutron structures determined from perdeuterated proteins reflects this. Although perdeuterated
proteins are nearly identical in structure to their hydrogenated counterparts, the proteins are often
found to have altered chemical properties, such as reduced solubility and stability. Deuterated cultures
also grow slowly, and perdeuterated protein yields are generally lower. Perdeuteration also introduces
the need to fine-screen or rescreen crystallization conditions under perdeuterated conditions [16].
As there is no radiation damage with neutrons used for protein crystallography applications,
it is routine to collect data at room temperature. This feature makes sample preparation for data
collection simple as crystals can be mounted in quartz capillaries and subjected to H/D exchange in
Molecules 2017, 22, 596
4 of 26
the
capillary prior to exposure to neutrons. To achieve this, crystals are mounted in capillaries4 and
Molecules 2017, 22, 596
of 25
a liquid plug of deuterated mother liquor is injected on either side of the crystal prior to sealing the
capillary.
capillary. The
The pitfalls
pitfalls of
of finding
finding protective
protective cryogenic
cryogenic solutions
solutions and
and the
the actual
actual freezing
freezing of
of crystals
crystals are
are
also
avoided
[11,16].
also avoided [11,16].
Globally
Globally there
there are
are only
only aa handful
handful of
of instruments
instruments available
available that
that are
are specifically
specifically optimized
optimized for
for
macromolecular
there
areare
currently
twotwo
instruments,
bothboth
located
at Oak
macromolecular crystallography.
crystallography.InInthe
theUSA
USA
there
currently
instruments,
located
at
Ridge
National
Laboratory
(ORNL,
Oak
Ridge,
TN,
USA).
The
IMAGINE
instrument
is
located
at
the
Oak Ridge National Laboratory (ORNL, Oak Ridge, TN, USA). The IMAGINE instrument is located
HFIR
source,
whilewhile
MaNDi
is found
at theatSpallation
Neutron
Source.
In Japan
there
is the
at the reactor
HFIR reactor
source,
MaNDi
is found
the Spallation
Neutron
Source.
In Japan
there
is
iBIX
instrument
at
the
Japan
Proton
Accelerator
Research
Complex
(J-PARC,
Tokai,
Japan).
In
Europe
the iBIX instrument at the Japan Proton Accelerator Research Complex (J-PARC, Tokai, Japan). In
there
arethere
currently
two instruments
operating
at reactor
neutron
sources
(Figure
1). Depending
on
Europe
are currently
two instruments
operating
at reactor
neutron
sources
(Figure
1). Depending
the
source
andand
how
thethe
instrument
is is
setsetup,
on the
source
how
instrument
up,data
datacan
canbe
becollected
collectedwith
with monochromated
monochromated (single
(single
wavelength)
method
or
with
a
white
Laue
(multi-wavelength)
method.
The
LADI-III
(ILL,
Grenoble,
wavelength) method or with a white Laue (multi-wavelength) method. The LADI-III (ILL, Grenoble,
France)
France) instrument
instrument uses
uses aa quasi-Laue
quasi-Laue spectrum,
spectrum, while
while BIODIFF
BIODIFF (MLZ,
(MLZ, Munich,
Munich, Germany)
Germany) uses
uses aa
monochromated
beam. However
However data
data is
is collected,
collected, the
the resulting
resulting neutron
neutron diffraction
diffraction data
data sets
monochromated beam.
sets from
from
different
instruments
are
remarkably
comparable
and
can
be
refined
using
standard
crystallographic
different instruments are remarkably comparable and can be refined using standard crystallographic
software
and
analysis
oneone
cancan
readily
observe
the
software packages.
packages. After
Aftercareful
carefulcrystallographic
crystallographicrefinement
refinement
and
analysis
readily
observe
three-dimensional
structural
details
of
H/D
atoms,
H
bonds,
water
molecules
and
their
interactions,
the three-dimensional structural details of H/D atoms, H bonds, water molecules and their
the
charged state
of aminostate
acidsofresidues
and their
interactions,
and
the details and
of ligand
binding.
interactions,
the charged
amino acids
residues
and their
interactions,
the details
of
All
this binding.
information
essential
for the study
of enzyme
andenzyme
drug binding,
making
neutron
ligand
Allisthis
information
is essential
for mechanism
the study of
mechanism
and
drug
protein
a powerful
tool in structural
biology [17,18].
binding,crystallography
making neutron
protein crystallography
a powerful
tool in structural biology [17,18].
Figure 1.1.Neutron
protein
crystallography
instrumentation.
(Left) photograph
of the LADI-III
Quasi-Laue
Figure
Neutron
protein
crystallography
instrumentation.
(Left) photograph
of the
LADI-III
diffractometer,
located
at
the
Institut
Laue-Langevin,
Grenoble
FR
[19];
(Right)
photograph
of the
Quasi-Laue diffractometer, located at the Institut Laue-Langevin, Grenoble FR [19]; (Right) photograph
BIODIFF
monochromatic
diffractometer,
located
at
Heinz
Maier-Leibnitz
Zentrum,
Munich
of the BIODIFF monochromatic diffractometer, located at Heinz Maier-Leibnitz Zentrum, Munich DE.
DE.
Re-used with
with permission
permission from
from the
the photographer,
photographer, Bernhard
BernhardLudewig.
Ludewig.
Re-used
In general the technique is still underutilized, with only ~100 PDB entries resulting from neutron
In general the technique is still underutilized, with only ~100 PDB entries resulting from neutron
diffraction studies. This is largely due to three major bottlenecks: (1) neutron scattering is a flux
diffraction studies. This is largely due to three major bottlenecks: (1) neutron scattering is a flux
limited technique and accordingly requires long data collections times, ~5–20 days depending on the
limited technique and accordingly requires long data collections times, ~5–20 days depending
on the
instrument; (2) quite large crystal volumes are required for most instruments (>0.5 mm33 on average);
instrument; (2) quite large crystal volumes are required for most instruments (>0.5 mm on average);
and (3) there are only a handful of instruments available worldwide to external users. Advances in
and (3) there are only a handful of instruments available worldwide to external users. Advances in
sources, beamlines, and detector technologies are easing the way forward with much shorter data
sources, beamlines, and detector technologies are easing the way forward with much shorter data
collection times from smaller crystals [11,16–18].
collection times from smaller crystals [11,16–18].
This review presents a number of topical examples of different kinds of structural and functional
This review presents a number of topical examples of different kinds of structural and functional
information that can be obtained with neutron crystallography, specifically how they relate to the
information that can be obtained with neutron crystallography, specifically how they relate to the study
study and importance of H bonds. Examples included are meant to span a broad area in structural
and importance of H bonds. Examples included are meant to span a broad area in structural biology
biology and include systems where ordered solvent and H bonded solvent network are intimately
linked to catalysis. Fluorescent and chromoproteins are also shown as excellent cases where optical
properties are solely determined by changes in H bonding around a chromophore. Several
enzymology examples are discussed in a later section, highlighting the importance of knowledge of
Molecules 2017, 22, 596
5 of 26
and include systems where ordered solvent and H bonded solvent network are intimately linked to
catalysis. Fluorescent and chromoproteins are also shown as excellent cases where optical properties
22, 596
5 of 25
areMolecules
solely 2017,
determined
by changes in H bonding around a chromophore. Several enzymology examples
are discussed in a later section, highlighting the importance of knowledge of H bonds and being able
H bonds and being able to determine how they change during catalysis. Due to the strong scattering
to determine how they change during catalysis. Due to the strong scattering of D, neutrons have also
of D, neutrons have also been used to “see” exotic solvent species such as the elusive hydronium
ion
been
used to “see” exotic solvent species such as the elusive hydronium ion (H O+ or D3 O+ ) an
and
(H3O+ or D3O+) and protons (H+ or D+ containing no electrons and therefore invisible3 with X-rays),
+ or D+ containing no electrons and therefore invisible with X-rays), an important proposed
protons
(H
important proposed partner in certain kinds of enzyme-catalyzed reactions. Concluding remarks
partner
certain kinds ofperspective
enzyme-catalyzed
reactions.
Concluding
remarks
offer a forward-looking
offer ainforward-looking
on the place
of neutrons
in current-day
structure
biology.
perspective on the place of neutrons in current-day structure biology.
3. Solvent Networks and H Bonding to Proteins
3. Solvent Networks and H Bonding to Proteins
3.1. Ultra High Resolution Neutron Crystallography
3.1. Ultra High Resolution Neutron Crystallography
Ultra-high resolution (1.1 Å) neutron diffraction data was collected on a large crystal of the small
Ultra-high
resolution (1.1 Å) neutron diffraction data was collected on a large crystal of the small
protein crambin (46 residues, 4.7 kDa), isolated from seeds of Crambe abyssinica. The crystal had been
protein
crambin
(46 residues,
4.7 kDa),
isolated
from
seeds
Crambe
abyssinica.
The high
crystal
had been
exchanged
against
fully deuterated
mother
liquor
prior
to of
data
collection.
The very
resolution
exchanged
against
fully
deuterated
mother
liquor
prior
to
data
collection.
The
very
high
resolution
of the data allowed for a nearly complete inventory (~95%) of all H and D atoms in the protein [20].
of Having
the dataunambiguous
allowed for astructural
nearly complete
inventory
(~95%)
of all
and D atoms
in the
protein [20].
positions
for the light
atoms
alsoHallowed
for easy
interpretation
Having
unambiguous
structural
positions
for
the
light
atoms
also
allowed
for
easy
interpretation
and
and orientation of solvent molecules, with the ability to directly observe H bonding networks
orientation
of
solvent
molecules,
with
the
ability
to
directly
observe
H
bonding
networks
(Figure
2).
(Figure 2).
Figure
2. Ultra-high
resolution
X-ray
and neutron
crystallographic
H/D exchanged
Figure
2. Ultra-high
resolution
X-ray
and neutron
crystallographic
studiesstudies
of H/Dofexchanged
crambin,
crambin, showing (A) details of possible H bonds between five ordered water molecules in crambin
showing (A) details of possible H bonds between five ordered water molecules in crambin as observed
as observed in the X-ray crystal structure; and (B) observed H bonds between water molecules as
in the X-ray crystal structure; and (B) observed H bonds between water molecules as seen in the
seen in the neutron crystal structure. Figure reproduced from [20].
neutron crystal structure. Figure reproduced from [20].
Even with the highest resolution X-ray structures, it remains difficult to visualize H atoms due
withZthe
highest
resolution
X-rayatoms,
structures,
to visualize
H atoms
due
to Even
their low
and
proximity
to heavier
and it
asremains
such, it difficult
is difficult
to accurately
model
Hto
their
low
Z
and
proximity
to
heavier
atoms,
and
as
such,
it
is
difficult
to
accurately
model
H
bonding
bonding networks. In the case of crambin there is a 0.48 Å resolution crystal structure of crambin
networks.
crambin
is a recently
0.48 Å resolution
crystal
structure
of crambin
available
availableIn
inthe
the case
PDBof
(3NIR),
andthere
also the
collected 0.38
Å X-ray
diffraction
data set
[21,22].in
theEven
PDBin(3NIR),
and also the recently
X-ray diffraction
set [21,22].
Even
these extraordinary
cases it collected
was still 0.38
onlyÅpossible
to observedata
around
two-thirds
ofin
allthese
H
extraordinary
still only
possible
to observe
two-thirds
H atoms
in the
atoms in thecases
Fo − Fitc was
difference
maps.
Inferring
H atomaround
positions
and the of
H all
bonds
they may
beFo
with maps.
can beInferring
misleading,
as exemplified
thethe
study
of crambin
(Figure
2A,B). There
− involved
Fc difference
H atom
positions in
and
H bonds
they may
be involved
withare
can
ordered water
molecules in
seen
ultra-high
resolution
X-ray2A,B).
crystalThere
structures
of crambin
befive
misleading,
as exemplified
theinstudy
of crambin
(Figure
are five
ordered with
water
very well-defined
oxygen atom
positions
but crystal
no definitive
information
on location
of the
H (or D)
molecules
seen in ultra-high
resolution
X-ray
structures
of crambin
with very
well-defined
atoms.
The
dashed
lines
in
Figure
2A
indicate
putative
H
bonds
between
these
waters.
Further
oxygen atom positions but no definitive information on location of the H (or D) atoms. The
dashed
inspection
of 2A
the indicate
nuclear density
maps
showed
that there
arewaters.
in fact only
two inspection
H bonds between
these
lines
in Figure
putative
H bonds
between
these
Further
of the nuclear
waters
(Figure
2B),
and
not
five
as
can
be
interpreted
based
on
the
X-ray
data
alone.
By
taking
density maps showed that there are in fact only two H bonds between these waters (Figure 2B), and
ofbe
theinterpreted
high resolution
the
data,
together
with the
scattering
exhibited
by
notadvantage
five as can
basedofon
the
X-ray
data alone.
By strong
taking neutron
advantage
of the high
resolution
D
atoms,
it
was
possible
to
model
the
anisotropic
vibrational
behavior
of
a
selected
number
of
D
of the data, together with the strong neutron scattering exhibited by D atoms, it was possible to model
atoms
in
crambin.
Qualitatively,
the
D
atoms
were
on
average
more
anisotropic
than
their
bonded,
the anisotropic vibrational behavior of a selected number of D atoms in crambin. Qualitatively, the D
neighboring heavy atom [20].
atoms were on average more anisotropic than their bonded, neighboring heavy atom [20].
3.2. H Bonded Water Networks in Proton Transfer
Human carbonic anhydrase II (HCA II) is one of 15 expressed isoforms found in humans. HCA II
(261 residues, ~29 kDa) is a Zn-dependent enzyme that catalyzes the reversible reaction of carbon
Molecules 2017, 22, 596
6 of 26
3.2. H Bonded Water Networks in Proton Transfer
Human carbonic anhydrase II (HCA II) is one of 15 expressed isoforms found in humans. HCA II
(261 residues, ~29 kDa) is a Zn-dependent enzyme that catalyzes the reversible reaction of carbon
dioxide hydration to form bicarbonate and a proton. In general, HCAs are found throughout the human
Molecules 2017, 22, 596
6 of 25
body, from blood to the brain, and are involved in many and diverse physiological reactions [23].
The HCA
active site
has bicarbonate
been studied
and defined
decades
of biochemical
dioxideIIhydration
to form
and aextensively
proton. In general,
HCAs areby
found
throughout
the human and
body, from
blood to the brain,
and are
involved
in many and
diverse physiological
reactions
biophysical
characterization,
including
X-ray
crystallography,
mutagenesis,
NMR, and
kinetics.[23].
HCA II
HCA
II active
site has
studied
defined
by
decades
of
biochemical
and
is oneThe
of the
fastest
enzymes
withbeen
a kcat
of 106 extensively
and kcat /Kand
approaching
the
diffusion
limit.
The
reaction
M
biophysical
characterization,
X-ray crystallography,
mutagenesis,
NMR,dioxide
and kinetics.
follows
a ping-pong
mechanismincluding
and is composed
of two half reactions:
carbon
hydration
HCA II is one of the fastest enzymes with a kcat of 106 and kcat/KM approaching the diffusion limit. The
followed by a proton transfer step. The rate-limiting proton transfer portion of the catalysis is
reaction follows a ping-pong mechanism and is composed of two half reactions: carbon dioxide
mediated by a very well-ordered H-bonded water network that spans over ~8 Å and connects catalytic
hydration followed by a proton transfer step. The rate-limiting proton transfer portion of the
(Zn-coordinated)
water to the bulk solvent via a proton shuttling residue, His64. The active site
catalysis is mediated by a very well-ordered H-bonded water network that spans over ~8 Å and
cavity
is
lined
with
several
hydrophilic
amino
acid
residues
for maintaining
properHis64.
active site
connects catalytic
(Zn-coordinated)
water
to the
bulk
solventcrucial
via a proton
shuttling residue,
chemistry,
geometry,
and
for
coordinating
the
solvent
network.
The
result
is
a
well-tuned
environment
The active site cavity is lined with several hydrophilic amino acid residues crucial for maintaining
that allows
and
efficientgeometry,
proton transfer
proper fast
active
sitevery
chemistry,
and for [23].
coordinating the solvent network. The result is a
well-tunedstudies
environment
thatIIallows
fastpH
and (pH
very 10)
efficient
proton
[23]. details of the ordered
Neutron
of HCA
at high
showed
thetransfer
H bonding
Neutron studies
of HCA
at high
pH (pHnon-physiological
10) showed the H bonding
details
of the
the organization
ordered
solvent network.
Consistent
withIIthe
decidedly
pH of the
study,
solvent
network.
Consistent
with
the
decidedly
non-physiological
pH
of
the
study,
the
organization
and H bonding in the solvent network was such that it could not support proton transfer between
and H bonding in the solvent network was such that it could not support proton transfer between
catalytic
water and the proton shuttling residue His64. In addition it was observed that the side chain
catalytic water and the proton shuttling residue His64. In addition it was observed that the side
of Tyr7, one of the residues that line the active site and coordinate the water network, was deprotonated
chain of Tyr7, one of the residues that line the active site and coordinate the water network, was
(Figure
3a) [24]. This
was3a)
surprising,
evensurprising,
though the
pH
of thethe
crystallization
was over 10,
deprotonated
(Figure
[24]. This was
even
though
pH of the crystallization
wasas the
theoretical
pK
of
Tyr
in
solution
is
~11.
Tyr7
was
the
only
deprotonated
Tyr
and
all
seven
other
over 10, asa the theoretical pKa of Tyr in solution is ~11. Tyr7 was the only deprotonated Tyr and all Tyr
residues
were
as expected.
seven
otherneutral,
Tyr residues
were neutral, as expected.
Figure
3. Three
different
statesofofTyr7
Tyr7in
inhuman
human carbonic
II. II.
(a) (a)
Tyr7
is deprotonated
at
Figure
3. Three
different
states
carbonicanhydrase
anhydrase
Tyr7
is deprotonated
at
high pH; (b) Tyr7 is neutral at neutral pH but involved in a bifurcated H bond; (c) at pH 6.0 Tyr7 is
high pH; (b) Tyr7 is neutral at neutral pH but involved in a bifurcated H bond; (c) at pH 6.0 Tyr7 is still
still neutral and involved in a conventional H bond, participating as both acceptor and donor. 2Fo − Fc
neutral and involved in a conventional H bond, participating as both acceptor and donor. 2Fo − Fc
nuclear density maps are shown in magenta and are contoured at 1.3σ level. PDB IDs used (a) 3kkx,
nuclear density maps are shown in magenta and are contoured at 1.3σ level. PDB IDs used (a) 3kkx,
(b) 3tmj, and (c) 4y0j.
(b) 3tmj, and (c) 4y0j.
Further neutron structures were solved at pH 7.8 and 6.0 to further probe the solvent network
and
changes
in protonation
Tyr7solved
(Figureat3b,c)
The
7.8 neutron
structure
showed
a
Further neutron
structuresofwere
pH [9,24].
7.8 and
6.0pH
to further
probe
the solvent
network
rearrangement
of
the
solvent
network,
with
all
the
waters
now
connected
by
H
bonds,
and
neutral
and changes in protonation of Tyr7 (Figure 3b,c) [9,24]. The pH 7.8 neutron structure showed a
His64 poised to accept a proton. This pH “switch” indicated the pathway that an excess H+ can use to
rearrangement of the solvent network, with all the waters now connected by H bonds, and neutral
leave the active site. At pH 7.8 Tyr7 was neutral and involved in a bifurcated H bond with two
His64 poised to accept a proton. This pH “switch” indicated the pathway that an excess H+ can use
solvent molecules (Figure 3b). At pH 6.0 the water network was unchanged compared to the neutral
to leave
the
active
At pH 7.8
neutral
involved
in aconventional
bifurcated H
bond with two
form,
but
Tyr7 site.
had changed
howTyr7
it Hwas
bonds
to the and
solvent
to a more
configuration
solvent
molecules
(Figure
3b).
At
pH
6.0
the
water
network
was
unchanged
compared
(Figure 3c). Another key observation was that of the proton shuttling residue, His64. At to
pHthe
6.0neutral
it
form,was
butdirectly
Tyr7 had
changed
how
it
H
bonds
to
the
solvent
to
a
more
conventional
configuration
observed that His64 was charged (doubly protonated) and flipped away from the active
(Figure
key aobservation
was that with
of the
proton
shuttling
residue,
His64.
At pH
site3c).
and Another
involved with
π-stacking interaction
Trp5.
Comparison
of the
low pH
form with
the 6.0 it
high resolution
X-ray
crystal
of HCA
II in complex
with substrate
carbon
dioxide,
wasactive
was directly
observed
that
His64structure
was charged
(doubly
protonated)
and flipped
away
fromitthe
seeninvolved
that the active
appeared interaction
similar in solvent
positions
and amino of
acid
orientations.
site and
with sites
a π-stacking
with Trp5.
Comparison
theresidue
low pH
form with the
This suggests that protonating the active site by lowering the pH provides a glimpse of how the
active site is poised to accept substrate.
Molecules 2017, 22, 596
7 of 26
high resolution X-ray crystal structure of HCA II in complex with substrate carbon dioxide, it was
seen that the active sites appeared similar in solvent positions and amino acid residue orientations.
This suggests that protonating the active site by lowering the pH provides a glimpse of how the active
site is poised to accept substrate.
The first neutron structural studies were followed up with NMR titration studies of 13 C-Tyr labeled
HCA II as a function of pH. It was possible to assign chemical shifts for all eight Tyr residues and
the data confirmed the observation of a perturbed pKa for Tyr7. Consistent with the crystallographic
data Tyr7 had a severely perturbed pKa at ~7, compared to ~10–12 for the seven other Tyr residues.
The combined results of the high pH (pH 10), neutral pH (pH 7.8) and low pH (pH 6) forms revealed
important details about proton transfer, amino acid residues protonation and H bonding, and how
ordered solvent networks can respond and adapt to changing conditions [9]. These structural features
could never be observed with X-ray crystallography alone as the ~1 Å X-ray structure did not indicate
any explicit H atom positions, even on well-ordered residues. This study used pH as a variable
parameter to mimic events that could occur during catalysis. Analysis of these data suggests an
alternate catalytic model for how HCA II works. With a pKa in the physiological range Tyr7 may very
well be part of catalysis under normal pH conditions.
3.3. Chromoprotein Dathail and H Bonding
Fluorescent proteins (FP) occur naturally and display different fluorescent characteristics and
are widely used in biological research as probes and reporters. Protein engineering, mutagenesis,
and directed evolution have dramatically increased the variety and spectral properties of FPs. Green
fluorescent protein (GFP) has a β-barrel structure made of 11 β strands with the chromophore located
in a central pocket. The chromophore is formed after expression by the cyclization of 3 amino acids
to form p-hydroxybenzylidenimidazolinone. The pocket around the chromophore is lined with
amino acid side chains and ordered water molecules; mutagenesis can dramatically affect the spectral
properties of variants. Some of these chromoproteins can photoswitch between fluorescent and
non-fluorescent states by simple illumination by a specific wavelength of light. From X-ray analysis
it is thought that switching from the fluorescent state is related to cis-to-trans isomerization of the
chromophore, with cis corresponding to the ground state (fluorescent). Due to the structural rigidity of
FPs it has been suggested that H bonds and related chemical environment around the chromophore
play a fundamental role [25].
Neutron, X-ray and spectroscopic studies of a newly engineered FP, Dathail, revealed some
interesting and unusual photoswitching behavior. Dathail has very high temperature stability, is less
prone to aggregation, and readily crystallizes. Even more interesting is that the crystals tolerate
photoswitching, allowing data to be collected of the “dark” (ground) and “light’ (metastable) states.
Dathail is negative switching, with highest fluorescence in the ground state that is turned off when
exposed to 497 nm light. Fluorescence emission spectra also showed these two states. The protein
returns to the dark state through a very slow thermal relaxation process which takes several hours at
10 ◦ C [26]. X-ray crystal structures of the dark state showed that the chromophore is very well ordered
in the cis conformation. There was a dramatic difference in the light state where the chromophore,
surrounding amino acid residues, and solvent were very mobile and disordered in this form.
X-ray structures alone did not provide insights into the photoswitching behavior of Dathail.
The structure of the Dathail in the ground state is remarkably similar to those of other FPs, especially
with regards to the chromophore region. The observed structural configuration is associated with
fluorescence in other chromoproteins, however, Dathail is non-fluorescent in this form. As the
differences seem to lie in H bonding patterns, neutron crystallographic data were collected from
a large 10 mm3 H/D exchanged crystal of the dark (ground) state [26]. The fact that the ground-state
chromophore of Dathail is clearly protonated and stable with the chromophore in the cis conformation
challenges the consensus view that a stable, co-planar chromophore in the cis isomer leads to stable
fluorescence, given that Dathail is essentially non-fluorescent.
Molecules 2017, 22, 596
8 of 26
Neutron studies showed that the chromophore is protonated in the ground state and showed
further details of water molecules, H bonds, and protonation states of certain residues in the
chromophore pocket. Furthermore, there is an H bonded water network from the chromophore
to the bulk solvent that may be involved with the transfer of excited state protons out of the pocket
during photo-activation and helps to explain the lack of fluorescence in the ground state.
The
precise
details
Molecules
2017,
22, 596 of the water network around the chromophore are unique to Dathail.
8 of 25 This is
due to the presence of a Ser to Pro mutation at position 142. The waters mediate the de- and
The of
precise
details of the water
around
the chromophore are
are unique
to Dathail.
This to
is the Ser
re-protonation
the chromophore.
In network
other FPs
the chromophores
thought
to H bond
due to the presence of a Ser to Pro mutation at position 142. The waters mediate the de- and
(or Thr) residue in this position, making it difficult to protonate. From these neutron studies it was also
re-protonation of the chromophore. In other FPs the chromophores are thought to H bond to the Ser
seen that
is neutral
the ground
state.
His193
an H bond
acceptor
from studies
a waterit molecule
in
(orHis193
Thr) residue
in thisinposition,
making
it difficult
to is
protonate.
From
these neutron
was
the network,
andthat
is in
turn ais donor
A Gly
atHis193
position
144Haffects
how His193
coordinated,
also seen
His193
neutralto
inGlu211.
the ground
state.
is an
bond acceptor
from ais water
especially
compared
to otherand
chromoproteins
there
is aatGlu
at this
imidazole
molecule
in the network,
is in turn a donorwhere
to Glu211.
A Gly
position
144position.
affects howThe
His193
coordinated,
especially
compared
to other chromoproteins
where
is chromoproteins
a Glu at this position.
ring of is
His193
is π-stacked
with
the chromophore
itself, similar
tothere
other
(Figure 4).
The
imidazole
ring
of
His193
is
π-stacked
with
the
chromophore
itself,
similar
to
other
chromoproteins
However, the loss of a H bond due to a Glu to Gly mutation makes His193 more dynamic and less able
(Figure 4). However, the loss of a H bond due to a Glu to Gly mutation makes His193 more dynamic
to order the chromophore. In this study neutrons provided valuable insights into the chromophore
and less able to order the chromophore. In this study neutrons provided valuable insights into the
pocket organization,
the role
of solvent
and
helpedand
identify
H bonding
interactions
chromophore pocket
organization,
the molecules,
role of solvent
molecules,
helped key
identify
key H bonding
that areinteractions
at the heart
ofare
this
behavior [26].
that
at unique
the heartphotoswitching
of this unique photoswitching
behavior [26].
4. The
chromophore pocket
of Dathail
shows interactions
to amino acidto
residues
andacid
organization
Figure Figure
4. The
chromophore
pocket
of Dathail
shows interactions
amino
residues and
and orientation
of key water
molecules.
generated
fromgenerated
PDB 4ebj [26].
organization
and orientation
of key
water Figure
molecules.
Figure
from PDB 4ebj [26].
4. Unusual H Bonds and Solvent Species as Observed in Neutron Crystal Structures
4. Unusual H Bonds and Solvent Species as Observed in Neutron Crystal Structures
4.1. Photoactive Yellow Protein: Short H Bonds vs. Low-Barrier H Bonds
4.1. Photoactive Yellow Protein: Short H Bonds vs. Low-Barrier H Bonds
Photoactive yellow protein (PYP) is a soluble 14 kDa protein from Halorhodospira halophila that
absorbs
light
via itsprotein
para-coumaric
(pCA) chromophore.
The absorption
spectra for PYP
is
Photoactive
yellow
(PYP) acid
is a soluble
14 kDa protein
from Halorhodospira
halophila
that
equivalent
to
the
negative
photo-dependent
response
of
the
organism,
implicating
PYP
in
its
absorbs light via its para-coumaric acid (pCA) chromophore. The absorption spectra for PYP is
biological behavior. PYP has a reversible photocycle during which pCA undergoes trans-to-cis
equivalent to the negative photo-dependent response of the organism, implicating PYP in its biological
isomerization, causing remodeling of the H bonded network and proton transfer processes.
behavior.
PYP has aofreversible
photocycle
which
pCA
undergoesregions
trans-to-cis
isomerization,
Isomerization
pCA culminates
in largeduring
structural
changes
of N-terminal
of the protein.
causingThis
remodeling
of
the
H
bonded
network
and
proton
transfer
processes.
Isomerization
means that H bonding is essential for photocycle kinetics in the protein. In the ground state of pCA
culminates
structural
changes
of N-terminal
ofoxygen
the protein.
This
meansTyr42
that and
H bonding
pCAin
is large
stabilized
by two H
bonds between
the pCAregions
phenolate
atom and
residues
Glu46.
High-resolution
X-ray
structural
studies
revealed
these
bonds
to
be
rather
short
at
2.49
Å
and
is essential for photocycle kinetics in the protein. In the ground state pCA is stabilized by two H
2.58 Å, respectively.
It was suggested
thatatom
these and
bonds
represent
short and
H bonds
(SHB)
[27].
bonds between
the pCA phenolate
oxygen
residues
Tyr42
Glu46.
High-resolution
X-ray
As the H bonds become shorter there is a lengthening of the donor O-H covalent bond. The barrier
structural studies revealed these bonds to be rather short at 2.49 Å and 2.58 Å, respectively. It was
decreases as the disorder over two protonation sites collapse to a centered single-well potential, the
suggested
that these
bonds
short
HLBHB
bondsthe(SHB)
so-called
low-barrier
H represent
bond (LBHB).
In an
proton[27].
is shared between donor and acceptor
Asatoms,
the H making
bonds become
shorter
there
is
a
lengthening
of the donor O-H covalent bond. The barrier
it behave like a covalent bond [28].
crystallography
was
used to study
PYPcollapse
in ground
in ordersingle-well
to investigatepotential,
these
decreases asNeutron
the disorder
over two
protonation
sites
tostate
a centered
the
short
H bonds and
elucidate
H atom
in the
the area
of the
core [28,29].
The
firstacceptor
so-called
low-barrier
H to
bond
(LBHB).
In positions
an LBHB
proton
ishydrophobic
shared between
donor
and
neutron crystal structure of PYP was determined to 2.5 Å resolution, from a ~0.8 mm3 crystal that
atoms, making
it behave like a covalent bond [28].
was H/D exchanged. From this first structure it was possible to elucidate some of the details around
the pCA pocket, with pCA being deprotonated and the phenolate oxygen accepting H bonds from
Glu46 and Tyr42. Tyr42 in turn is H bonded to Arg52. The observed H bond lengths to pCA were
Molecules 2017, 22, 596
9 of 26
Neutron crystallography was used to study PYP in ground state in order to investigate these
short H bonds and to elucidate H atom positions in the area of the hydrophobic core [28,29]. The first
neutron crystal structure of PYP was determined to 2.5 Å resolution, from a ~0.8 mm3 crystal that was
H/D exchanged. From this first structure it was possible to elucidate some of the details around the
pCA pocket, with pCA being deprotonated and the phenolate oxygen accepting H bonds from Glu46
Molecules 2017, 22, 596
of 25
and Tyr42.
Tyr42 in turn is H bonded to Arg52. The observed H bond lengths to pCA were 9consistent
with the high-resolution X-ray work, remarkable given the moderate resolution (2.5 Å) of the neutron
consistent with the high-resolution X-ray work, remarkable given the moderate resolution (2.5 Å) of
structure
(Figure
5) [29]. (Figure 5) [29].
the neutron
structure
Figure
5. Nuclear density maps around the pCA of PYP. (Left) H bonds around the pCA in PYP as
Figure 5. Nuclear density maps around the pCA of PYP. (Left) H bonds around the pCA in PYP as
observed
in the
2.52.5
Å structure
withpermission
permission
from
International
of
observed
in the
Å structure[29].
[29].Reproduced
Reproduced with
from
the the
International
UnionUnion
of
Crystallography;
(Right)
High
resolution
neutron
studies
to
1.5
Å
enable
accurate
determination
of
Crystallography; (Right) High resolution neutron studies to 1.5 Å enable accurate determination of
H/DH/D
atom
positions
and
assignment
of
LBHB
in
PYP.
Figure
reproduced
from
[28].
atom positions and assignment of LBHB in PYP. Figure reproduced from [28].
In 2009 a much higher resolution neutron structure of PYP was reported, to 1.5 Å resolution.
In 2009 a much higher resolution neutron structure of PYP was reported, to 1.5 Å resolution.
Now it was possible to fully investigate the short H bonds around pCA to see if they could
Now it was possible to fully investigate the short H bonds around pCA to see if they could correspond
correspond to LBHB [28]. From detailed investigations of omit Fo − Fc and 2Fo − Fc nuclear density
to LBHB
From
detailed
of omit
2Fao LBHB,
− Fc nuclear
density
mapsHit was
o − Fcould
c andbe
maps[28].
it was
found
that theinvestigations
SHB between Glu46
andFpCA
while the
Tyr42-pCA
found
thatwas
thebetter
SHB between
andionic
pCAHcould
be a LBHB, while the Tyr42-pCA H bond was better
bond
classified Glu46
as a short
bond (SIHB).
classifiedItas
short ionic
H bond
(SIHB).
is athought
that the
deprotonated/charged
pCA represents an isolated charge in the otherwise
hydrophobic
ofdeprotonated/charged
PYP with Arg52 acts aspCA
a counter-ion.
in charge
the high-resolution
It is thoughtpocket
that the
representsHowever,
an isolated
in the otherwise
neutron
structure
this
Arg
was
neutral.
It
was
postulated
that
the
SIHB
and
LBHB
compensate for
hydrophobic pocket of PYP with Arg52 acts as a counter-ion. However, in the high-resolution
neutron
this and
provide
structural
stability
PYP. LBHBsthat
havethe
been
implicated
in other
systems forfor
being
structure
this
Arg was
neutral.
It wastopostulated
SIHB
and LBHB
compensate
this and
part of catalysis and/or stabilization of intermediates. In this case the proposed LBHB may stabilize
provide
structural stability to PYP. LBHBs have been implicated in other systems for being part of
an isolated charge and participate in proton transfer during the photocycle of PYP [28]. Comparison
catalysis and/or stabilization of intermediates. In this case the proposed LBHB may stabilize an
of the two studies also shows, unsurprisingly, that higher resolution neutron crystallographic data
isolated
charge and participate in proton transfer during the photocycle of PYP [28]. Comparison of
allows for more accurate and detailed studies.
the two studies also shows, unsurprisingly, that higher resolution neutron crystallographic data allows
for more
accurate
and detailed
studies.
4.2. Unusual
H Bonds
as Seen with
Neutrons
Recent neutron structures have revealed a wide variety of H bonding details, including LBHBs,
4.2. Unusual
H Bonds as Seen with Neutrons
SIHBs, and resonance-assisted H bonds. Despite over 40 years of structural, mechanistic, and
computational
work,
there remains
much atowide
understand
this fundamental,
yet varied
Recent neutron
structures
have revealed
variety about
of H bonding
details, including
LBHBs,
interaction,
found in protein structures
and enzyme
SIHBs,
and resonance-assisted
H bonds.
Despitemechanisms.
over 40 years of structural, mechanistic, and
To study
whether
LBHB was
involved
in catalysis,about
neutron
studies
were carried
on a interaction,
serine
computational
work,
therea remains
much
to understand
this
fundamental,
yetout
varied
protease,
porcine
pancreatic
elastase
(PPE),
in
complex
with
a
peptidic
inhibitor
FR130180
[30].
found in protein structures and enzyme mechanisms.
The residues His57-Asp102-Ser195 compose the catalytic triad and are conserved in the serine
To study whether a LBHB was involved in catalysis, neutron studies were carried out on a
proteases. In the first step of catalysis, the -OD1 of Ser195 makes a nucleophilic attack on a substrate
serine protease, porcine pancreatic elastase (PPE), in complex with a peptidic inhibitor FR130180 [30].
carbonyl group. Next, a tetrahedral intermediate is formed through a covalent bond to Ser195 and
The residues
His57-Asp102-Ser195
the catalytic
triad and
conserved
in theelectrostatic
serine proteases.
the substrate
carbonyl group. compose
The tetrahedral
intermediate
is are
stabilized
through
In the
first step that
of catalysis,
-OD1
Ser195amides
makesof
a nucleophilic
attackTogether
on a substrate
carbonyl
interactions
involve Hthe
bonds
to of
backbone
Ser195 and Gly193.
this forms
a
group.
Next, aoxyanion
tetrahedral
intermediate
is formed
through a covalent
bond toa Ser195
and the side
substrate
proposed
hole.
Ser195 is activated
for nucleophilic
attack through
SSHB between
chainsgroup.
of His57The
andtetrahedral
Asp102 [31].intermediate
The tetrahedralisintermediate
is resolved
through proton
transfer that
carbonyl
stabilized through
electrostatic
interactions
from
His57
to
the
leaving
group,
leaving
an
acyl-enzyme
intermediate.
The
H
bond
between
Ser195,
involve H bonds to backbone amides of Ser195 and Gly193. Together this forms a proposed
oxyanion
His57, and Asp102 has been proposed to be a LBHB and has been investigated with high resolution
X-ray crystallography and NMR spectroscopy, in different types of proteases such as trypsin and
subtilisin. The studies were inconclusive and in some cases instead showed a SIHB where the short
Molecules 2017, 22, 596
10 of 26
hole. Ser195 is activated for nucleophilic attack through a SSHB between side chains of His57 and
Asp102 [31]. The tetrahedral intermediate is resolved through proton transfer from His57 to the leaving
group, leaving an acyl-enzyme intermediate. The H bond between Ser195, His57, and Asp102 has
been proposed to be a LBHB and has been investigated with high resolution X-ray crystallography
and NMR spectroscopy, in different types of proteases such as trypsin and subtilisin. The studies
were inconclusive and in some cases instead showed a SIHB where the short H bond (~2.6 Å) is
Molecules 2017, 22, 596
10 of 25
observed. However, the shared H atom remains with the donor, and is not found equidistant between
donor/acceptor
in aHowever,
LBHB [32,33].
H bond (~2.6as
Å)expected
is observed.
the shared H atom remains with the donor, and is not found
FR130180
to donor/acceptor
PPE with theassame
tetrahedral
transition
equidistantbinds
between
expected
in a LBHB
[32,33]. state intermediate as seen with
binds
to PPEmaps
with show
the same
transition state
intermediate
as seen in
with
substrate.FR130180
The nuclear
density
thattetrahedral
His57 is protonated
(charged)
and involved
a short
substrate.
The
nuclear
density
maps
show
that
His57
is
protonated
(charged)
and
involved
in
a
short
H bond with Asp102, however the shared D density is clearly located on -ND1 of His57, and is most
Asp102, however
the shared
D densitythe
is clearly
located
on -ND1
His57,
and is is
most
likelyHa bond
SIHB.with
In addition
the neutron
data showed
carboxyl
oxygen
(O32)ofof
FR130180
present
likely
a
SIHB.
In
addition
the
neutron
data
showed
the
carboxyl
oxygen
(O32)
of
FR130180
is
present
as an oxygen anion (oxyanion) and was the first oxyanion in a tetrahedral intermediate observed at an
as an oxygen anion (oxyanion) and was the first oxyanion in a tetrahedral intermediate observed at
oxyanion hole. At a substrate binding subsite, S4, a π-cation interaction is observed between Arg217
an oxyanion hole. At a substrate binding subsite, S4, a π-cation interaction is observed between
and the benzoyl ring of FR130180. Finally, there is an H bond between the –OH group of Thr175 and
Arg217 and the benzoyl ring of FR130180. Finally, there is an H bond between the –OH group of
the carboxyl
group
benzoic
acidofinbenzoic
the inhibitor.
In this
interaction,
the carboxyl
is protonated
Thr175 and
the of
carboxyl
group
acid in the
inhibitor.
In this interaction,
thegroup
carboxyl
group
and the
O-D
distance
is
longer
than
expected
(1.21
vs.
1.02
Å),
while
being
only
1.41
Å
is protonated and the O-D distance is longer than expected (1.21 vs. 1.02 Å), while being onlyaway
1.41 Åfrom
the -OD1
Thr175.
This
is consistent
with another
of very
strong
H bond,
awayof
from
the -OD1
ofconfiguration
Thr175. This configuration
is consistent
with type
another
type of
very strong
H the
bond, assisted
the resonance
assisted
H bond
(RAHB) [30].
resonance
H bond
(RAHB)
[30].
The ultra-high
resolution
(1.1Å)Å)ofofthe
thecrambin
crambin neutron
made
it possible
to clearly
The ultra-high
resolution
(1.1
neutronstructure
structure
made
it possible
to clearly
resolve
H
atom
positions
and
the
degree
of
exchange
with
D
[20].
More
accurately
than
in
other
resolve H atom positions and the degree of exchange with D [20]. More accurately than in other
neutron structures resolved to date, the relative occupancies of H/D at exchangeable positions could
neutron
structures resolved to date, the relative occupancies of H/D at exchangeable positions could
be refined, and solvent molecules and side chains could be identified and oriented. Of note were two
be refined, and solvent molecules and side chains could be identified and oriented. Of note were two
particular regions of the protein with unexpected characteristics. In Gly31, the two Hs appear to
particular
regions of the protein with unexpected characteristics. In Gly31, the two Hs appear to have
have different degrees of exchange, based on the observed nuclear density for the two H atoms
different
degrees
(Figure 6). of exchange, based on the observed nuclear density for the two H atoms (Figure 6).
Figure 6. Nuclear 2Fo − Fc density for crambin residue Gly31, with negative (pink) density for H and
Figure 6. Nuclear 2Fo − Fc density for crambin residue Gly31, with negative (pink) density for H and
positive (green) for other atoms, showing different degrees of H/D exchange for the two alpha
positive (green) for other atoms, showing different degrees of H/D exchange for the two alpha carbons.
carbons. Figure reproduced from [20].
Figure reproduced from [20].
One interpretation of this is that there was partial exchange of the Cα-H against the deuterated
solvent,
leading to partial
possibility
of Cα-H…O
H bonding.
can be
One
interpretation
of thisDisoccupancy,
that there and
was the
partial
exchange
of the Cα-H
againstThis
the deuterated
attributed
to
differing
chemical
environments
in
the
crystal,
corroborated
by
solution
state
NMR
studies.
solvent, leading to partial D occupancy, and the possibility of Cα-H . . . O H bonding. This can be
two Cα-H
atoms inenvironments
Gly31 were observed
to havecorroborated
distinct NMR
chemical
shifts studies.
of
attributedThe
to differing
chemical
in the crystal,
byproton
solution
state NMR
3.96 and 5.56 ppm in solution, reflecting different chemical environments for the two alpha
The two Cα-H atoms in Gly31 were observed to have distinct NMR proton chemical shifts of 3.96
hydrogens [34]. Since Gly does not have a side chain, they are relatively strong carbon acids (pKa ~ 22),
and 5.56 ppm in solution, reflecting different chemical environments for the two alpha hydrogens [34].
comparable to the side chain hydroxyl groups of Ser and Thr (pKa ~ 18–20).
Since Gly Another
does notseries
haveofapotential
side chain,
they are
relatively
carbon
(pKregion
comparable
a ~22),of
Cα-H…O
H bonds
werestrong
observed
in theacids
β-strand
crambin, to
the side
hydroxyl
groupsoxygen
of Ser of
and
Thrand
(pKthe
a ~18–20).
e.g.,chain
between
the carbonyl
Thr1
Cα-H of Ile34. These observations add to the
Another
series
of
potential
Cα-H
.
.
.
O
H
bonds
were
in the β-strand
of crambin,
evidence that these weak, but numerous, stabilizing H observed
bond interactions
serve toregion
contribute
to
e.g., between
the
carbonyl
oxygen
of
Thr1
and
the
Cα-H
of
Ile34.
These
observations
add to the
secondary and tertiary structure stability. Secondly, Tyr44 in crambin is a well-ordered surface
residue
is solvated
by anumerous,
network of stabilizing
waters on the
one side,
and makesserve
a vantoder
Waals to
evidence
thatthat
these
weak, but
H bond
interactions
contribute
interaction with Ile33 on the other side. As shown in Figure 7, one of these ordered water molecules,
W1023, has its D1 atom pointing towards the center of the aromatic ring, indicating a potential
O-H…π H bonding interaction [20]. With improvements in X-ray and neutron instrumentation and
Molecules 2017, 22, 596
11 of 26
secondary and tertiary structure stability. Secondly, Tyr44 in crambin is a well-ordered surface residue
that is solvated by a network of waters on the one side, and makes a van der Waals interaction with
Ile33 on the other side. As shown in Figure 7, one of these ordered water molecules, W1023, has
its D1
atom2017,
pointing
Molecules
22, 596 towards the center of the aromatic ring, indicating a potential O-H
11 of. .25. π H
bonding interaction [20]. With improvements in X-ray and neutron instrumentation and crystallization
crystallization
ultra high-resolution
crystallography
willHallow
for H bonding
interactions
to
approaches,
ultra approaches,
high-resolution
crystallography
will allow for
bonding
interactions
to be studied
be
studied
in
greater
detail
and
to
better
accuracy.
in greater detail and to better accuracy.
Figure 7. Crambin residue Tyr44 is making van der Waals contact with Ile33 on one side of the
Figure
7. Crambin residue Tyr44 is making van der Waals contact with Ile33 on one side of the aromatic
aromatic ring and is solvated on the other side. W1023 interacts with the π-system of residue Tyr44.
ring and is solvated on the other side. W1023 interacts with the π-system of residue Tyr44. D1 of W1023
D1 of W1023 is directly pointing Gly31 and different degrees of H/D exchange. Figure reproduced
is directly pointing Gly31 and different degrees of H/D exchange. Figure reproduced from [20].
from [20].
−−), and Hydronium Ions (H +O+ ) in Proteins
+), Hydroxides
4.3. Protons
(H+ ),
(OH
4.3. Protons
(HHydroxides
(OH
),
Hydronium Ions (H3O3) in Proteins
+ important
Water
(H2O),
protons
), hydroxide(OH
(OH−−)) and
(H(H
3O+)O
players
Water
(H2 O),
protons
(H+(H
),+hydroxide
andhydronium
hydroniumions
ions
important
players
3 are) are
in
many
chemical
and
biochemical
processes.
They
may
be
involved
in
a
variety
of
different
enzyme
in many chemical and biochemical processes. They may be involved in a variety of different enzyme
+
catalytic
processes
difficult,ororininthe
thecase
case of
of H
to to
directly
observe
usingusing
X-rays.
catalytic
processes
butbut
areare
difficult,
H+, ,impossible
impossible
directly
observe
X-rays.
In electron density maps derived from X-ray crystallography, most of these hydrated proton forms
In electron density maps derived from X-ray crystallography, most of these hydrated proton forms
are indistinguishable from each other as only the oxygen peak of these species are visible in electron
are indistinguishable from each other as only the oxygen peak of these species are visible in electron
density maps. Neutrons provide the appropriate tool to “see” the light atoms in solvent and make it
density
maps.
provide
the
appropriate
tool towater.
“see” the light atoms in solvent and make it
possible
to Neutrons
discern exotic
solvent
species
from ordinary
possible to
discern exotic solvent
ordinary
Structure-function
studies species
of xylosefrom
isomerase
(XI) water.
using neutrons made it possible to observe
Structure-function
studies
of xylose
(XI)inusing
neutronssystem
made it[35,36].
possible
observe
the first hydronium,
hydroxyl,
and isomerase
proton ions
a biological
XI tofrom
the first
hydronium,
hydroxyl,
andkDa
proton
ions in a biological
system
[35,36].
from Streptomyces
Streptomyces
rubiginosus
is a 172
homotetramer
with 2 metal
sites that
bind XI
divalent
cations,
usuallyismagnesium,
in the active site.
XI catalyzes
the interconversion
of D-glucose
D-fructose,
or
rubiginosus
a 172 kDa homotetramer
with
2 metal sites
that bind divalent
cations, to
usually
magnesium,
D
-xylose
to
D
-xylulose,
and
is
of
considerable
industrial
interest
for
production
of
sweeteners
and
in
in the active site. XI catalyzes the interconversion of D-glucose to D-fructose, or D-xylose to D-xylulose,
the
biofuel
field
[37].
Metal
site
M1
is
coordinated
by
a
number
of
waters,
and
Glu
and
Asp
amino
and is of considerable industrial interest for production of sweeteners and in the biofuel field [37].
acid residues. Metal site M2 is coordinated by His, Asp, Glu side chains and a catalytic water
Metal site M1 is coordinated by a number of waters, and Glu and Asp amino acid residues. Metal
molecule. During catalysis there are multiple steps: substrate binding and coordination adjacent to
site M2 is coordinated by His, Asp, Glu side chains and a catalytic water molecule. During catalysis
the two metal sites, sugar ring opening, movement of a H from sugar C2 to sugar C1 through a
thereproposed
are multiple
steps:
substrate
binding
coordination
adjacent
to theThe
twoprocess
metal sites,
sugar ring
hydride
shift
or donation
of aand
hydride
anion, and
ring closure.
culminates
opening,
movement
of
a
H
from
sugar
C2
to
sugar
C1
through
a
proposed
hydride
shift
or
with the product leaving the active site. These steps necessarily involve complex changes in waterdonation
and
of a hydride
anion,
ring closure.
process culminates
theprotonation/deprotonation
product leaving the active
amino side
chainand
positions,
as well The
as movement
of H atoms with
through
of site.
active site involve
residues complex
[35,38]. changes in water and amino side chain positions, as well as
Thesewater
stepsand
necessarily
Using
combination
of protonation/deprotonation
different metal substitutions and
by soaking
in substrate
or product
movement
of Haatoms
through
of water
and active
site residues
[35,38].
molecules,
it was possible
to trap various
stages
occur during
catalysis or
by product
XI.
Using
a combination
of different
metalintermediate
substitutions
and that
by soaking
in substrate
In
one
such
structure
of
the
natural
magnesium-containing
enzyme
in
complex
with
the
product
molecules, it was possible to trap various intermediate stages that occur during catalysis by XI. In one
xylulose, a hydroxyl anion was observed in place of the catalytic water at the M2 site. This observation
such structure of the natural magnesium-containing enzyme in complex with the product xylulose,
supports the idea that the catalytic water molecule donated a proton during the isomerization step of
a hydroxyl anion was observed in place of the catalytic water at the M2 site. This observation supports
catalysis and gave important clues as to the molecular events that occur upon isomerization [35].
the idea that
the catalytic
water
molecule
a proton
during
of catalysis
Subsequent
neutron
studies
lookeddonated
at metal free
XI at both
pHthe
5.9 isomerization
and 7.7 to help step
understand
and gave
important
clues
as
to
the
molecular
events
that
occur
upon
isomerization
[35].
why the enzyme is inactive at low pH. XI is most active at ~pH 8, however for industrial applications
it is desirable to engineer the enzyme to also be active at lower pH. It is known that at low pH the
metals from M1 and M2 are expelled due to electrostatic changes in the area and this was also
Molecules 2017, 22, 596
12 of 26
Subsequent neutron studies looked at metal free XI at both pH 5.9 and 7.7 to help understand
why the enzyme is inactive at low pH. XI is most active at ~pH 8, however for industrial applications it
is desirable to engineer the enzyme to also be active at lower pH. It is known that at low pH the metals
from Molecules
M1 and2017,
M222,are
596expelled due to electrostatic changes in the area and this was also observed
12 of 25in the
neutron crystal structures. Neutron studies were conducted to study how the active site looks in the
observed
in the
neutron
crystal
structures.
studies
were as
conducted
to study
how the active
absence
of metal,
and
how the
charged
stateNeutron
of residues
change
a function
of pH.
site looks in
metal, andathow
residues
change
as a function
of pH. by a
Analysis
of the
theabsence
neutronofstructure
pH the
7.7 charged
showedstate
thatof
the
metal at
position
M1 is replaced
Analysis
of
the
neutron
structure
at
pH
7.7
showed
that
the
metal
at
position
M1
is
replaced
by for
hydronium molecule (Figure 8) [36]. The hydronium ion templates the appropriate size and shape
a hydronium molecule (Figure 8) [36]. The hydronium ion templates the appropriate size and shape
metal binding at M1 and is coordinated by Glu181, Glu217, Asp245, and Asp287. In the pH 5.9 structure
for metal binding at M1 and is coordinated by Glu181, Glu217, Asp245, and Asp287. In the pH 5.9
the hydronium ion has been dehydrated to leave a proton only in the M1 site. The proton is involved
structure the hydronium ion has been dehydrated to leave a proton only in the M1 site. The proton is
in a very
shortintrifurcated
bond to coordinating
residues and
located
Glu217.
At low pH
involved
a very shortHtrifurcated
H bond to coordinating
residues
andclosest
locatedto
closest
to Glu217.
the amino
acid
residues
have
collapsed
around
the
site
and
does
not
have
the
proper
geometry
At low pH the amino acid residues have collapsed around the site and does not have the proper to
accept
an incoming
metal
ligand. Soaking
in solutions
metal salts
were
also
unsuccessful,
providing
geometry
to accept
an incoming
metal ligand.
Soaking inofsolutions
of metal
salts
were
also unsuccessful,
providing
further
lowof
pH
form
of M1
is not
bind metal,
explaining
further
evidence
thatevidence
the lowthat
pHthe
form
M1
is not
able
to able
bindtometal,
thus thus
explaining
thethe
lack of
lackofofXI
activity
XI at
low pH [37].
activity
at lowofpH
[37].
Figure
Atomicdetails
details of
of metal
metal binding
at two
different
pHs. pHs.
Figure
8. 8.Atomic
binding site
siteM1
M1using
usingneutrons
neutrons
at two
different
+ located in metal site M1 at pH 7.7 (from PDB ID 3kcj);
3O+
(Left)
Coordination
of
hydronium
ion
D
(Left) Coordination of hydronium ion D3 O located in metal site M1 at pH 7.7 (from PDB ID 3kcj);
+ located in metal site M1 at pH 5.9 (from PDB ID 3qza).
(Right)
Coordination
of proton
(Right)
Coordination
of proton
D+Dlocated
in metal site M1 at pH 5.9 (from PDB ID 3qza).
4.4. The Role of D3O+ Ions in Redox Processes
4.4. The Role of D3 O+ Ions in Redox Processes
Rubredoxins are highly thermostable, small, mononuclear iron-sulfur cluster proteins found in
Rubredoxins
areand
highly
thermostable,
mononuclear
proteins
both prokaryotes
eukaryotes.
As redoxsmall,
proteins
they serve as iron-sulfur
an importantcluster
model system
forfound
the in
both study
prokaryotes
and
eukaryotes.
As
redox
proteins
they
serve
as
an
important
model
system
of proton transfer in catalysis. Iron-sulfur proteins vary greatly in their redox potentials, for
the study
of proton
transfer
in catalysis.
Iron-sulfur
vary
greatly
in their
redox
potentials,
reflecting
differences
in solvation,
electrostatics,
and proteins
H bonding.
Neutron
studies
were
initiated
to
obtaindifferences
high resolution
structureselectrostatics,
of both oxidized
on rubredoxin,
in order
to to
reflecting
in solvation,
andand
H reduced
bonding.forms
Neutron
studies were
initiated
study
water
molecules,
H
bonding,
and
to
investigate
the
charged
state
of
chemical
groups
in
the
obtain high resolution structures of both oxidized and reduced forms on rubredoxin, in order to study
as a function
of oxidation
[39].
waterprotein
molecules,
H bonding,
and tostate
investigate
the charged state of chemical groups in the protein as
Inspection of the ~1.3 and 1.4 Å resolution nuclear density maps revealed several interesting
a function of oxidation state [39].
structural features involved with protein stability in both oxidized and reduced forms of the protein.
Inspection of the ~1.3 and 1.4 Å resolution nuclear density maps revealed several interesting
Four hydronium ions were observed in both oxidized and reduced rubredoxin structures. The first
structural
features involved with protein stability in both oxidized and reduced forms of the protein.
molecule, HYD1, was located near Leu51 amide group and involved with amide-imide
Four tautomerization.
hydronium ions
were
observed
in both
oxidized
and reduced rubredoxin
structures.
The first
The
second
one, HYD2,
is found
by Pro44-Pro19-Ser46.
HYD2 is ~10
Å away from
molecule,
HYD1, was
located
Leu51
amide groupHand
involved
with amide-imide
tautomerization.
the iron-sulfur
cluster
andnear
forms
water-mediated
bonds
to iron-sulfur
coordinating
Cys and
The second
HYD2, isInfound
by HYD2
Pro44-Pro19-Ser46.
is H
~10
Å away
from
the
iron-sulfur
adjacentone,
Ser residues.
this way
is at the center HYD2
of a tight
bonded
group
that
connects
water
to
the
metal
cluster
coordinating
residues.
HYD3
also
found
near
the
iron-sulfur
cluster
(~9
Å),
cluster and forms water-mediated H bonds to iron-sulfur coordinating Cys and adjacent Ser residues.
but
near
Pro25-Ser24.
The
last
one,
HYD4,
is
H
bonded
to
Ala16
carbonyl
group.
Occupancy
In this way HYD2 is at the center of a tight H bonded group that connects water to the metal cluster
refinement
of one of the
D atoms
water/hydronium
equilibrium.
coordinating
residues.
HYD3
also indicates
found near
the iron-sulfur
cluster (~9 Å), but near Pro25-Ser24.
In summary then, three of the four hydronium ions are observed close to the protein main chain
The last one, HYD4, is H bonded to Ala16 carbonyl group. Occupancy refinement of one of the D
and another one forms the core of an intricate H bonded water network. One of these hydronium
atoms indicates water/hydronium equilibrium.
ions is located near main chain amide of Leu51 and is probably involved with redox driven
In
summary then, three of the four hydronium ions are observed close to the protein main chain
tautomerization of amide to imide forms of the residue. In addition to the several hydronium ions
and another
one forms the
core
of an
intricate
H bonded
water
network.shifts
Oneobserved,
of these hydronium
seen in rubredoxin,
there
were
two
additional
important
protonation
in residues ions
Glu47, Asp13, and Asp15. Glu47 was seen to be unprotonated in the reduced form while being
Molecules 2017, 22, 596
13 of 26
is located near main chain amide of Leu51 and is probably involved with redox driven tautomerization
of amide to imide forms of the residue. In addition to the several hydronium ions seen in rubredoxin,
there were two additional important protonation shifts observed, in residues Glu47, Asp13, and Asp15.
Glu47 was seen to be unprotonated in the reduced form while being protonated in the oxidized form.
This shifting D atom is also associated with HYD2. Asp13 on the other hand is unprotonated in the
oxidized form while Asp15 is protonated. The observation of four hydronium ions and three shifting
residues in close proximity to the iron-sulfur cluster suggest that these entities and surrounding solvent
play important roles in charge transfer processes that occur in the change between oxidization and
reduction in rubredoxin and associated redox processes [39].
With the growing use of neutron protein crystallography, it can be expected that more of these
unusual solvent types will be directly observed and shed light on intricate and complex protein functions.
5. Structural Enzymology and H Bonds
Hydrogen atoms are central to enzyme mechanisms and catalysis often relies on the movement
of H atoms, either through proton shuttling or hopping or through the breaking/making of H
bonds. There are many examples of neutron crystallography helping to elucidate enzyme mechanistic
questions and all cannot be covered here. In this section a few examples are discussed where using
neutrons made a significant contribution with regards to hydrogen atoms and their interactions.
5.1. Cytochrome C Peroxidase
Heme peroxidases are widespread in nature and support a wide range of biological functions.
They employ peroxide to oxidize a number of different substrates. Catalysis by heme peroxidases
require the formation of two transient, highly oxidized ferryl (Fe(IV)) intermediates, the so-called
Compounds I and II. These intermediates are applied widely in other oxygen-dependent catalytic
heme-containing enzymes, including the cytochrome P450s, nitric oxide synthases, terminal oxidases
and the heme dioxygenases [40]. It is important to understand the protonation state around the heme
as this determines the reactivity and involvement of the enzyme.
Compound I and Compound II form sequentially during catalysis and represent different Fe-IV
(ferryl) species. They are different in their oxidation state, specifically the porphyrin ring. It is thought
that the species are Fe(IV)=O or Fe(IV)-OH and it has been challenging to characterize the identities
definitively [40,41]. Photo-reduction of the iron has been observed in both Raman spectroscopy
and X-ray crystallography, rendering the interpretation of the oxidation states derived from these
methods unreliable. Neutron crystallography can provide the high resolution, unambiguous structural
information required to characterize the iron oxidation state, without inducing photo-reduction in
the samples [40,41]. To unravel the mechanism of the peroxidases, two related heme peroxidases
have been studied with neutron crystallography. Cytochrome c peroxidase (Ccp) oxidizes the small
protein cytochrome c by hydrogen peroxide, and the mechanistically and structurally related ascorbate
peroxidase (APX), the oxidation of ascorbate [40,41].
Cytochrome c peroxidase is a well-studied model system for heme-containing oxidases, which
play a key role in many metabolic processes. Various species along the reaction path can be
characterized by spectroscopic methods and trapped by cryo-cooling for crystallographic structure
determination. However, the protonation states—which are crucial for understanding the electronic
structure—are not easy to infer even with the combination of spectroscopy and crystallography.
The structure from neutron diffraction study of the ferric state of the heme, determined at 2.4 Å
resolution, showed a heavy water molecule (i.e., D2 O) bound at the start of the reaction (Figure 9, left).
Crystals were then treated with peroxide, known to induce formation of Compound I. The treated
crystals were cryo-cooled to trap the intermediate and studied with neutrons. The neutron structure,
determined to 2.5 Å resolution, clearly shows the oxygen to be unprotonated, consistent with the
H bonding structure of the active site (Figure 9, right). The ferryl iron is then double bonded to an
Molecules 2017, 22, 596
14 of 26
oxygen (Fe(IV)=O). The Trp 191 radical is protonated, so the radical species is a (protonated) π-cation
radical
[41]. 2017, 22, 596
Molecules
14 of 25
Figure
9. Neutron
crystal
structures
peroxidaseinin
the
ferric
Compound
I form.
Figure
9. Neutron
crystal
structuresofofcytochrome
cytochrome cc peroxidase
the
ferric
andand
Compound
I form.
2
O,
is
bound
to
the
heme;
(Right)
nuclear
(Left)
nuclear
density
maps
show
that
a
heavy
water,
D
(Left) nuclear density maps show that a heavy water, D2 O, is bound to the heme; (Right) nuclear
density
maps
show
oxygen
boundtotothe
the heme
heme in
I intermediate
form.
Figure
density
maps
show
thatthat
oxygen
is is
bound
in the
theCompound
Compound
I intermediate
form.
Figure
reproduced
from
[41].
reproduced from [41].
Another recent neutron crystallographic study focused on determining the oxidation and
Another
recent
crystallographic
focused
on determining
the oxidation
coordination
state ofneutron
the Compound
II intermediate study
in APX [40].
APX catalyzes
the peroxide-dependent
and oxidation
coordination
state of
II intermediate
in II
APX
[40]. APX
of ascorbate
by the
usingCompound
both Compound
I and Compound
intermediates.
APXcatalyzes
and Ccp the
are highly conserved
and Ccp
has servedby
as ausing
model
system
for mechanistic
studies in heme
enzyme
peroxide-dependent
oxidation
of ascorbate
both
Compound
I and Compound
II intermediates.
catalysis.
The
experimental
challenges
of
studying
Compound
II
are
more
easily
overcome
in APX
APX and Ccp are highly conserved and Ccp has served as a model system for mechanistic
studies
compared
to
Ccp.
In
APX
Compound
I
appears
as
a
ferryl
heme
and
a
porphyrin
π-cation
radicaleasily
in heme enzyme catalysis. The experimental challenges of studying Compound II are more
and is readily distinguishable from Compound II, which contains only a ferryl species. This is
overcome
in APX compared to Ccp. In APX Compound I appears as a ferryl heme and a porphyrin
different from Ccp where Compound I is a ferryl heme and a Trp radical. Careful X-ray analysis and
π-cation radical and is readily distinguishable from Compound II, which contains only a ferryl species.
microspectrophotometry, along with neutron crystallography, revealed the heme ligand to be
This ishydroxyl.
differentAtfrom
Ccp where Compound I is a ferryl heme and a Trp radical. Careful X-ray analysis
the resolution limits of the crystallographic studies, it is not justified to assert that the
and microspectrophotometry,
along with
revealed
theanalysis
heme ligand
refined distance (1.88 Å) corresponds
to aneutron
hydroxylcrystallography,
bound to the heme.
However,
of omitto be
hydroxyl.
At the
thetocrystallographic
it is notIIjustified
assert that
maps and
lackresolution
of residual limits
densityofled
the determinationstudies,
that Compound
in APX to
is indeed
the refined
distance
(1.88 Å)these
corresponds
to anew
hydroxyl
the heme.
However,
analysis of
Fe(IV)-OH
[40]. Together
results offer
insightsbound
on the to
mechanism
of heme
peroxidases,
the entire
family
of the
enzymes.
Importantly,
the first time,
they is
offer
omit with
mapsimplications
and lack of for
residual
density
led to
determination
thatfor
Compound
II in APX
indeed
unambiguous
insights
into
the
coordination
and
protonation
state
around
the
heme
groups
where
Fe(IV)-OH [40]. Together these results offer new insights on the mechanism of heme peroxidases, with
other techniques
have been
unsuitable.
implications
for the entire
family
of enzymes. Importantly, for the first time, they offer unambiguous
insights into the coordination and protonation state around the heme groups where other techniques
5.2. Urate Oxidase
have been unsuitable.
Urate oxidase (EC 1.7.3.3) is an enzyme involved in purine metabolism. It is a co-factorless oxidase
that catalyzes
5.2. Urate
Oxidase the reaction between uric acid and molecular oxygen to form 5-hydroxy-isourate
(5-HIU), which in turn is hydrolyzed to allantoin.
UrateThe
oxidase
1.7.3.3)
is an
involved
in purine in
metabolism.
It ispredominant
a co-factorless
reaction(EC
between
oxygen
andenzyme
a urate dianion
is spontaneous
solution, but the
oxidase
that
catalyzes
the
reaction
between
uric
acid
and
molecular
oxygen
to
form
5-hydroxy-isourate
protonation state of uric acid in solution is a monoanion with N3 deprotonated (Figure 10) [42].
(5-HIU),
whichthe
in principal
turn is hydrolyzed
to allantoin.
Therefore
role of the enzyme
is to deprotonate the urate substrate in proximity of the
oxygen.
Numerous
X-ray
structures
determined
with inhibitors
that mimic
as well
The
reaction
between
oxygen
and awere
urate
dianion isboth
spontaneous
in solution,
but urate
the predominant
as oxygen
at atomic
The results
difficult towith
reconcile
with the biochemical
protonation
state
of uricresolution.
acid in solution
is awere
monoanion
N3 deprotonated
(Figuredata,
10) [42].
since
the
site
of
the
next
deprotonation
in
solution,
N7,
is
H
bonded
to
a
main
chain
amide
and
Therefore the principal role of the enzyme is to deprotonate the urate substrate in proximitynoof the
feasible
general bases
be found were
in the determined
vicinity. The neutron
structure
at 2.35that
Å resolution
(Figure
oxygen.
Numerous
X-raycan
structures
both with
inhibitors
mimic urate
as11)
well as
showed that in fact the species in the active site was the enol tautomer of urate, 8-hydroxyxanthine.
oxygen at atomic resolution. The results were difficult to reconcile with the biochemical data, since
the site of the next deprotonation in solution, N7, is H bonded to a main chain amide and no feasible
general bases can be found in the vicinity. The neutron structure at 2.35 Å resolution (Figure 11)
showed that in fact the species in the active site was the enol tautomer of urate, 8-hydroxyxanthine.
Molecules 2017, 22, 596
Molecules 2017, 22, 596
Molecules 2017, 22, 596
15 of 26
15 of 25
15 of 25
Figure
10. The
reaction catalyzed
by urate
oxidase.
Figure
reproduced
from [42].
[42].
Figure
oxidase.Figure
Figurereproduced
reproducedfrom
from
Figure10.
10.The
Thereaction
reactioncatalyzed
catalyzed by
by urate
urate oxidase.
[42].
Figure
The
2.35ÅÅresolution
resolution2mF
2mFoo-DF
-DFcc nuclear
nuclear density
with
Figure
11.11.
The
2.35
density map
mapof
ofurate
urateoxidase
oxidase(A)
(A)compared
compared
with
Figure 11. The 2.35 Å resolution 2mFo -DFc nuclear density map of urate oxidase (A) compared with
1.05
X-raymap
map(B);
(B);(C)
(C)aaschematic
schematic representation
representation of
substrate.
thethe
1.05
ÅÅ
X-ray
of the
thesurroundings
surroundingsofofthe
theurate
urate
substrate.
the 1.05 Å X-ray map (B); (C) a schematic representation of the surroundings of the urate substrate.
Panels A
and
reproducedfrom
from[42].
[42].
Panels
and
BB
reproduced
Panels A
A and
B
reproduced
from [42].
It was
shownbybyquantum
quantumchemical
chemical calculations
calculations that
unstable
It was
shown
that despite
despitethe
theenol
enoltautomer
tautomerbeing
being
unstable
It
was shown
by quantum
chemical
calculations
that
despite
theclearly
enol tautomer
being
unstable
in
in
solution,
the
active
site
can
stabilize
it.
Interestingly,
the
O-H
was
pointing
away
from
thethe
in solution, the active site can stabilize it. Interestingly, the O-H was clearly pointing away from
solution,
the
active
site
can stabilize
it. Interestingly,
the
O-H was
clearly
pointing
away
from theThere
purine
purine
plane,
even
though
there
was
no
H
bonding
partner
observed
to
explain
this
behavior.
purine plane, even though there was no H bonding partner observed to explain this behavior. There
plane,
though
there
was no
H bonding
partner
observed
explain
this
behavior.
There
was even
however
enough
empty
space
for a water
molecule,
even though
none
were
observed
in the
was
however
enough
empty
space
for
a water
molecule,
even to
though
none
were
observed
inwas
the
X-ray maps.
quantum
calculations
wereinperformed
to
however
enoughFurther
empty space
for amechanics/molecular
water molecule, even mechanics
though none
were observed
the X-ray maps.
X-ray maps. Further quantum mechanics/molecular mechanics calculations were performed to
understand
this. It was not possible to mechanics
obtain good geometry consistent
with thetohigh-resolution
Further
quantum
were performed
this.
understand
this. Itmechanics/molecular
was not possible to obtain goodcalculations
geometry consistent
with the understand
high-resolution
X-ray
structure
unless
a
water
molecule
was
placed
in
proper
H
bonding
geometry
with
respect
to
It was structure
not possible
to obtain
good
geometry
consistent
the high-resolution
X-ray structure
unless
X-ray
unless
a water
molecule
was
placed inwith
proper
H bonding geometry
with respect
to
the O-H.
This illustrates
the in
synergies
between
crystallography
andrespect
computational
chemistry
where
a
water
molecule
was
placed
proper
H
bonding
geometry
with
to
the
O-H.
This
illustrates
the O-H. This illustrates the synergies between crystallography and computational chemistry where
methods
can predict and
or fill
in details when
crystalwhere
structures
give incomplete
thecomputational
synergies between
crystallography
computational
chemistry
computational
methods
computational
methods
can predict
or fill
in detailsinwhen
crystal instructures
give
incomplete
information.
A
water
molecule
was
indeed
observed
this
position
an
X-ray
structure
from
can predict orAfill
in details
whenwas
crystal
structures
giveinincomplete
information.
A water
molecule
information.
water
molecule
indeed
observed
this position
in an X-ray
structure
from
anaerobically
grown
crystals
[43].
was
indeed observed
in this position
in an X-ray structure from anaerobically grown crystals [43].
anaerobically
grown
crystals
[43].
This result allowed a more plausible mechanism to be postulated where the deprotonation
This result
allowed
a amore
plausible
mechanism
to be postulated
wherewhere
the deprotonation
occurs
This
result
allowed
more
plausible
mechanism
be postulated
deprotonation
occurs
on
the oxygen
instead
of the
nitrogen.
One of the to
unsolved
questions
in thethe
mechanism
was
on
the
oxygen
instead
of
the
nitrogen.
One
of
the
unsolved
questions
in
the
mechanism
was
always
occurs
on
the
oxygen
instead
of
the
nitrogen.
One
of
the
unsolved
questions
in
the
mechanism
was
always how an H atom from the N7 position can reach the oxygen site above the ring plane.
how
an
H
atom
from
the
N7
position
can
reach
the
oxygen
site
above
the
ring
plane.
A
surprising
always
how anobservation
H atom from
N7 position
oxygennetwork
site above
the ring
A surprising
in thethe
neutron
structurecan
was reach
that anthe
H bonded
extended
fromplane.
N7
observation
in the neutroninstructure
wasstructure
that an− Hwas
bonded
network
extended
fromextended
N7 all the
wayN7
to
A
surprising
the(that
neutron
that
an H bonded
network
from
all the way observation
to the oxygen site
contains
Cl
in
this
structure),
and
the
H
atoms
were
disordered
− in this structure), and the H atoms were disordered (Figure 12).
the
oxygen
site
(that
contains
Cl
12).
all(Figure
the way
to the oxygen site (that contains Cl− in this structure), and the H atoms were disordered
Nevertheless
the
oxygen atoms
atomswere
werewell
wellordered.
ordered.This
This
Nevertheless
theX-ray
X-raystructure
structure showed
showed that
that the
the oxygen
is is
(Figure 12).
indicative
of of
a proton
chain
that
can
easily
pass
the
proton
through
three
water
molecules,
a aLys
indicative
a proton
relay
chain
that
can
easily
pass
the
proton
through
three
water
molecules,
Nevertheless
therelay
X-ray
structure
showed
that
the
oxygen
atoms
were
well
ordered.
This
is
and
a
Ser
side
chain.
The
existence
of
such
a
relay
chain
also
explains
how
the
oxygen
site
can
Lys and aofSer
side chain.
The
existence
of such
a relay
also explains
howthree
the oxygen
site can be be
indicative
a proton
relay
chain
that can
easily
passchain
the proton
through
water molecules,
a
provided
with
a
second
proton
to
eventually
form
H
O
.
2
2
provided
with
a
second
proton
to
eventually
form
H
2
O
2
.
Lys and a Ser side chain. The existence of such a relay chain also explains how the oxygen site can be
provided with a second proton to eventually form H2O2.
Molecules 2017, 22, 596
16 of 26
Molecules 2017, 22, 596
16 of 25
Figure 12. The postulated proton relay chain as seen in the 2mFo-DFc nuclear density map. Figure
Figure 12. The postulated proton relay chain as seen in the 2mFo -DFc nuclear density map. Figure
reproduced from
from [42].
[42].
reproduced
5.3. β-Lactamase
5.3. β-Lactamase
Bacteria have developed antibiotic resistance to β–lactam antibiotics by expressing an enzyme,
Bacteria have developed antibiotic resistance to β–lactam antibiotics by expressing an enzyme,
β-lactamase, which breaks the amide bond within the lactam ring and renders the compound ineffective.
β-lactamase, which breaks the amide bond within the lactam ring and renders the compound
This class of antibiotics inhibits cell wall synthesis by binding to bacterial penicillin-binding proteins
ineffective. This class of antibiotics inhibits cell wall synthesis by binding to bacterial penicillin-binding
(PBPs). The β-lactamases are divided into classes A through D. All but class B use a serine-reactive
proteins (PBPs). The β-lactamases are divided into classes A through D. All but class B use a
mechanism. The class A CTX-M extended-spectrum-β-lactamase (ESBL) has high hydrolytic activity
serine-reactive mechanism. The class A CTX-M extended-spectrum-β-lactamase (ESBL) has high
against first, second, and third generation cephalosporins and monobactams. The enzymes work by
hydrolytic activity against first, second, and third generation cephalosporins and monobactams.
using a general base mechanism to break the amide bond within the β-lactam ring [44]. One class A
The enzymes work by using a general base mechanism to break the amide bond within the β-lactam
enzyme, CTX-M-44, was chosen for neutron studies to further investigate the role of active site
ring [44]. One class A enzyme, CTX-M-44, was chosen for neutron studies to further investigate
Glu166, Ser70, and Lys73 in catalysis. CTX-M-44 is composed of 262 residues and has two conserved
the role of active site Glu166, Ser70, and Lys73 in catalysis. CTX-M-44 is composed of 262 residues
domains, with the active site located at the interface between the domains. It has been proposed that
and has two conserved domains, with the active site located at the interface between the domains.
Glu166 acts as a general catalytic base in the acylation part of the reaction through deprotonation of
It has been proposed that Glu166 acts as a general catalytic base in the acylation part of the reaction
Ser70. Ser70 then attacks the carbonyl carbon of the β-lactam ring. Glu166 is also thought to be
through deprotonation of Ser70. Ser70 then attacks the carbonyl carbon of the β-lactam ring. Glu166
involved in the second part de-acylation part of the reaction by activating a hydrolytic water
is also thought to be involved in the second part de-acylation part of the reaction by activating a
molecule. Interestingly, the CTX-M-44 Glu166Ala (E166A) mutant is still able to degrade β-lactams,
hydrolytic water molecule. Interestingly, the CTX-M-44 Glu166Ala (E166A) mutant is still able to
but with the rates several orders of magnitude slower [45]. This suggests there is more than one way to
degrade β-lactams, but with the rates several orders of magnitude slower [45]. This suggests there is
degrade β-lactams, including some that have been proposed to use Lys73 [46]. Neutron studies were
more than one way to degrade β-lactams, including some that have been proposed to use Lys73 [46].
conducted on the E166A mutant as it degrades the β-lactam slow enough to allow crystallographic
Neutron studies were conducted on the E166A mutant as it degrades the β-lactam slow enough to
investigation of the acyl-intermediate. Studies were done on both the ligand-free enzyme and in
allow crystallographic investigation of the acyl-intermediate. Studies were done on both the ligand-free
complex with cefotaxime. The data on the complex revealed that upon inhibitor binding, an
enzyme and in complex with cefotaxime. The data on the complex revealed that upon inhibitor binding,
oxyanion hole is formed involving the backbone amides of Ser237 and Ser70 with stabilizing H
an oxyanion hole is formed involving the backbone amides of Ser237 and Ser70 with stabilizing H
bonds to the carbonyl oxygen of cefotaxime. This interaction serves to stabilize the negative charge that is
bonds to the carbonyl oxygen of cefotaxime. This interaction serves to stabilize the negative charge that
formed in the tetrahedral intermediates during the acylation and deacylation stages of hydrolysis.
is formed in the tetrahedral intermediates during the acylation and deacylation stages of hydrolysis.
Lys73 is highly conserved in class A β-lactamases and has been thought to act as general base,
Lys73 is highly conserved in class A β-lactamases and has been thought to act as general base,
in competition with Glu166, by abstracting a proton from Ser70. The ability of E166A mutant to still
in competition with Glu166, by abstracting a proton from Ser70. The ability of E166A mutant to
produce the acyl-enzyme complexes, which supports the idea that Lys73 is involved, albeit less
still produce the acyl-enzyme complexes, which supports the idea that Lys73 is involved, albeit less
effectively than Glu166. In the ligand-free neutron structure, it is seen that Lys73 is charged (-D3+)
effectively than Glu166. In the ligand-free neutron structure, it is seen that Lys73 is charged (-D3 + ) and
and occupies a single conformation that is identical to that observed in the form where Glu166 is
occupies a single conformation that is identical to that observed in the form where Glu166 is present.
present. In the cefotaxime-complex structures however, Lys73 is observed in two alternative
In the cefotaxime-complex structures however, Lys73 is observed in two alternative conformations
conformations that refine to near equal occupancy, presumably due to different charged states
that refine to near equal occupancy, presumably due to different charged states (Figure 13). These
(Figure 13). These observations together support a role for Ly73 in proton transfer and represents a
observations together support a role for Ly73 in proton transfer and represents a viable alternative
viable alternative pathway when Glu166 is mutated [45].
pathway when Glu166 is mutated [45].
Molecules 2017, 22, 596
17 of 26
Molecules 2017, 22, 596
Molecules 2017, 22, 596
17 of 25
17 of 25
Figure
13. Proposed
catalytic
residuesLys73
Lys73in
in class
class A—lactamase
actact
as aasgeneral
base.base.
In theIn the
Figure
13. Proposed
catalytic
residues
A—lactamasecould
could
a general
acyl-enzyme complex, Lys73 can be modeled in two conformations. The nuclear density map is
acyl-enzyme complex, Lys73 can be modeled in two conformations. The nuclear density map is shown
shown in pink mesh and is contoured at 0.8σ. Reprinted (adapted) with permission from [45].
in pink mesh and is contoured at 0.8σ. Reprinted (adapted) with permission from [45]. Copyright 2016
Copyright 2016 American Chemical Society.
American Chemical Society.
Figure 13. Proposed catalytic residues Lys73 in class A—lactamase could act as a general base. In the
5.4. Diisopropyl
Fluorophosphatase
acyl-enzyme
complex, Lys73(DFPase)
can be modeled in two conformations. The nuclear density map is
5.4. Diisopropyl
Fluorophosphatase
shown
in pink mesh and(DFPase)
is contoured
at 0.8σ. Reprinted
with permission
[45]. in the
The enzyme diisopropyl fluorophosphatase
(DFPase; (adapted)
314 residues,
35 kDa),from
found
Copyright 2016 American Chemical Society.
Mediterranean
squid Loligo vulgaris,
is a calcium-dependent
that can hydrolyze
The
enzyme diisopropyl
fluorophosphatase
(DFPase;enzyme
314 residues,
35 kDa),a variety
found of
in the
toxic
organophosphorus
compounds
through
hydrolysis
of
a
P-F
bond.
The
compounds
include
the
Mediterranean
squidFluorophosphatase
Loligo vulgaris,(DFPase)
is a calcium-dependent enzyme that can hydrolyze a variety
5.4. Diisopropyl
nerve
agents Sarin, Soman,
and cyclosarin,
and hydrolysis
the pesticideofdiisopropyl
fluorophosphate
(DFP).
of toxic
organophosphorus
compounds
through
a P-F bond.
The compounds
include
The enzyme
diisopropylinhibitors
fluorophosphatase
(DFPase; 314 and
residues,
kDa), found
in the
These agents
act as irreversible
of acetylcholinesterase
serine35proteases.
The neutron
the nerveMediterranean
agents Sarin,
Soman,
cyclosarin,
and the pesticide
diisopropyl
fluorophosphate
squid
Loligoand
vulgaris,
a calcium-dependent
enzyme
thatstructures
can hydrolyze
a variety and
of (DFP).
structure of DFPase,
together
with a is
variety
of high-resolution
X-ray
of wild-type
These agents
act as irreversible
inhibitors
of acetylcholinesterase
andThe
serine
proteases.
Thethe
neutron
toxic organophosphorus
compounds
through
hydrolysis of a P-F bond.
compounds
include
site-directed mutants, and enzymology have greatly clarified the mechanism [47,48].
agents Sarin,
Soman,
and acyclosarin,
andhigh-resolution
the 18pesticide diisopropyl
fluorophosphate
(DFP). and
structurenerve
of
DFPase,
together
with
variety
of
X-ray
structures
of
wild-type
Single- and multiple-turnover studies utilizing an O substrate, together with mass spectroscopy,
These agents
actand
as irreversible
inhibitors
of
acetylcholinesterase
serine proteases.
The 18neutron
site-directed
enzymology
have
greatly
clarified
theand
mechanism
[47,48].
16O-containing
showed amutants,
product under
single-turnover
conditions,
and incorporation
of O under
structure of DFPase, together with a variety of high-resolution
X-ray structures of wild-type and
18
Singlemultiple-turnover
studies
utilizing
an O the
substrate,
with mass
spectroscopy,
multipleand
turnover
conditions. After
multiple
turnovers,
enzymetogether
was digested
to determine
site-directed
mutants, and enzymology have greatly clarified the mechanism [47,48].
1618
showed
a
O-containing
product
under
single-turnover
conditions,
and
incorporation
of 18 O under
where SingleO hadand
been
incorporated studies
and results
showed
that the together
proteolytic
in question
multiple-turnover
utilizing
an 18O substrate,
withpeptide
mass spectroscopy,
contained
the
residue
Asp229. These
labeling
demonstrated
that
multiple
turnover
conditions.
multiple
turnovers,
the
enzymeexperiments
wasincorporation
digested
toof
determine
where
18O under
showed
a 16calcium-coordinating
O-containing After
product
under
single-turnover
conditions,
and
18 O had
the multiple
reaction
proceeded
through
aAfter
direct
attack that
ofturnovers,
anthe
amino
acidenzyme
on the
substrate,
andto
notdetermine
through
a the
been incorporated
and results
showed
proteolytic
peptide
in question
contained
turnover conditions.
multiple
the
was digested
18O had
metal-activated
water
as
previously
thought.
In
the
neutron
structure,
Asp229
was
found
to
be
where
been
incorporated
and
results
showed
that
the
proteolytic
peptide
in
question
calcium-coordinating residue Asp229. These labeling experiments demonstrated that the reaction
deprotonated,
consistent
with the proposed
(Figure
14A)experiments
[49].
contained the
calcium-coordinating
residuemechanism
Asp229. These
labeling
demonstrated that
proceeded through a direct attack of an amino acid on the substrate, and not through a metal-activated
the reaction proceeded through a direct attack of an amino acid on the substrate, and not through a
water as previously thought. In the neutron structure, Asp229 was found to be deprotonated, consistent
metal-activated water as previously thought. In the neutron structure, Asp229 was found to be
with the deprotonated,
proposed mechanism
(Figure
14A) [49].
consistent with
the proposed
mechanism (Figure 14A) [49].
Figure 14. DFPase active site as observed in the neutron crystal structure. (A) 2Fo − Fc nuclear density
(gray) shows a deprotonated Asp229; (B) Detail of the calcium-binding environment, showing the
Fo −Figure
Fc omit14.
density (gray)
for
the
two D atoms
inneutron
W33. This
establishes
the identity
of W33 density
as water
site
as observed
in the
crystal
structure.
2Fo2F
− Fc −
nuclear
Figure 14.
DFPaseDFPase
activeactive
site as
observed
in the
neutron
crystal
structure.(A)(A)
Fc nuclear density
o
and(gray)
not hydroxide.
Details surrounding
calcium
and associated
water molecule
shows a deprotonated
Asp229; the
(B) active
Detail site
of the
calcium-binding
environment,
showingenable
the
(gray)mechanistic
shows a deprotonated
Asp229; (B)from
Detail
of the calcium-binding environment, showing the
studies.
Figure
[47].
Fo − Fc omit
density
(gray)reproduced
for the two D atoms
in W33. This establishes the identity of W33 as water
Fo − Fc omit
density (gray) for the two D atoms in W33. This establishes the identity of W33 as water
and not hydroxide. Details surrounding the active site calcium and associated water molecule enable
and not
DetailsFigure
surrounding
thefrom
calcium andwith
associated
water
molecule
enable
Thehydroxide.
catalytic
ion
isreproduced
involved
inactive
7-fold
coordination,
a solvent
molecule
bound
in
mechanisticcalcium
studies.
[47].site
mechanistic
studies.
Figure
reproduced
from [47].
the active site
pocket.
While
X-ray structures
were not able to discern the species of the solvent
The
calcium
ion revealed
is involved
in 7-fold
coordination,
with
solvent
moleculeascertained
bound in
molecule,
thecatalytic
neutron
structure
it to
be a water
molecule,
anda not
a hydroxyl,
the active
site pocket.
While
X-ray
structures
were maps
not able
to discern
the
species
of the solvent
by
the
elongated
nuclear
density
and
F
o − Fc omit
(Figure
14B)
[47].
This
unambiguously
The catalytic calcium ion is involved in 7-fold coordination, with a solvent molecule bound in
molecule, the neutron structure revealed it to be a water molecule, and not a hydroxyl, ascertained
the active
pocket. While
to discern
species
of the solvent
by site
the elongated
nuclear X-ray
densitystructures
and Fo − Fcwere
omit not
mapsable
(Figure
14B) [47].the
This
unambiguously
molecule, the neutron structure revealed it to be a water molecule, and not a hydroxyl, ascertained by
the elongated nuclear density and Fo − Fc omit maps (Figure 14B) [47]. This unambiguously showed
Molecules 2017, 22, 596
Molecules 2017, 22, 596
18 of 26
18 of 25
that the mechanism did not involve direct water activation by the calcium ion, as had been previously
showed by
that
the mechanism
didbut
notwas
involve
directwith
waterthe
activation
bythe
therole
calcium
ion, as had Of
been
suggested
other
investigators,
consistent
Asp229 in
as nucleophile.
note
previously
suggested
by
other
investigators,
but
was
consistent
with
the
Asp229
in
the
role
as
was the orientation of the bound water (D2 O) molecule in relation to the position of the catalytic
nucleophile.
Of anote
was
thedistance
orientation
of Å)
thebetween
bound water
2O)Dmolecule
relation
thethe
calcium
ion, with
very
close
(2.08
one of(D
the
atoms ofinthe
water to
and
position
the a
catalytic
ion,
a very close distance (2.08 Å) between one of the D atoms
calcium
ion,ofand
Ca-O-Dcalcium
angle of
53with
degrees.
of the water and the calcium ion, and a Ca-O-D angle of 53 degrees.
5.5. Fungal Lytic Polysaccharide Monooxygenase
5.5. Fungal Lytic Polysaccharide Monooxygenase
Fungal lytic polysaccharide monooxygenases (LPMO) are Cu-containing metalloenzymes
Fungal
lyticdegradation
polysaccharide
monooxygenases
(LPMO)
are Cu-containing
metalloenzymes
involved
with the
of polysaccharides
through
the insertion
of one O into
the glycosidic
involved
with
the
degradation
of
polysaccharides
through
the
insertion
of
one
O
into
glycosidic
bond carbon, leading to chain cleavage. There is significant interest in the mechanism ofthe
these
enzymes
bond carbon, leading to chain cleavage. There is significant interest in the mechanism of these
as there is potential application in cellulose degradation for biofuel production. The general mechanism
enzymes as there is potential application in cellulose degradation for biofuel production. The
starts with a single electron reducing the resting state Cu(II) to Cu(I). Cu(I) has a high affinity for
general mechanism starts with a single electron reducing the resting state Cu(II) to Cu(I). Cu(I) has a
molecular oxygen (O2 ). Oxygen binding to Cu(I) quickly regenerates Cu(II) through oxidation with
high affinity for molecular oxygen (O2). Oxygen binding to Cu(I) quickly regenerates Cu(II) through
the formation of Cu-superoxide. From this intermediate there are multiple possible pathways that
oxidation with the formation of Cu-superoxide. From this intermediate there are multiple possible
may involve superoxide, hydroperoxyl, or oxyl that are responsible for abstracting an H atom from the
pathways that may involve superoxide, hydroperoxyl, or oxyl that are responsible for abstracting an
glycosidic
the absence
of substrate
the superoxide
species
can leave Cu(II)
H atom carbon.
from theInglycosidic
carbon.
In the absence
of substrate
the superoxide
speciesand
cangenerate
leave
peroxide.
There
are many
uncertainties
in many
how these
enzymesinactivate
oxygen
and the
mechanism
Cu(II) and
generate
peroxide.
There are
uncertainties
how these
enzymes
activate
oxygen of
peroxide
no substrate
is present
and theproduction
mechanismwhen
of peroxide
production
when [50,51].
no substrate is present [50,51].
Neurospora
crassa
LPMO
is
a
223
residue
(23.3
enzyme
withCu
one
per monomer.
Neurospora crassa LPMO is a 223 residue (23.3
kDa)kDa)
enzyme
with one
perCu
monomer.
High
High
resolution
crystallography
DFT calculations
were supplemented
with neutron
resolution
X-rayX-ray
crystallography
and DFT and
calculations
were supplemented
with neutron crystallography
crystallography
in order
to better
understand
mono-Cu
activation
of molecular
oxygen and
in order to better
understand
mono-Cu
activation
of molecular
oxygen
and H2O2 production
in H
the
2 O2
production
the absence of substrate.
absence ofinsubstrate.
X-ray
crystal
structuresofofthe
theresting
restingCu(II)
Cu(II) state
state showed
showed that
a conserved
X-ray
crystal
structures
thatCu
Cuisiscoordinated
coordinatedbyby
a conserved
“His
brace”made
madeofof the
the backbone
backbone amide
of His1
and and
the imidazole
of His84.
The
“His
brace”
amideand
andimidazole
imidazole
of His1
the imidazole
of His84.
hydroxyl
group
of
Tyr168
is
pointing
towards
with
Cu
site
with
the
remaining
axial
and
equatorial
The hydroxyl group of Tyr168 is pointing towards with Cu site with the remaining axial and equatorial
coordination
positions
occupiedby
bytwo
twowaters,
waters,H
H22O
Oax
ax and
eq (Figure 15) [51].
coordination
positions
occupied
andHH2O
2 Oeq (Figure 15) [51].
Figure
Cu-binding
site coordination
and coordination
in different
structures
offrom
LPMO
from
Figure
15. 15.
Cu-binding
site and
in different
crystal crystal
structures
of LPMO
Neurospora
2− ) bound;
bound;
(b)structure
X-ray
Neurospora
crassa.
(a) X-ray
structure
of crystal
with peroxo
ascorbate,
(O22−)(b)
crassa.
(a) X-ray
structure
of crystal
treated
withtreated
ascorbate,
(O2peroxo
X-ray
structure
of crystal
withmolecular
ascorbate, oxygen
molecular
bound in “pre-bound”
position;
of crystal
treated
withtreated
ascorbate,
(O2oxygen
) bound(Oin2) “pre-bound”
position; (c)
neutron
(c)
neutron
structure
with
His157
modeled
as
neutral
and
pointing
at
“pre-bound”
position.
Figure
structure with His157 modeled as neutral and pointing at “pre-bound” position. Figure reproduced
reproduced
with
permission
from [51].
with
permission
from
[51].
The coordination sphere for Cu indicates the presence of Cu(II), despite the possibility of
The coordination
for Cu indicates
Cu(II),
despite
the presence
possibility
photoreduction
uponsphere
X-ray exposure.
Treatingthe
thepresence
crystals of
with
ascorbate
in the
of of
photoreduction
upon prior
X-raytoexposure.
Treating
the crystals
ascorbateofinthe
theCu(II)-dioxo
presence of
atmospheric oxygen
freezing leads
to reduction
to Cu(I)with
and formation
2−
atmospheric
oxygen
prior
to
freezing
leads
to
reduction
to
Cu(I)
and
formation
of
the
Cu(II)-dioxo
complex. The crystal structure shows that H2Oeq is displaced by a peroxo (O2 ) species in the equatorial
complex.
crystal
structure
shows
that H2 Oeq isisdisplaced
a peroxo
(O2 2− ) species
position The
(bond
length
~1.44 Å).
This configuration
consistentbywith
displacement
by waterintothe
equatorial
position in
(bond
length ~1.44
Å). ThisInconfiguration
is consistentsymmetry
with displacement
by water
release peroxide
the absence
of substrate.
the non-crystallographic
related monomer
molecular
oxygeninwas
instead observed,
adjacent
H2Oax. The molecular
oxygen in
this position
to release
peroxide
the absence
of substrate.
In the to
non-crystallographic
symmetry
related
monomer
was proposed
be in
a “pre-bound”
and
not
as wasoxygen
seen ininthe
adjacent
molecular
oxygentowas
instead
observed,position
adjacent
to H
The molecular
this
position
ax . activated
2 Oyet
molecule [51].
Molecules 2017, 22, 596
19 of 26
was proposed to be in a “pre-bound” position and not yet activated as was seen in the adjacent
molecule [51].
The neutron crystal structure of the resting state showed a nearby His157 to be neutral. At lower
pH this His157 would be protonation and positioned to coordinate molecular oxygen in this
“pre-bound” location. This implies that His157 could possibly promote oxygen binding and activation.
DFT calculations were performed on different His157 protonation states and conformations and
showed a strong thermodynamically favored addition of oxygen when His157 is charged. Overall
these studies revealed additional details about oxygen binding and activation not considered before
and the observation of “pre-bound” molecular oxygen indicate a mechanism for H2 O2 formation in
the absence of substrate [51].
6. Ligand Binding and Substrate Specificity through H Bonds
Inhibitor binding to target proteins or enzymes can involve H atoms in different ways: through
direct H bonding, water-mediated H bonding, C-H . . . π interactions, hydrophobic interactions, and
electrostatic interactions (e.g., deprotonated Lys or His residues). In addition to the role of H atoms
in the protein and solvent, the ligand’s own protonation state will have a large effect on solubility
and binding properties. While X-ray crystallography has been a powerful tool in ligand binding
studies and for rational drug design, it has limitations on the details of the information it can provide.
Neutrons are severely underutilized in this field and can fill the knowledge gap by providing atomic
information on ligand protonation and detailed information on how they bind to the target protein.
6.1. Rational Drug Design Using Neutrons as a Guide
HIV protease (PR) is an aspartic protease and an integral part of the viral life cycle. PR cleaves the
pre-cursor proteins Gag and Gag-Pol into functional, catalytic enzymes involved in viral replication
and the structure of the viral capsid. Upon PR inhibition, the resulting virus particles are immature and
non-infectious. As such PR is an important drug target for suppression and control of HIV infection
and progression to AIDS. PR serves as a standout example of successful structure-based drug design
efforts using X-ray crystallography. Since the first reported PR crystal structure, many PR inhibitors
(PI) have been developed and are in use today [15]. Neutron studies of PR:PI complexes can give
useful information to guide next generation structure-based drug design. Such analysis can reveal
unfilled pockets in the active site, the role of water-mediated PI binding, and areas where H-bonding
can be introduced, or removed, by chemical modification of new ligands.
PR is a homodimer with each monomer containing 99 amino acids. The PR active site is made
up of a conserved triad, Asp25-Thr26-Gly27 and Asp250 -Thr260 -Gly270 from the adjacent monomer.
The two Asp residues from opposite monomers point at each other and substrate (peptides/proteins
or inhibitors) binds between the catalytic Asp residues and the flexible flap regions. The first neutron
structure of HIV PR in complex with a clinical drug, amprenavir (APV), was compared to another HIV
PR neutron structure in complex with a non-clinical inhibitor KNI-272 [15,52]. In the APV complex
neutron structure Asp25 is protonated and donates an H bond to the -OH group of APV. In contrast,
Asp250 accepts an H bond from the same -OH. It is interesting that overall the distribution of observed
H bonds in the active site varies from the PR:KNI-272 structure. This shows that the location of H or D
atoms on catalytic Asp residues, as well as the identity of H bond donors/acceptors, can be influenced
strongly by the structure and chemistry of the PI.
Previous studies on ultra-high resolution X-ray structures enabled identification of a number of
proposed H bonds between the ligand and PR [53]. These interactions are expected to enhance, or even
enable, binding interactions. When the nuclear density maps for the PR:APV complex were studied,
the inventory of actual H bonds were quite different from those inferred from the ultra-high resolution
X-ray studies. In fact, some of the details seen in the neutron structure contradicted some of assigned
H bonds and their relative importance in determining ligand binding. The neutron data shows only
two strong H bonds. These are found between APV and catalytic Asp residues, and between the
Molecules 2017, 22, 596
20 of 26
tetrahydrofuran moiety and backbone amides of Asp29 and Asp30. There is also a water-mediated
H bond between Asp300 and the flap regions. The result is three total H bonds and not four as was
deduced from the X-ray data alone [15]. Previous 13 C-NMR data showed that the catalytic Asp residues
have mismatched pKa , and that this affects ionization and subsequent H bond potential. By direct
investigation of the protonation states of Asp25 and Asp250 , the nuclear density maps support the
NMR data in that only Asp25 is protonated [54]. Binding interactions can be improved by ligand
derivatization to match the pKa ’s. This can be achieved by incorporating a strong electron-accepting
group in the APV molecule. Another strategy is to include chemical groups on the PI to directly
engage PR in H bonds and minimize water-mediated H bonding. Such a strategy will also improve
the enthalpy of binding.
In general the results from neutron studies highlight some issues with structure-based drug design.
The overarching one is that X-ray data alone can be misleading from a rational drug design view.
Neutrons can fill this gap by giving information on ionization/protonation state of residues, water,
and H bonds involved. The data can also be used to estimate the importance of interactions based
on distances and angles between donors/acceptors. Another issue is that the chemical and physical
basis as to why some drugs bind better than others may be very subtle [15]. For these reasons it is
important to use multiple methods or approaches to ligand/inhibitor binding interactions. Neutrons
clearly have an important future role in structure-based drug design and this may be one of the best
applications of the technique in structural biology. With new sources and improved instruments the
practical feasibility of using NPX for drug design studies will only increase.
6.2. Ordered Water Network and H Bonds Provide a Template and Specificity for Substrate Binding
Carbohydrate binding modules (CBMs) are domains that are part of carbohydrate-metabolizing
enzymes. They confer substrate specificity and increase enzymatic throughput by aiding enzymatic
recognition of substrates and also by tethering the catalytic part of the enzyme close to the substrate [55].
Some CBMs only bind to a particular carbohydrate while others display more promiscuous binding and
understanding the details will be useful in designing enzyme with specific properties, especially for
industrial application (e.g., biorefinery processes). Neutron protein crystallography is a great tool for
the investigation of substrate specificity. For example, it yields important clues in the case of one such
engineered CBM, X-2 L110F, from a xylanase that is expressed in Rhodothermus marinus [56]. Wild type
X-2 specifically binds to xylan. Mutation of a single residue Leu110 to a Phe (X-2 L110F) changes the
substrate specificity to bind not only xylan, but also xyloglucan, and β-glucan. For the neutron studies
the CBM was complexed to a branched heptasaccharide (XXXG) to probe the interactions to substrate
that enable broader binding to more substrates. Even though H atoms are known to be important
mediators of CBM binding to carbohydrates, there is little direct evidence of their involvement
in H bonds, water mediated binding, π-stacking, and other electrostatic interactions. XXXG is a
heptasaccharide that contains a chain of four glucose molecules and three branched xylose molecules.
Two of the glucose molecules in XXXG are involved in H bonds and a π-stacking interaction to
two Phe residues while the other two are not engaged in any strong or visible interactions. An H
bond was observed between Tyr149 and one of the branching xyloses, XYS1173, and explains why
mutation of this residue disrupts binding ability to branched sugars. XYS1173 is also involved in two
additional H bonds to His146 and Arg115 and in a π-stacking interaction with Pro68. Neutron data
shows that Arg115 is oriented to donate the H bond through a network of interactions that results in
optimal positioning for this interaction. The neutron studies revealed the interactions surrounding
branched xylose XYS1173 and support the inability of the L110F mutant to bind sugars that are further
substituted at this position (e.g., XLLG) [57].
High resolution X-ray crystallographic studies of wild type X-2 and apo X-2 L110F show a number
of ordered water molecules engaged in a network occupy the ligand binding cleft [58]. At least
ten of these waters are displaced upon ligand binding giving an entropic enhancement to binding.
In addition, the oxygen atoms of these waters are precisely replaced by sugar oxygen atoms upon
Molecules 2017, 22, 596
21 of 26
binding, suggesting that waters serve as a chemical template for ligand binding. Xylan and xyloglucan
Molecules 2017, 22, 596
21 of 25
displace a similar number of waters, but XXXG removes an additional three water molecules and
causes compensatory amino acid side chain movements. Three conserved water molecules that
these waters and the H bonds to the protein serve to orient the residues to be in optimal H bonding
form an H bonded network were observed between BGC1171 and several ligand binding residues.
configurations.
Together these waters and the H bonds to the protein serve to orient the residues to be in optimal H
This work highlighted several important considerations in carbohydrate binding interactions.
bonding configurations.
Mainly it is proper positioning of interacting amino acid side chains and H bonds to ordered water
This work highlighted several important considerations in carbohydrate binding interactions.
that give clues as to how a carbohydrate ligand may bind. These factors are also involved in
Mainly it is proper positioning of interacting amino acid side chains and H bonds to ordered water that
understanding substrate specificity and could be exploited to design CBM for specific carbohydrate
give clues as to how a carbohydrate ligand may bind. These factors are also involved in understanding
recognition [57].
substrate specificity and could be exploited to design CBM for specific carbohydrate recognition [57].
6.3. Water Displacement, Drug Ionization, and H Bonds in Mediating Drug Binding
6.3. Water Displacement, Drug Ionization, and H Bonds in Mediating Drug Binding
Human
carbonic
anhydrases
are expressed
in diverse
and
cellthroughout
types throughout
Human
carbonic
anhydrases
are expressed
in diverse
tissuestissues
and cell
types
the body.the
body.
Due
to
the
fundamental
nature
of
the
catalyzed
reaction,
carbon
dioxide
hydration,
CAs are
Due to the fundamental nature of the catalyzed reaction, carbon dioxide hydration, CAs are involved
involved in a number of physiological processes. They are at the center of many pH homeostasis
in a number of physiological processes. They are at the center of many pH homeostasis processes,
processes, gluconeogenesis, bone remodeling, cerebrospinal fluid production and so on. One human
gluconeogenesis, bone remodeling, cerebrospinal fluid production and so on. One human isoform,
isoform, HCA IX, has been found overexpressed in numerous cancer cells and are now an attractive
HCA IX, has been found overexpressed in numerous cancer cells and are now an attractive drug target
drug target for cancer detection, imaging, staging, and treatment [59,60]. Carbonic anhydrase
for cancer detection, imaging, staging, and treatment [59,60]. Carbonic anhydrase inhibitors (CAIs)
inhibitors (CAIs) have been prescribed for decades as diuretics and for the treatment of glaucoma.
have been prescribed for decades as diuretics and for the treatment of glaucoma. Today there are over
Today there are over twenty CAI in clinical use. In recent years their use has been expanded for use
twenty CAI in clinical use. In recent years their use has been expanded for use as anti-epilepsy and
as anti-epilepsy and anti-obesity drugs. Clinically used CAIs have a common motif, a sulfonamide
anti-obesity drugs. Clinically used CAIs have a common motif, a sulfonamide group that coordinates
group that coordinates directly to the active site Zn+2, displacing a critical catalytic water molecule in
directly to the active site Zn+2 , displacing a critical catalytic water molecule in the process [60,61].
the process [60,61]. HCAs share a high degree of sequence and structural identity, making it very
HCAs share a high degree of sequence and structural identity, making it very challenging to design
challenging to design inhibitors that display isoform selectivity. In the case of HCA IX, it would be
inhibitors that display isoform selectivity. In the case of HCA IX, it would be more clinically effective to
more clinically effective to have an inhibitor specific for HCA IX to minimize binding to the other
have an inhibitor specific for HCA IX to minimize binding to the other HCAs with important, central
HCAs with important, central physiological functions. Acetazolamide (AZM) is a classic CAI with a
physiological functions. Acetazolamide (AZM) is a classic CAI with a Ki of 12 nM for HCA II. AZM
Ki of 12 nM for HCA II. AZM has three possible protonation states in solution. It was unknown
has three possible protonation states in solution. It was unknown which form binds to CA under
which form binds to CA under physiological pH conditions. High-resolution X-ray crystallographic
physiological pH conditions. High-resolution X-ray crystallographic studies did not reveal the exact
studies did not reveal the exact nature of the molecule in terms of ionization state or how specifically
nature of the molecule in terms of ionization state or how specifically it binds to the active site of
it binds to the active site of HCA II.
HCA II.
Figure
Comparison
of AZM
MZM
binding
water
displacement
as observed
in neutron
Figure
16. 16.
Comparison
of AZM
andand
MZM
binding
andand
water
displacement
as observed
in neutron
crystallographic
studies.
(A)
AZM
complex;
(B)
MZM
complex
[60].
Figure
reproduced,
with
crystallographic studies. (A) AZM complex; (B) MZM complex [60]. Figure reproduced, with
permission
from
International
Union
of Crystallography
from
permission
from
the the
International
Union
of Crystallography
from
[60].[60].
A 2.0
Å neutron
structure
of HCA
in complex
AZM
revealed
in complex
this complex
A 2.0
Å neutron
structure
of HCA
II inIIcomplex
withwith
AZM
revealed
that that
AZMAZM
in this
is
is
anionic
with
the
unprotonated
sulfonamide
group
coordinated
to
the
active
site
zinc.
In
contrast
anionic with the unprotonated sulfonamide group coordinated to the active site zinc. In contrast
to the to
the negatively
sulfonamide
group,
the acetamido
group is protonated
with
clearly
visible
negatively
chargedcharged
sulfonamide
group, the
acetamido
group is protonated
with clearly
visible
nuclear
nuclear
density
for
the
exchanged
D
atom
at
this
position
(Figure
16A).
The
single
D
atom
on the
density for the exchanged D atom at this position (Figure 16A). The single D atom on the sulfonamide
sulfonamide
group
is
engaged
as
H
bond
donor
to
Thr199.
Thr200
acts
as
a
bifurcated
H
bond
donor
group is engaged as H bond donor to Thr199. Thr200 acts as a bifurcated H bond donor to an ordered
to anmolecule
ordered and
solvent
molecule and
the backbone
of Pro201.
In this
way the
solventtomolecule
solvent
the backbone
of Pro201.
In this way
the solvent
molecule
is poised
act as H is
poised to act as H bond acceptor from the acetamido group of AZM. A total of four water molecules
are displaced upon AZM binding and this will lead to entropic gains in drug binding. Finally the
data also showed two C-H…π interactions. The first is between the AZM thiadiazole ring and
Leu198, the other between the AZM methyl group and Pro201 itself [59,60].
Molecules 2017, 22, 596
22 of 26
bond acceptor from the acetamido group of AZM. A total of four water molecules are displaced upon
AZM binding and this will lead to entropic gains in drug binding. Finally the data also showed two
C-H . . . π interactions. The first is between the AZM thiadiazole ring and Leu198, the other between
the AZM methyl group and Pro201 itself [59,60].
Another neutron study investigated the drug methazolamide (MZM) bound to HCA II [60].
MZM is a methyl derivative of AZM and has superior properties in terms of shelf life, stability, and is
less water-soluble than AZM, making it better suited for use in glaucoma treatment. Unsurprisingly the
data showed that MZM was also unprotonated on the sulfonamide moiety and has similar coordination
to the zinc as AZM (Figure 16B). MZM displaced a total of seven water molecules, compared to four
by AZM. Three of the water molecules are common between the two structures. This means that in the
AZM complex there are still a number of ordered, H bonded water molecules that are absent in the
room temperature MZM complex. In the MZM situation there is an enthalpic trade-off that involves
the breaking of so many H bonds to expel the waters, this balances the entropic gain of releasing water
to the bulk solvent. This explains why the Ki for HCA II of MZM is still very similar as AZM (i.e.,
~10 nM). The added methyl group on the thiadiazole group of MZM disrupts the water-mediated H
bond that was observed between AZM and Thr200 and Pro201 [59,60]. Detailed knowledge of H bonds
and water networks in ligand binding is valuable and neutron crystallography will have increasing
importance in rational drug design that can exploit this knowledge. Having quantitative information
on H bonds can be used for example to modify or derivatize compounds to introduce compensatory
interactions or to produce a desired ionization form of the ligand.
7. Conclusions
This review highlights specific areas of structural enzymology and general protein biochemistry
where neutrons have made a large impact. Due to their unique nature and interactions with biological
materials, neutrons have the ability to give high resolution, detailed, unambiguous information on
H atoms and their all-important interaction, the H bond. Neutron crystallography has provided the
opportunity to “see” into the heart of the enzyme active sites or ligand binding regions, and often
brings new insights. Neutrons have contributed direct evidence of amino acid involvement in catalysis
or drug binding, how waters are organized and oriented, and even how some waters are activated to
perform catalysis. Arguably the most important thing neutrons can show us is the H bond and how it
ties the protein, solvent, and/or ligand together. As illustrated in this review, neutron crystallography
complements results from NMR spectroscopy, X-ray crystallography, modeling, and computational
chemistry. In recent years, significant advances have been made in the fields of DFT calculations
and 1 H-NMR spectroscopy for the study of H bonds [62,63]. It can be expected that theoretical and
experimental information gained from all these methods will impact how protein structures are refined
and analyzed.
Neutron crystallography is under-utilized by the general structural biology community due
to perceived issues with sample preparation and acquiring beam time. With current molecular
biology tools and equipment for protein production, purification and crystallization, there are now far
fewer technical limitations to preparing sufficient material to grow large enough crystals for neutron
diffraction experiments. Current instruments routinely collect high quality data from crystals that are
0.1–1 mm3 and smaller in volume, a range that is readily attainable for many systems. The low number
of instruments and highly competitive nature of being awarded beam time, especially compared to
macromolecular X-ray beamlines, represents the true bottleneck for current and new users. With the
advent of the new European Spallation Source (Lund, Sweden), and the planned Macromolecular
Diffractometer NMX there, data collection and sample requirements will dramatically improve for the
user community.
Acknowledgments: SZF and EO are partially funded by Interreg ESS & MAX IV Cross Border Science and Society
and SZF by the SINE2020 grant (under Horizon 2020 grant agreement No. 654000).
Molecules 2017, 22, 596
23 of 26
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Hubbard, R.E.; Haider, M.K. Hydrogen bonds in proteins: Role and strength. In Encyclopedia of Life Science;
John Wiley & Sons Ltd.: Chichester, UK, 2010.
Martin, T.W.; Derewenda, S. The name is bond—H bond. Nat. Struct. Biol. 1999, 6, 403–405. [CrossRef]
[PubMed]
McDonald, I.K.; Thornton, J.M. Satisfyin hydrogen bonding potential in proteins. J. Mol. Biol. 1994, 238,
777–793. [CrossRef] [PubMed]
Tuttle, T.; Grafenstein, J.; Wu, A.; Kraka, E.; Cremer, D. Analysis of the NMR spin-spin coupling mechanism
across a Hbond: Nature of the H-bond in proteins. J. Phys. Chem. 2004, 108, 1115–1129. [CrossRef]
Baker, E.N.; Hubbard, R.E. Hydrogen bonding in globular proteins. Prog. Biophys. Mol. Biol. 1984, 44, 97–179.
[CrossRef]
Kumar, M.; Balaji, P.V. C-H . . . pi interactions in proteins: Prevalence, pattern of occurrence, residue
propensities, location, and contribution to protein stability. J. Mol. Model. 2014, 20, 2136–2150. [CrossRef]
[PubMed]
Plevin, M.J.; Bryce, D.L.; Boisbouvier, J. Direct detection of CH/π interactions in proteins. Nat. Chem. 2010, 2,
466–471. [CrossRef] [PubMed]
Torshin, I.Y.; Weber, I.T.; Harrison, R.W. Geometric criteria of hydrogen bonds in proteins and identification
of “bifurcated” hydrogen bonds. Prot. Eng. 2002, 15, 359–363. [CrossRef]
Michalczyk, R.; Unkefer, C.J.; Bacik, J.P.; Schrader, T.E.; Ostermann, A.; Kovalevsky, A.Y.; McKenna, R.;
Fisher, S.Z. Joint neutron crystallographic and NMR solution studies of Tyr residue ionization and hydrogen
bonding: Implications for enzyme-mediated proton transfer. Proc. Natl. Acad. Sci. USA 2015, 112, 5673–5678.
[CrossRef] [PubMed]
Yokoyama, T.; Mizuguchi, M.; Ostermann, A.; Kusaka, K.; Niimura, N.; Schrader, T.E.; Tanaka, I. Protonation
State and Hydration of Bisphosphonate Bound to Farnesyl Pyrophosphate Synthase. J. Med. Chem. 2015, 58,
7549–7556. [CrossRef] [PubMed]
Blakeley, M.P.; Hasnain, S.S.; Antonyuk, S.V. Sub-atomic resolution X-ray crystallography and neutron
crystallography: Promise, challenges and potential. IUCrJ 2015, 2, 464–474. [CrossRef] [PubMed]
O’Dell, W.B.; Bodenheimer, A.M.; Meilleur, F. Neutron protein crystallography: A complementary tool for
locating hydrogens in proteins. Arch. Biochem. Biophys. 2016, 602, 48–60. [CrossRef] [PubMed]
NIST Website for Neutron Scattering Lengths. Available online: http://www.ncnr.nist.gov/resources/nlengths/ (accessed on 17 January 2017).
Niimura, N.; Bau, R. Neutron protein crystallography: Beyond the folding structure of biological
macromolecules. Acta Crystallogr. A 2008, 64, 12–22. [CrossRef] [PubMed]
Weber, I.T.; Waltman, M.J.; Mustyakimov, M.; Blakeley, M.P.; Keen, D.A.; Ghosh, A.K.; Langan, P.;
Kovalevsky, A.Y. Joint X-ray/neutron crystallographic study of HIV-1 protease with clinical inhibitor
amprenavir: Insights for drug design. J. Med. Chem. 2013, 56, 5631–5635. [CrossRef] [PubMed]
Fisher, S.J.; Blakeley, M.P.; Howard, E.I.; Petit-Haertlein, I.; Hartlein, M.; Mitschler, A.; Cousido-Siah, A.;
Salvay, A.G.; Popov, A.; Muller-Dieckman, C.; et al. Perdeuteration: Improved visualization of solvent
structure in neutron macromolecular crystallography. Acta Crystallogr. D 2014, 70, 3266–3272. [CrossRef]
[PubMed]
Langan, P.; Chen, J.C.-H. Seeing the chemistry in biology with neutron crystallography. Phys. Chem.
Chem. Phys. 2013, 15, 13705–13712. [CrossRef] [PubMed]
Chen, J.C.-H.; Unkefer, C.J. Fifteen years of the Protein Crystallography Station: The coming of age of
macromolecular neutron crystallography. IUCrJ 2017, 4, 72–86. [CrossRef] [PubMed]
Blakeley, M.P.; Teixeira, S.C.; Petit-Haertlein, I.; Hazemann, I.; Mitschler, A.; Haertlein, M.; Howard, E.;
Podjarny, A.D. Neutron macromolecular crystallography with LADI-III. Acta Crystallogr. D 2010, 66,
1198–1205. [CrossRef] [PubMed]
Chen, J.C.-H.; Hanson, B.L.; Fisher, S.Z.; Langan, P.; Kovalevsky, A.Y. Direct observation of hydrogen atom
dynamics and interactions by ultrahigh resolution neutron protein crystallography. Proc. Natl. Acad. Sci. USA
2012, 109, 15301–15306. [CrossRef] [PubMed]
Molecules 2017, 22, 596
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
24 of 26
Schmidt, A.; Teeter, M.; Weckert, E.; Lamzin, V.S. Crystal structure of small protein crambin at 0.48 Å
resolution. Acta Crystallogr. F 2011, 67, 424–428. [CrossRef] [PubMed]
Rosenbaum, G.; Ginell, S.L.; Chen, J.C.-H. Energy optimization of a regular macromolecular crystallography
beamline for ultra-high-resolution crystallography. J. Synchrotron Radiat. 2015, 22, 172–174. [CrossRef]
[PubMed]
Silverman, D.N.; McKenna, R. Solvent-mediated proton transfer in catalysis by carbonic anhydrase.
Acc. Chem. Res. 2007, 40, 669–675. [CrossRef] [PubMed]
Fisher, Z.; Kovalevsky, A.Y.; Mustyakimov, M.; Silverman, D.N.; McKenna, R.; Langan, P. Neutron structure
of human carbonic anhydrase II: A hydrogen-bonded water network “switch” is observed between pH 7.8
and 10.0. Biochemistry 2011, 50, 9421–9423. [CrossRef] [PubMed]
Dedecker, P.; de Schryver, F.C.; Hofkens, J. Fluorescent proteins: Shine on, you crazy diamond. J. Am.
Chem. Soc. 2013, 135, 2387–2402. [CrossRef] [PubMed]
Langan, P.S.; Close, D.W.; Coates, L.; Rocha, R.C.; Ghosh, K.; Kiss, C.; Waldo, G.; Freyer, J.; Kovalevsky, A.;
Bradbury, A.R. Evolution and characterization of a new reversibly photoswitching chromogenic protein,
Dathail. J. Mol. Biol. 2016, 428, 1776–1789. [CrossRef] [PubMed]
Anderson, S.; Crosson, S.; Moffat, K. Short hydrogen bonds in photoactive yellow protein. Acta Crystallogr. D
2004, 60, 1008–1016. [CrossRef] [PubMed]
Yamaguchi, S.; Kamikubo, H.; Kurihara, K.; Kuroki, R.; Niimura, N.; Shimizu, N.; Yamazaki, Y.; Kataoka, M.
Low-barrier hydrogen bond in photoactive yellow protein. Proc. Natl. Acad. Sci. USA 2009, 106, 440–444.
[CrossRef] [PubMed]
Fisher, S.Z.; Anderson, K.; Henning, R.; Moffat, K.; Langan, P.; Thiyagarajan, P.; Schultz, A.J. Neutron and
X-ray structural studies of short hydrogen bonds in photoactive yellow protein (PYP). Acta Crystallogr. D
2007, 63, 1178–1184. [CrossRef] [PubMed]
Tamada, T.; Kinoshita, T.; Kurihara, K.; Adachi, M.; Ohhara, T.; Imai, K.; Kuroki, R.; Tada, T. Combined
high-resolution neutron and X-ray analysis of inhibited elastase confirms the active site oxyanion hole but
rules against a low-barrier hydrogen bond. J. Am. Chem. Soc. 2009, 131, 11033–11040. [CrossRef] [PubMed]
Cleland, W.W.; Kreevoy, M.M. Low-barrier hydrogen bonds and enzymic catalysis. Science 1994, 264,
1887–1890. [CrossRef] [PubMed]
Kuhn, P.; Knapp, M.; Solitis, S.M.; Ganshaw, G.; Thoene, M.; Bott, M. The 0.78 Å structure of a serine protease:
Bacillus lentus subtilisin. Biochemistry 1998, 37, 13446–13452. [CrossRef] [PubMed]
Schmidt, A.; Jelsch, C.; Ostergaard, P.; Rypniewski, W.; Lamzin, V.S. Trypsin revisited: Crystallography
at sub-atomic resolution and quantum chemistry revealing details of catalysis. J. Biol. Chem. 2003, 278,
43357–43362. [CrossRef] [PubMed]
Xu, Y.; Wu, J.; Gorenstein, D.; Braun, W. Automated 2D NOESY assignment and structure calculation
of crambin(S22/I25) with the self-correcting distance geometry based NOAH/DIAMOD programs.
J. Magn. Reson. 1999, 136, 76–85. [CrossRef] [PubMed]
Kovalevsky, A.Y.; Hanson, L.; Fisher, S.Z.; Mustyakimov, M.; Mason, S.A.; Forsyth, V.T.; Blakeley, M.P.;
Keen, D.A.; Wagner, T.; Carrell, H.L.; et al. Metal ion roles and the movement of hydrogen during reaction
catalyzed by D-xylose isomerase: A joint X-ray and neutron diffraction study. Structure 2010, 18, 688–699.
[CrossRef] [PubMed]
Kovalevsky, A.Y.; Hanson, B.L.; Mason, S.A.; Yoshida, T.; Fisher, S.Z.; Mustyakimov, M.; Forsyth, V.T.;
Blakeley, M.P.; Keen, D.A.; Langan, P. Identification of the elusive hydronium ion exchanging roles with
a proton in an enzyme at lower pH values. Angew. Chem. Int. Ed. Engl. 2011, 50, 7520–7523. [CrossRef]
[PubMed]
Asboth, B.; Naray-Szabo, G. Mechanism of action of D-xylose isomerase. Curr. Protein Pept. Sci. 2000, 1,
237–254. [CrossRef] [PubMed]
Carrell, H.L.; Glusker, J.P.; Burger, P.V.; Manfre, F.; Tritsch, D.; Biellmann, J.F. X-ray analysis of D-xylose
isomerase at 1.9 Å: Native enzyme in complex with substrate and with a mechanism-designed inactivator.
Proc. Natl. Acad. Sci. USA 1989, 86, 4440–4444. [CrossRef] [PubMed]
Cuypers, M.G.; Mason, S.A.; Blakeley, M.P.; Mitchell, E.P.; Haertlein, M.; Forsyth, V.T. Near-atomic resolution
neutron crystallography on perdeuterated Pyrococcus furiosus rubredoxin: Implication of hydronium ions
and protonation state equilibria in redox changes. Angew. Chem. 2013, 52, 1022–1025. [CrossRef] [PubMed]
Molecules 2017, 22, 596
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
25 of 26
Kwon, H.; Basran, J.; Casadei, C.M.; Fielding, A.J.; Schrader, T.E.; Ostermann, A.; Devos, J.M.; Aller, P.;
Blakeley, M.P.; et al. Direct visualization of a Fe(IV)-OH intermediate in a heme enzyme. Nat. Commun. 2016,
7, 13445. [CrossRef] [PubMed]
Casadei, C.M.; Gumiero, A.; Metcalfe, C.L.; Murphy, E.J.; Basran, J.; Concilio, M.G.; Teixeira, S.C.;
Schrader, T.E.; Fielding, A.J.; Ostermann, A.; et al. Heme enzymes. Neutron cryo-crystallography captures
the protonation state of ferryl heme in a peroxidase. Science 2014, 345, 193–197. [CrossRef] [PubMed]
Oksanen, E.; Blakeley, M.P.; El-Hajji, M.; Ryde, U.; Budayova-Spano, M. The neutron structure of urate
oxidase resolves a long-standing mechanistic conundrum and reveals unexpected changes in protonation.
PLoS ONE 2014, 9, e86651. [CrossRef] [PubMed]
Gabison, L.; Chopard, C.; Colloc’h, N.; Peyrot, F.; Castro, B.; Hajji, M.E.; Altarsha, M.; Monard, G.;
Chiadmi, M.; Prangé, T. X-ray, ESR, and quantum mechanics studies unravel a spin well in the cofactor-less
urate oxidase. Proteins 2011, 79, 1964–1976. [CrossRef] [PubMed]
Chen, Y.; Delmas, J.; Sirot, J.; Shoichet, B.; Bonnet, R. Atomic resolution structures of the CTX-M β-lactamases:
Extended spectrum activities form increased mobility and decreased stability. J. Mol. Biol. 2005, 348, 349–362.
[CrossRef] [PubMed]
Vandavasi, V.G.; Weiss, K.L.; Cooper, J.B.; Erskine, P.T.; Tomanicek, S.J.; Ostermann, A.; Schrader, T.E.;
Ginell, S.L.; Coates, L. Exploring the Mechanism of β-Lactam Ring Protonation in the Class A β-lactamase
Acylation Mechanism Using Neutron and X-ray Crystallography. J. Med. Chem. 2016, 59, 474–479. [CrossRef]
[PubMed]
Strynadka, N.C.; Adachi, H.; Jensen, S.E.; Johns, K.; Sielecki, A.; Betzel, C.; Sutoh, K.; James, M.N. Molecular
structure of the acyl-enzyme intermediate in beta-lactam hydrolysis at 1.7 A resolution. Nature 1992, 359,
700–705. [CrossRef] [PubMed]
Blum, M.-M.; Mustyakimov, M.; Ruterjans, H.; Kehe, K.; Schoenborn, B.P.; Langan, P.; Chen, J.C.-H. Rapid
determination of hydrogen positions and protonation states of diisopropyl fluorophosphatase by joint
neutron and X-ray diffraction refinement. Proc. Natl. Acad. Sci. USA 2009, 106, 713–718. [CrossRef] [PubMed]
Blum, M.-M.; Lohr, F.; Richardt, A.; Ruterjans, H.; Chen, J.C.-H. Binding of a designed substrate analogue to
diisopropyl fluorophosphatase: Implications for the phosphotriesterase mechanism. J. Am. Chem. Soc. 2006,
128, 12750–12757. [CrossRef] [PubMed]
Chen, J.C.-H.; Mustyakimov, M.; Schoenborn, B.P.; Langan, P.; Blum, M.M. Neutron structure and mechanistic
studies of diisopropyl fluorophosphatase (DFPase). Acta Crystallogr. D 2010, 66, 1131–1138. [CrossRef]
[PubMed]
Kjaergaard, C.H.; Qayyum, M.F.; Wong, S.D.; Xu, F.; Hemsworth, G.R.; Walton, D.J.; Young, N.A.; Davies, G.J.;
Walton, P.H.; Johansen, K.S.; et al. Spectroscopic and computational insight into the activation of O2 by the
mononuclear Cu center in polysaccharide monooxygenases. Proc. Natl. Acad. Sci. USA 2014, 111, 8797–8802.
[CrossRef] [PubMed]
O’Dell, W.B.; Agarwal, P.K.; Meilleur, F. Oxygen Activation at the Active Site of a Fungal Lytic Polysaccharide
Monooxygenase. Angew. Chem. Int. Ed. 2017, 56, 767–770. [CrossRef] [PubMed]
Adachi, M.; Ohhara, T.; Kurihara, K.; Tamada, T.; Honjo, E.; Okazaki, N.; Arai, S.; Shoyama, Y.; Kimura, K.;
Matsumura, H.; et al. Structure of HIV-1 protease in complex with potent inhibitor KNI-272 determined
by high-resolution X-ray and neutron crystallography. Proc. Natl. Acad. Sci. USA 2009, 106, 4641–4646.
[CrossRef] [PubMed]
Kovalevsky, A.Y.; Liu, F.; Leshchenko, S.; Ghosh, A.K.; Louis, J.M.; Harrison, R.W.; Weber, I.T. Ultrahigh
resolution crystal structure of HIV-1 protease mutant reveals two binding sites for clinical inhibitor TMC114.
J. Mol. Biol. 2006, 363, 161–173. [CrossRef] [PubMed]
Wang, Y.X.; Freedberg, D.I.; Yamazaki, T.; Wingfield, P.T.; Stahl, S.J.; Kaufman, J.D.; Kiso, Y.; Torchia, D.A.
Solution NMR evidence that the HIV-1 protease catalytic aspartyl groups have different ionization states
in the complex formed with the asymmetric drug KNI-272. Biochemistry 1996, 35, 9945–9950. [CrossRef]
[PubMed]
Hervé, C.; Rogowski, A.; Blake, A.W.; Marcus, S.E.; Gilbert, H.J.; Knox, J.P. Carbohydrate-binding modules
promote the enzymatic deconstruction of intact plant cell walls by targeting and proximity effects. Proc. Natl.
Acad. Sci. USA 2010, 107, 15293–15298. [CrossRef] [PubMed]
Molecules 2017, 22, 596
56.
57.
58.
59.
60.
61.
62.
63.
26 of 26
Karlsson, E.N.; Bartonek-Roxå, E.; Holst, O. Cloning and sequence of a thermostable multidomain xylanase
from the bacterium Rhodothermus marinus. Biochim. Biophys. Acta Gene Struct. Expr. 1997, 1353, 118–124.
[CrossRef]
Fisher, S.Z.; von Schantz, L.; Håkansson, M.; Logan, D.T.; Ohlin, M. Neutron crystallographic studies reveal
hydrogen bond and water-mediated interactions between a carbohydrate-binding module and its bound
carbohydrate ligand. Biochemistry 2015, 54, 6435–6438. [CrossRef] [PubMed]
Von Schantz, L.; Håkansson, M.; Logan, D.T.; Walse, B.; Österlin, J.; Karlsson, E.N.; Ohlin, M. Structural basis
for carbohydrate-binding specificity—A comparative assessment of two engineered carbohydrate-binding
modules. Glycobiology 2012, 22, 948–961. [CrossRef] [PubMed]
Fisher, S.Z.; Aggarwal, M.; Kovalevsky, A.Y.; Silverman, D.N.; McKenna, R. Neutron diffraction of
acetazolamide-bound human carbonic anhydrase II reveals atomic details of drug binding. J. Am. Chem. Soc.
2012, 134, 14726–14729. [CrossRef] [PubMed]
Aggarwal, M.; Kovalevsky, A.Y.; Velazquez, H.; Fisher, S.Z.; Smith, J.C.; McKenna, R. Neutron structure of
human carbonic anhydrase II in complex with methazolamide: Mapping the solvent and hydrogen-bonding
patterns of an effective clinical drug. IUCrJ 2016, 3, 319–325. [CrossRef]
Supuran, C.T.; Scozzafava, A.; Conway, J. Carbonic Anhydrase—Its Inhibitors and Activators; CRC Press:
Boca Raton, FL, USA, 2004; Volume 1.
Deringer, V.L.; Hoepfner, V.; Dronskowski, R. Accurate hydrogen positions in organic crystals: Assessing
a quantum-chemical aide. Cryst. Growth Des. 2012, 12, 1014–1021. [CrossRef]
Siskos, M.G.; Choudhary, M.I.; Gerothanassis, I.P. Hydrogen Atomic Positions of O–H . . . O Hydrogen
Bonds in Solution and in the Solid State: The Synergy of Quantum Chemical Calculations with 1H-NMR
Chemical Shifts and X-ray Diffraction Methods. Molecules 2017, 22, 415. [CrossRef] [PubMed]
© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).