Article
pubs.acs.org/Biomac
Structure and Hydration of Highly-Branched, Monodisperse
Phytoglycogen Nanoparticles
Jonathan D. Nickels,†,‡,∥,∇ John Atkinson,§,∇ Erzsebet Papp-Szabo,§ Christopher Stanley,∥
Souleymane O. Diallo,⊥ Stefania Perticaroli,§ Benjamin Baylis,§ Perry Mahon,§ Georg Ehlers,#
John Katsaras,†,‡,∥ and John R. Dutcher*,§
†
Joint Institute for Neutron Sciences, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States
Department of Physics, University of Tennessee, Knoxville, Tennessee 37996, United States
§
Department of Physics, University of Guelph, Guelph, Ontario N1G 2W1, Canada
∥
Biology and Soft Matter Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States
⊥
Chemical and Engineering Materials Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, United States
#
Quantum Condensed Matter Division, Oak Ridge National Laboratory, P.O. Box 2008, Oak Ridge, Tennessee 37831, United States
‡
S Supporting Information
*
ABSTRACT: Phytoglycogen is a naturally occurring polysaccharide nanoparticle made up of extensively branched glucose monomers. It has a number of
unusual and advantageous properties, such as high water retention, low viscosity,
and high stability in water, which make this biomaterial a promising candidate for
a wide variety of applications. In this study, we have characterized the structure
and hydration of aqueous dispersions of phytoglycogen nanoparticles using
neutron scattering. Small angle neutron scattering results suggest that the
phytoglycogen nanoparticles behave similar to hard sphere colloids and are
hydrated by a large number of water molecules (each nanoparticle contains
between 250% and 285% of its mass in water). This suggests that phytoglycogen
is an ideal sample in which to study the dynamics of hydration water. To this
end, we used quasielastic neutron scattering (QENS) to provide an independent
and consistent measure of the hydration number, and to estimate the retardation
factor (or degree of water slow-down) for hydration water translational motions.
These data demonstrate a length-scale dependence in the measured retardation factors that clarifies the origin of discrepancies
between retardation factor values reported for hydration water using different experimental techniques. The present approach can
be generalized to other systems containing nanoconfined water.
■
INTRODUCTION
Phytoglycogen is a highly branched polysaccharide, produced in
the form of monodisperse nanoparticles by some varieties of
plants, in which it is used to store glucosea role similar to
that of glycogen in animals. Both phytoglycogen and glycogen
consist of regularly branched, flexible linear chains of glucose,
with growth proceeding in a manner similar to that for
synthetic dendrimers.1 The molecules grow outward, until the
density of glucose units is such that the enzymes that are
required to catalyze further growth and branching are sterically
blocked. Computer simulations,1−3 self-consistent field theory
calculations4,5 and small angle neutron scattering (SANS)
experiments6 on high generation dendrimer molecules have all
shown that further growth involves the folding of the flexible
chains toward the center of the molecule, resulting in high
molecular weight, monodisperse molecules with a highly
branched outer surface and a dense core.
The dendrimeric structure of phytoglycogen (Figure 1) is
markedly different from other natural polysaccharides of similar
© 2016 American Chemical Society
molecular weights. Significant work has been done to
understand the basic structure of glycogen-like molecules,8−14
as well as the role of enzymes in their growth.15−17 Their
structure7 leads to a number of remarkable physical properties
such as monodispersity, high water retention, low viscosity and
exceptional stability in water (measured in years), the ability to
stabilize and disperse bioactive compounds, and the capacity to
form films on surfaces. These properties highlight the potential
of this biological nanomaterial to be the foundation of new
technologies and therapies.
In addition to its practical applications, aqueous dispersions
of monodisperse phytoglycogen nanoparticles can be used to
test theories of colloidal dispersions and study hydration water
physics. The dynamics and thermodynamics of the hydration
water population can drive important processes in biological
Received: October 15, 2015
Revised: January 5, 2016
Published: January 30, 2016
735
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
Figure 1. (a) Schematic of phytoglycogen structure. The molecule is made up of glucose monomers joined by α-1,4-glycosidic linkages with
branching, on average, every 13 monomers7 through α-1,6-glycosidic linkages. (b) Atomic force microscopy topography images of phytoglycogen
nanoparticles on mica. The main image was measured in air and contains a large number of nanoparticles, with the image width corresponding to 2.5
μm. The inset shows an isolated nanoparticle measured in water, with a line profile through its center indicating the flattened nature of the adsorbed
nanoparticles (range of the horizontal axis is 90 nm, and range of the vertical axis is 12.5 nm).
materials, such as the inverse temperature transition in elastin.18
In addition, there is a discrepancy in the reported retardation
factors, ξ, for hydration water (or slow-down of molecular
motions relative to bulk water). NMR and dielectric spectroscopy measurements have reported a 2−3 fold decrease in the
rotation of water molecules near biomolecules19−22 and small
molecules,21,23 compared to bulk water. However, differences of
orders of magnitude in the slow-down of hydration water have
been reported using time-resolved fluorescence spectroscopy24−28 and other dielectric measurements.29 The large
population of hydration water within hydrated phytoglycogen
enables neutron measurements to resolve two water
populations, namely, bulk water and hydration water. Because
neutron scattering probes multiple length scales, these
measurements may help to explain these discrepancies by
allowing the determination of ξ as a function of length scale.
In the current study we used neutron scattering to investigate
the relationship between phytoglycogen and water by
measuring the structure and dynamics of aqueous dispersions
of phytoglycogen as a function of temperature and concentration. We carried out structural measurements to measure
nanoparticle size, evaluate their density, quantify the particle-toparticle spacing, and determine their water content. Our results
show that the nanoparticles have a radius of 17.4 ± 1.6 nm, and
are of uniform density, a feature consistent with the back
folding of the glucose chains toward the center of the molecule.
This has been observed at very high generations of dendrimeric
growth in computer simulations and self-consistent field theory
calculations.1 Analysis of the particle-to-particle spacing
indicates the onset of jamming30,31 for nanoparticle concentrations near 25% w/w. Taken together with our observation of
uniform scattering length density, these results suggest that the
nanoparticles can be treated as hard spheres, a notion that is
consistent with the dramatic increase in viscosity at
concentration values >25% w/w (Figure S3). By measuring
the neutron scattering length density of phytoglycogen, we
were able to compute the monomer volume and, in turn,
estimate the water content of the fully hydrated nanoparticles
(22.5 water molecules per monomer, or 5.85 × 105 water
molecules per nanoparticle). We also exploited the high water
content of phytoglycogen nanoparticles to analyze the
dynamics of hydration water, a population that is normally
difficult to measure. This analysis provided an independent,
complementary, and consistent determination of the water
content of the phytoglycogen nanoparticles. Dynamics experi-
ments also show that the hydration water translation is
subdiffusive, occurring, on average, ∼ 5.8 times slower than
that of bulk water. These experiments indicate a clear qdependence, showing that the measured retardation factor
depends on the length scale being probed. This length-scale
dependence of the retardation factor had not been demonstrated previously within a single sample, and may help to
reconcile the often conflicting range of hydration water
retardation factors reported in the literature.32 The present
experiments not only provide key physical properties of
phytoglycogen that are necessary for the development of new
biobased technologies and therapies, but also demonstrate the
utility of phytoglycogen as a model system able to provide a
detailed understanding of hydration water properties.
■
MATERIALS AND METHODS
Extraction and Purification of Phytoglycogen Nanoparticles.
Frozen sweet corn kernels (75% moisture content) were mixed with
deionized water at 20 °C, and pulverized in a blender at 3000 rpm for
3 min. The resulting mush was centrifuged at 12 000 × g for 15 min at
4 °C. The combined supernatant fraction was subjected to cross-flow
filtration using using poly(ether sulfone) (PES) microfiltration crossflow filters (Sartorius Stedim Biotech, 3021860604O-SG, 0.1 mm pore
size), which removed fibrous material, as well as most of the proteins
and lipids. The retentate fraction was mixed with 2.5 volumes of 95%
ethanol and centrifuged at 8000 × g for 10 min at 4 °C. The pellet
containing phytoglycogen was dried in an oven at 50 °C for 24 h and
then milled to 45 mesh. The nanoparticle dispersions were further
purified by adding phytoglycogen prepared using the above procedure
to Millipore water (1% w/w), mixing for several hours, and flowing
across Omega modified PES tangential flow ultrafiltration membranes
(Pall, OA500C12 500 kDa and OA300C12 300 kDa filters). The
resulting retentate was lyophilized. The overall yield of the
ultrafiltration process was ∼0.2 (mass of output ultrafiltered
phytoglycogen divided by mass of input microfiltered phytoglycogen).
To assess the purity of the phytoglycogen nanoparticles obtained from
the ultrafiltration procedure, we performed the Bradford protein
binding microassay,32 using bovine serum albumin (BSA) as a
standard. The residual protein content was less than 0.1%. Details of
additional characterization of the phytoglycogen nanoparticles using
size exclusion chromatography, atomic force microscopy, and dynamic
light scattering are also provided in the Supporting Information.
Preparation of Phytoglycogen Nanoparticle Dispersions.
Nanoparticles were dispersed in H2O/D2O at phytoglycogen
concentrations ranging from 0% to 24.4% w/w. Samples were allowed
to equilibrate for at least 1 h before experimentation. The sample
volume was ∼0.3 mL for the SANS measurement, and ∼2 mL for the
quasielastic neutron scattering (QENS) measurement (annular sample
736
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
cell, 1 mm inner spacer). Sample concentrations are reported as a
percentage in weight/weight (% w/w), defined as the phytoglycogen
mass divided by the total mass of phytoglycogen and solvent. Samples
used for neutron studies as a function of % D2O were formulated at
the same concentrations in weight per volume for H2O and D2O, and
are reported as their equivalent % w/w at 100% D2O.
Neutron Scattering. SANS measurements were performed on
phytoglycogen dispersions in a range of concentrations and H2O/D2O
ratios at the Extended Q-range Small Angle Neutron Spectrometer
(EQ-SANS),33 which is located at the Spallation Neutron Source
(SNS) [Oak Ridge National Laboratory (ORNL)]. Samples were
loaded in 2 cm diameter by 1 mm thick circular quartz cells (Hellma),
which were then placed in a temperature regulated 45-position sample
holder. Data were collected in 60 Hz mode using three instrumental
configurations, namely: 1.3 m sample-to-detector distance, using 5.5−9
Å wavelength neutrons; 4.0 m sample-to-detector distance using 10−
13.4 Å wavelength neutrons; and 4.0 m sample-to-detector distance
using 12.5−15 Å wavelength neutrons. These three configurations
yielded a total usable q-range between ∼0.005 and 0.35 Å−1. Twodimensional scattering data were reduced using the Mantid34 software
environment and were normalized to a Porasil standard to establish an
absolute scale. Data were corrected for pixel sensitivity, dark current,
and sample transmission. Background scattering from the respective
solvents was subtracted from the one-dimensional intensity versus q
data.
Scattering data were analyzed in three ways; (1) the total scattering
invariant, Q*, to obtain information about particle density and
monomer volume, (2) the form factor, F(q), for particle shape and
size, and (3) the structure factor, S(q), for interparticle spacing. The
total scattering invariant was evaluated from scattering measurements
made in the q-range from 0.005 Å−1 < q < 0.07 Å−1, according to
Q* =
∫ q2I(q) dq
the atomic nuclei in the sample; the NSLD is a convenient
quantity used to describe this atomic property. NSLD is
calculated as the sum of the atomic scattering lengths bi of all
“i” atoms in a molecule divided by the molecular volume vm
(NSLD = Σbi/vm). We are thus able to predict the NSLD of a
given molecule based on its volume and atomic content, and
compare it to that measured by SANS, or use the SANS data to
compute molecular volumes based on the known chemical
composition.
Structural information was obtained by performing SANS on
dispersions of phytoglycogen in 100% D2O as a function of
phytoglycogen concentration (0% to 24.4% w/w) at 295 K
(Figure 2). We found that the best fit to the low concentration
data (1.0% w/w) was obtained by using a model of a sphere of
uniform density and a Shultz distribution for the particle
polydispersity.39 We also considered a number of other models
(1)
The scattering intensity I(q) is defined as
I(q) = S(q)|F(q)|2 + bkg
(2)
where S(q) represents the scattering from interparticle interference,
F(q) is the form factor, and bkg describes the q-independent
incoherent background. The form factor describes the sample
structure and scales with the difference in neutron scattering length
density ρ between the solvent and particle, as well as concentration.
This implies that the total scattering is expected to follow a ρ2
dependence for systems with uniform neutron scattering length
density (NSLD). Fits to the SANS data were performed using the
SasView package.35
Dynamics measurements were performed on dispersions of
phytoglycogen at 10.5%, 19.7%, and 24.4% w/w in D2O, as well as
pure D2O, in the temperature range of 100 to 300 K. The
Backscattering Spectrometer (BASIS)36 at SNS was used for QENS
measurements in an energy range of ±120 μeV, with a resolution of
3.5 μeV and a q-range of 0.2−2.0 Å−1. BASIS was also used to obtain
elastic fixed window scattering data at a resolution of ±3.5 μeV. The
Cold Neutron Chopper Spectrometer (CNCS)37 at SNS was used for
measurements of D2O at 295 K in the energy range up to 47 meV,
with a resolution of ∼50 μeV in the q-range from 0.5 to 5 Å−1. All
BASIS and CNCS spectra were corrected for a background and sample
holder, and were normalized to the scattering from vanadium. No
multiple scattering corrections were used. Data reduction for BASIS
and CNCS spectra was performed using the Mantid34 software
environment and fitted in intensity using the DAVE software suite.38
Figure 2. SANS data for phytoglycogen dispersions. (a) Scattered
intensity versus scattering wavevector q as a function of phytoglycogen
concentration. Intensity data were normalized by the concentration, C.
The data for the lowest concentration were fitted to obtain the radius
and NSLD of the phytoglycogen nanoparticles (error reported is σ2).
(b) The emergence of a structure factor from interparticle scattering
was used to determine the interparticle distance. The inset shows
offset S(q) curves along with their respective fits. (c) The interparticle
distance approached the nanoparticle diameter at high phytoglycogen
concentrations.
■
RESULTS AND DISCUSSION
Structure of Phytoglycogen. SANS was used to interrogate the structure of aqueous dispersions of phytoglycogen.
In SANS, elastically scattered neutrons are used to determine
structural information on length scales ranging from ∼1 to
∼100 nm. The scattered intensity depends on a number of
parameters, one of which is the neutron scattering lengths of
737
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
Figure 3. Neutron contrast series data used to determine the NSLD of phytoglycogen. SANS data for two phytoglycogen concentrations: (a) 12.9%
w/w and (b) 22.4% w/w, for different % D2O. (c) The total scattering, as indicated by the total scattering invariant Q*, shows a clear minimum that
is independent of concentration. (d) Contrast matching used to predict the NSLD of phytoglycogen as a function of the H2O/D2O ratio.
different H 2 O/D2 O ratios, and at two phytoglycogen
concentrations (12.9% and 22.4% w/w). This type of
experiment is commonly known as a contrast series, and the
data are shown in Figure 3. From these data, three pieces of
information can be obtained. First, the match point in the
NSLD of phytoglycogen and the solvent occurs at 48.8% D2O.
Second, the match point is not significantly different for the two
concentrations, implying that there was no change in
nanoparticle density as the separation between nanoparticles
was significantly decreased (also supported by EISF analysis
shown in Figure S6). Finally, the data demonstrated a roughly
uniform density of the phytoglycogen nanoparticles, as implied
by the quadratic shape of the total scattering invariant Q*
versus %D2O (Figure 3c, calculated using eq 1). This is a result
of the dependence of the scattering intensity on the square of
the difference between the NSLD of the nanoparticle and the
solvent (Q* ∝ (Δρ)2). The quadratic shape also suggests that
there is a linear relationship between the NSLD of the
nanoparticle and the solvent (Figure 3d). This reaffirms our use
of the hard sphere model for the nanoparticle form factor, a
model that is consistent with the dense core structure
associated with dendrimeric particles.1−6
Determination of the NSLD match point of phytoglycogen
in the aqueous medium allows us to calculate the volume of the
glucose monomers within the phytoglycogen nanoparticles.
Dividing the sum of the atomic scattering lengths42 bi for an
average monomer dissolved in 48.8% D2O (C6H8.54D1.46O5), by
the match point NSLD (0.28 fm/Å3), results in a molecular
volume per monomer of 166.8 Å3. This in turn allows us to
calculate the water content of the fully hydrated phytoglycogen
nanoparticles. We do this by solving for the number of water
molecules associated with each glucose monomer, also called
the hydration number nH. This is the number of water
(core−shell, and a sphere with a continuous radial change in its
NSLD), which did not adequately fit the data. The use of the
uniform sphere model was also supported by the contrast series
experiments, which indicated nanoparticles with a uniform
NSLD. These measurements allowed the determination of the
radius (17.4 ± 1.6 nm) and NSLD (0.58 fm/Å3) of the
phytoglycogen nanoparticles in 100% D2O. The measured
nanoparticle radius is in good agreement with reported values
from AFM, electron microscopy, light scattering, and size
exclusion chromatography measurements.8,12−14
At higher phytoglycogen concentrations, SANS is sensitive to
interparticle correlations.40 These are described by the structure
factor S(q) that modifies the scattered intensity. According to
eq 2, dividing the high concentration data by the data for the
dilute (1% w/w) dispersion, which does not contain scattering
from interparticle interference, reveals the structure factor
(Figure 2b). One should note that this division has been
performed on concentration normalized data, shown in Figure
2a, rather than the raw I(q) curves obtained from the
experiment. This approach assumes that there is no change in
the nanoparticle size as a function of concentration, which
appears to be valid in this case based on the results of the
contrast series experiments described below (Figure 3). By
fitting the structure factor to a hard sphere model,41 the
interparticle spacing was obtained (Figure 2c). This analysis
reveals that the interparticle spacing approaches the average
diameter of the phytoglycogen nanoparticles near the maximum
concentration of 24.4% w/w. This result is consistent with the
onset of a jamming transition,30,31 which is observed as a
marked increase in the zero shear viscosity of phytoglycogen
dispersions at a concentration of ∼25% w/w (Figure S3).
Because neutrons scatter differently from hydrogen’s isotopes
(we now refer to 1H as hydrogen or ‘H’ and 2H as deuterium or
‘D’), we performed SANS measurements as a function of
738
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
interpretation is further supported by the calculated scattering
cross sections of bulk water, hydration water, and phytoglycogen (Table S1), which show that the cross sections of
hydration water and phytoglycogen are comparable to that of
bulk water.
QENS experiments also produce energy resolved spectra,
providing a more descriptive picture of the dynamics in the
30−300 ps time frame. We collected QENS spectra for aqueous
D2O dispersions of phytoglycogen at concentrations of 0%
(bulk D2O), 10.5% w/w, 19.7% w/w and 24.4% w/w, and
temperatures of 100, 280, and 295 K. To obtain an overall
picture of the dynamics, it is informative to sum the spectra
across the entire q-range and plot this as the dependence of the
imaginary part of the dynamic susceptibility as a function of
energy transfer. This is calculated as
molecules required to obtain the best-fit NSLD value at 100%
D2O (0.58 fm/Å3) using the equation
NSLD100%D2O =
bphytoglycogen + bD2OnH
vphytoglycogen + vD2OnH
(3)
where vD2O is taken to be 30.4 Å3.39 Additionally, we note that
the exchangeable protons within the glucose monomers must
be taken into account, which changes the value of bphytoglycogen
from 46.7 fm, at 48.8% D2O, to 62.7 fm, at 100% D2O. Solving
eq 3 for nH, we find that it would take 22.5 ± 2.5 water
molecules per glucose monomer to reach an NSLD of 0.58 fm/
Å3 in 100% D2O.
SANS data can also be used to estimate the molecular weight
of the phytoglycogen nanoparticles. We estimate the number of
glucose monomers per nanoparticle at N = 2.59 × 104 by
dividing the total nanoparticle volume of 2.21 × 10−23 m3 by
the average volume per monomer (including the associated
water) of 8.53 × 10−28 m3. Using nH and N, we can also
calculate the total number of water molecules per phytoglycogen nanoparticle as NH = 5.85 × 105. By multiplying the
number of monomers by the monomer molecular weight (162
g/mol, based on the formula C6H10O5 for phytoglycogen in
H2O), we obtain the molecular weight of a completely dry
phytoglycogen nanoparticle to be 4.16 × 106 g/mol. By
including the mass of 22.5 water molecules per glucose
monomer, we obtain a molecular weight of 14.7 × 106 g/mol
for phytoglycogen nanoparticles fully hydrated in H2O, with
each nanoparticle containing approximately 250% of its mass in
water. We note that there are a number of assumptions made in
this analysis such as assuming that the volume of hydration
water is unchanged from that of bulk water, that water
properties do not change substantially with deuteration, and
that all of the water inside the nanoparticles is hydration water.
We also note that the analysis of the QENS datawhich we
describe belowyields a similar value for the hydration
number nH.
Dynamics of Hydration Water. The high water content of
hydrated phytoglycogen suggests that the motions of both
hydration water and bulk water in aqueous dispersions of
phytoglycogen can be studied using QENS. In this technique,
gains and losses in the energy of scattered neutrons are
measured, which allows the direct determination of the motion
of atoms within a sample. Our measurements on the BASIS
instrument are sensitive to atomic motions that are faster than
∼300 ps, within the q-range of 0.2 to 2.0 Å−1 (∼3−30 Å in real
space). Data were collected for D2O and phytoglycogen
dispersions in D2O at concentrations of 10.5%, 19.7%, and
24.4% w/w, and temperatures from 100 to 300 K. We have
analyzed the QENS data in several different ways.
First, we evaluated the temperature dependence of the elastic
scattering intensity at a fixed energy window (Figure S4). This
is a convenient approach to look at overall differences in the
dynamics at each q-value occurring at a given time scale. The
data not only show a dramatic increase in dynamics
(corresponding to a significant decrease in the elastic
scattering) associated with the melting of bulk water, but also
show reductions in the melting temperature that increase with
increasing phytoglycogen concentration. We interpret this latter
result as being due to the hydration water population confined
within the phytoglycogen particles, which is consistent with the
suppression of the melting temperature of hydration water
populations associated with lipids and proteins.43−47 This
χ″(q , E) =
S(q , E )
= S(q , E)*[e E / kBT − 1]
nB(T , E)
(4)
where nB is the temperature-dependent Bose occupation
number. In Figure 4, the QENS spectra for the phytoglycogen
Figure 4. QENS data for phytoglycogen dispersions of different
concentrations plotted as the dynamic susceptibility χ″ at 295 K (for
which the data has been summed over all q) versus energy transfer, E.
BASIS instrument data for phytoglycogen are superimposed on spectra
of bulk D2O that were collected using the BASIS and CNCS
instruments. The excess scattering at low E reflects the dynamics of
hydration water and phytoglycogen nanoparticles.
dispersions demonstrate an excess in scattering when superimposed on the spectrum of D2O collected at the BASIS and
CNCS instruments. This excess scattering increases with
increasing phytoglycogen concentration and is particularly
evident at lower energy transfers. Previous analyses of oligoand polysaccharide solutions48,49 indicate that three relaxation
processes contribute to the spectra below 1 meV: bulk water
translation, hydration water translation, and the tail of a
relaxation process associated with the sugar. The enhanced
values of χ″ for phytoglycogen dispersions at low energy
transfers suggest that our spectra contain significant contributions from the translation of hydration water.
A more detailed analysis of the QENS data can be performed
as a function of q. For this analysis, the data are considered in
terms of the dynamic structure factor S(q,E). In this
representation the data can be fit using Lorentzian functions
Γ(q,E) to represent the inelastic scattering of the various
dynamical processes, while the elastic scattering term is
represented by a delta function δ(E).50 The elastic component
is scaled by the Elastic Incoherent Structure Factor, or
EISF(q,E), associated with each process, and is defined as the
ratio of elastic intensity to total scattered intensity. These terms
739
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
Figure 5. Fits to the QENS data. (a) A representative fit is shown for 24.4% w/w phytoglycogen in D2O at 280 K for q = 1.3 Å−1. The fit required
three Lorentzian functions, in addition to the instrumental resolution function, to account for the elastic scattering. The fwhm for each Lorentzian
component is shown for data collected at (b) 280 K and (c) 295 K. The q-range is truncated for the phytoglycogen process, due to it leaving the
dynamic window of the instrument (errors reported are 2σ). The data in panels b and c are shown for different phytoglycogen concentrations: 10.5%
(top), 19.7% (middle) and 24.4% (bottom). In each plot, data is shown for bulk water (black), hydration water (red), and phytoglycogen (blue). In
panels b and c, the dashed line corresponds to the best fit of the bulk water to a q2 dependence.
are then scaled by the factor Pi denoting the relative population
that participates in a given relaxation, i. From these terms we
can derive the theoretical scattering function:
We treated the data as consisting of three inelastic
components, each represented by Lorentzian functions. The
fwhm of the bulk water term was fixed to the value obtained
from the best fit to the D2O spectra collected at 280 and 295 K
to a single Lorentzian function. The fwhm of the hydration
water term was allowed to vary, along with the amplitudes of
both processes. The best fits to the water spectra reproduced
the data well in the q-ranges of 0.2−1.6 Å−1 at 280 K, and 0.2−
1.2 Å−1 at 295 K. Accordingly, we have restricted our analysis of
all other samples to these q-ranges. We point out that the use of
two components to treat the water dynamics is one of several
approaches used in the literature. The approach that we have
chosen is based on experimental48,49 and simulation51 results
that suggest two distinct water populations. However, others
have utilized a single feature to account for both bulk and
hydration water.52,53 In addition, we point out that the use of
Lorentzian functions may not be ideal, as hydration water
relaxations often exhibit a stretched spectral shape.43,52,54−56
The third Lorentzian feature is assigned to the tail of the
phytoglycogen relaxation, similar to that used for oligo- and
polysaccharide spectra obtained from depolarized light
scattering in the range of 1−100 GHz (0.004−0.4 meV).49
We do not present a detailed analysis of the phytoglycogen
process used in the fitting of the QENS data to avoid
overinterpretation of the data, since the majority of this process
resides outside of our experimental window. However, it is not
unreasonable to expect a small contribution of phytoglycogen
dynamics contributing to the low energy side of the
spectra.48,49,57
We observe a clear trend in the relative amplitude terms of
bulk and hydration water as a function of phytoglycogen
concentration in Figure 6. There is a decrease of aBulk Water and a
corresponding increase in aHydration Water as a function of
increasing phytoglycogen concentration. This result is expected,
as a larger fraction of the water population is needed to hydrate
an increasing number of nanoparticles. We can use these
amplitudes to provide a second, independent estimate of the
hydration water content in phytoglycogen. We solve for the
total number of water molecules per phytoglycogen nanoparticle, NH, of water using the relation:
n
STheo(q , E) =
∑ Pi(EISF(i q)δ(E) + [(1 − EISF(i q))·Γi(q , E)]
i=1
(5)
which can be fit against experimental data as follows:
Sexp(q , E) = DWF(q)*[STheo(q , E) ⊗ R(q , E)] + B(q , E)
(6)
R(q,E) is the instrument resolution function that is determined
from a low temperature (100 K) measurement of each sample,
B(q,E) is the instrumental background, and DWF(q) is the
Debye−Waller Factor (an example of the fit is shown in Figure
5a). Analysis of S(q,E) provides the full-width at half-maximum
(fwhm) (Figure 5b,c) and the normalized amplitude terms ai
(Figure 6a,b) of the Lorentzian functions associated with each
of the dynamic processes. Each fwhm term is related to the
relaxation time of the corresponding dynamic process, while the
amplitude terms describe the relative contributions of each
process.
Figure 6. Best fit values of the normalized area terms ai for each of the
Lorentzian functions used to fit the QENS data at (a) 280 K and (b)
295 K. No significant q-dependence is observed in the relative
contributions of the bulk and hydration water components. The
phytoglycogen process, however, shows increasing intensity as a
function of q (errors reported are 2σ).
NH =
740
fsol *aHydrationWater
aBulkWater + aHydrationWater
(7)
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
Figure 7. Translational motions of bulk water and hydration water in phytoglycogen dispersions show different q-dependences. (a) The exponents
correspond to the average values for measurements at both 280 and 295 K and all three concentrations. (b) Retardation factors, ξ, for hydration
water show a clear q-dependence. This dependence is a result of the comparison of the dynamics in the subdiffusive hydration water population, to
those in the diffusive bulk water population. This illustrates how the retardation factor can be influenced by the the intrinsic length-scale of a given
experimental technique.
be quantified as the retardation factor, ξ, defined as the ratio of
the bulk water fwhm to the hydration water fwhm. Here, we
report an average value of ξ = 5.8 ± 1.2 for phytoglycogen,
which was calculated across a q-range of 0.2−1.6 Å−1 at 280 K,
and 0.2−1.2 Å−1 at 295 K. This is comparable to the retardation
factor reported for glucose, the monomer subunit of
phytoglycogen, and other small saccharides.48,49,61 It is
remarkable that the increased molecular weight and structural
complexity of phytoglycogen, relative to that of glucose, do not
drastically modify the degree of perturbation of the hydration
water dynamics.49 This perturbation is consistent with the
effect of short-range interactions between the monomer and
water, where water−sugar and water−water hydrogen bonds
have comparable strengths,62 and the sugar molecules are able
to partially replace water in the tetrahedral hydrogen bond
network.58
It is important to note, however, that ξ has an implied
dependence on q. This is because the motions of hydration
water are subdiffusive, whereas the motions of bulk water are
diffusive (q2.6 vs q2 dependence). In Figure 7b, we demonstrate
this dependence by plotting ξ as a function of q. We observe a
clear trend in the apparent perturbation of hydration water with
respect to the length scale used to interrogate the system. At
small q, or long distance, hydration water dynamics are more
perturbed than at shorter distances (large q). This result shows
the power of neutron scattering measurements, which can
provide dynamical information at multiple length scales. This
observation is significant, as it demonstrates the length-scale
dependence of the retardation factor ξ, making it clear that
experimental techniques that probe ξ on different length scales
could result in very different values. As such, reported
retardation factors should specify the length scale of the
dynamic process probed in the experiment.
where fsol is the solute/water molar ratio.49,57 Taking the
molecular weights of dry phytoglycogen and D2O to be 4.21 ×
106 g/mol and 20 g/mol, respectively, we obtain NH = 6.7 ×
105 ± 1.2 × 105 water molecules per phytoglycogen
nanoparticle (average of all concentrations and q-values). We
obtain the hydration number nH per glucose monomer by
dividing NH by the average number of glucose monomers in the
phytoglycogen nanoparticles, 2.59 × 104. This yields a value of
nH = 25.8 ± 4.6 water molecules per glucose monomer, which
agrees well with the value of nH obtained from our analysis of
the SANS data (22.5 ± 2.5 water molecules per glucose
monomer). Taken together, these data indicate that phytoglycogen nanoparticles contain between 250% and 285% of their
mass in water.
The value of the hydration number is remarkable when
compared to the number of perturbed water molecules per
glucose molecule in solution, which is ∼14 as measured by
extended depolarized light scattering.48,49 As mentioned, the
effects of sugars on water are thought to be short-range
interactions, with the sugar molecules partially replacing water
in the tetrahedral hydrogen bond network,58 perturbing only
the first shell of water around the glucose.59 For phytoglycogen,
we observe between 22 and 26 water molecules per glucose
monomer, a significantly higher number than for glucose in
solution. This is better illustrated by comparing the number of
hydration waters per free −OH group. Oligo- and polysaccharides in solution typically associate ∼3 water molecules
per −OH group.48,49 On the other hand, phytoglycogen has
more than 7 water molecules perturbed per free −OH group.
We ascribe this increased number of perturbed or “bound”
water to the crowded environment between the closely packed
chains within the phytoglycogen nanoparticles.
The fwhm values are related to the inverse of the relaxation
time for a given process. The data in Figure 7a shows that bulk
water displays the expected q2 dependence for the fwhm
associated with translational diffusion. By contrast, the
hydration water process shows an average q-dependence of qB
with B = 2.6 ± 0.3 (Figure 7a). This indicates that the motions
of hydration water are subdiffusive, which is consistent with
other neutron scattering measurements of hydration water.43,60
Hydration water also translates more slowly than bulk water, as
can be seen by the smaller fwhm values. This perturbation can
■
CONCLUSIONS
In this study we have used neutron scattering to determine the
structure and hydration of phytoglycogen nanoparticles. SANS
measurements show that the nanoparticles are relatively
monodisperse with a radius of 17.4 ± 1.6 nm. Moreover, the
interparticle spacing suggests that the dispersions approach the
jamming regime at our highest measured concentration of
741
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
have supplied the monodisperse phytoglycogen nanoparticles
for the present study.
24.4% w/w. This is consistent with the divergence of the zero
shear viscosity at concentrations greater than 25% w/w.
Neutron contrast variation experiments showed that there is
no significant effect of crowding of the nanoparticles at high
phytoglycogen concentrations, finding the same value of NSLD
for concentrations of 12.9% and 22.4% w/w. These measurements also allowed us to calculate the monomer volume (166.8
Å3), and to estimate the hydration number (22.5 ± 2.5 water
molecules per glucose monomer). This hydration number value
is significantly larger than that measured for other oligo- and
polysaccharides, with fully hydrated phytoglycogen nanoparticles containing between 250% and 285% of their own
mass in water.
The large water content of the phytoglycogen nanoparticles
shows promise as a model system for investigations of
hydration water in hydrophilic environments. QENS measurements provided detailed information regarding the dynamics of
water in phytoglycogen dispersions as a function of
phytoglycogen concentration. These measurements allowed
us to obtain an independent estimate of the amount of water
associated with each nanoparticle (25.7 ± 4.6 water molecules
per glucose monomer), as well as the retardation of hydration
water motion (average value of 5.8 ± 1.2). The observed qdependence of the translational motion of bulk water showed
the expected q2 dependence, whereas the hydration water
process was subdiffusive with a q-dependence given by a power
law exponent of 2.6 ± 0.3. This result leads to an important
observation: the retardation factor ξ of the hydration water is qdependent, i.e., the measured value of ξ depends on the length
scale probed in the experiment. This analysis can be applied to
hydration water associated with other polysaccharides as well as
proteins and lipids, and will improve our understanding of the
dynamics of nanoconfined water.
Collectively, the results of the present study highlight the
intimate interactions of phytoglycogen nanoparticles with
water, resulting in many of its unique properties such as high
levels of water retention, low viscosity in water, and exceptional
stability of aqueous dispersions. These insights will enable
promising new technologies and therapies based on this
naturally occurring nanomaterial.
■
■
ACKNOWLEDGMENTS
The authors would like to thank Phil Whiting, Anton
Korenevski, and Michael Grossutti for discussions that have
greatly benefitted this work. Mirexus Biotechnologies, Inc.
generously supplied the monodisperse phytoglycogen nanoparticles. This work was funded by grants from the Ontario
Ministry of Agriculture (OMAF) and the Natural Sciences and
Engineering Research Council (NSERC) of Canada (J.R.D.).
J.R.D. is the recipient of a Senior Canada Research Chair in
Soft Matter and Biological Physics. J.D.N. is partially supported
by the U.S. DOE BES through EPSCoR Grant No. DE-FG0208ER46528. J.K. is supported through the Scientific User
Facilities Division of the DOE Office of Basic Energy Sciences
under US DOE Contract No. DE-AC05-00OR22725. Facilities
as the Spallation Neutron Source are managed by UT-Battelle,
LLC under US DOE Contract No. DE-AC05-00OR22725.
■
ASSOCIATED CONTENT
S Supporting Information
*
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.biomac.5b01393.
Additional characterization of the nanoparticles, along
with six supporting data figures and one supporting data
table (PDF)
■
REFERENCES
(1) Ballauff, M.; Likos, C. N. Angew. Chem., Int. Ed. 2004, 43, 2998−
3020.
(2) Lescanec, R. L.; Muthukumar, M. Macromolecules 1990, 23,
2280−2288.
(3) Mansfield, M. L.; Klushin, L. I. Macromolecules 1993, 26, 4262−
4268.
(4) Boris, D.; Rubinstein, M. Macromolecules 1996, 29, 7251−7260.
(5) Zook, T. C.; Pickett, G. T. Phys. Rev. Lett. 2003, 90, 015502−
015504.
(6) Nägele, G.; Kellerbauer, O.; Krause, R.; Klein, R. Phys. Rev. E:
Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 1993, 47, 2562−
2574.
(7) Meléndez, R.; Meléndez-Hevia, E.; Cascante, M. J. Mol. Evol.
1997, 45, 446−455.
(8) Huang, L.; Yao, Y. Carbohydr. Polym. 2011, 83, 1665−1671.
(9) Manners, D. J. Carbohydr. Polym. 1991, 16, 37−82.
(10) Meléndez, R.; Meléndez-Hevia, E.; Canela, E. I. Biophys. J. 1999,
77, 1327−1332.
(11) Meléndez, R.; Meléndez-Hevia, E.; Mas, F.; Mach, J.; Cascante,
M. Biophys. J. 1998, 75, 106−114.
(12) Powell, P. O.; Sullivan, M. A.; Sheehy, J. J.; Schulz, B. L.;
Warren, F. J.; Gilbert, R. G. PLoS One 2015, 10, e0121337.
(13) Wong, K.-S.; Kubo, A.; Jane, J.-l.; Harada, K.; Satoh, H.;
Nakamura, Y. J. Cereal Sci. 2003, 37, 139−149.
(14) Yun, S.-H.; Matheson, N. K. Carbohydr. Res. 1993, 243, 307−
321.
(15) Kajiura, H.; Takata, H.; Kuriki, T.; Kitamura, S. Carbohydr. Res.
2010, 345, 817−824.
(16) Putaux, J.-L.; Potocki-Véronese, G.; Remaud-Simeon, M.;
Buleon, A. Biomacromolecules 2006, 7, 1720−1728.
(17) Takata, H.; Kajiura, H.; Furuyashiki, T.; Kakutani, R.; Kuriki, T.
Carbohydr. Res. 2009, 344, 654−659.
(18) Perticaroli, S.; Ehlers, G.; Jalarvo, N.; Katsaras, J.; Nickels, J. D. J.
Phys. Chem. Lett. 2015, 6, 4018−4025.
(19) Mattea, C.; Qvist, J.; Halle, B. Biophys. J. 2008, 95, 2951−2963.
(20) Oleinikova, A.; Smolin, N.; Brovchenko, I.; Geiger, A.; Winter,
R. J. Phys. Chem. B 2005, 109, 1988−1998.
(21) Qvist, J.; Persson, E.; Mattea, C.; Halle, B. Faraday Discuss.
2009, 141, 131−144.
(22) Sasisanker, P.; Weingärtner, H. ChemPhysChem 2008, 9, 2802−
2808.
(23) Qvist, J.; Halle, B. J. Am. Chem. Soc. 2008, 130, 10345−10353.
(24) Li, T.; Hassanali, A. A.; Kao, Y.-T.; Zhong, D.; Singer, S. J. J. Am.
Chem. Soc. 2007, 129, 3376−3382.
(25) Pal, S. K.; Peon, J.; Bagchi, B.; Zewail, A. H. J. Phys. Chem. B
2002, 106, 12376−12395.
AUTHOR INFORMATION
Corresponding Author
*Address: Department of Physics, University of Guelph,
Guelph Ontario, Canada N1G 2W1. Tel: 519-824-4120, ext.
53950; Fax: 519-836-9967; E-mail: [email protected].
Author Contributions
∇
These authors contributed equally.
Notes
The authors declare the following competing financial
interest(s): We declare that two of the authors (J.R.D. and
E.P.-S.) are founders of Mirexus Biotechnologies, Inc., who
742
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743
Article
Biomacromolecules
(26) Pal, S. K.; Peon, J.; Zewail, A. H. Proc. Natl. Acad. Sci. U. S. A.
2002, 99, 1763−1768.
(27) Zhang, L.; Wang, L.; Kao, Y.-T.; Qiu, W.; Yang, Y.; Okobiah, O.;
Zhong, D. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 18461−18465.
(28) Zhang, L.; Yang, Y.; Kao, Y.-T.; Wang, L.; Zhong, D. J. Am.
Chem. Soc. 2009, 131, 10677−10691.
(29) Mashimo, S.; Kuwabara, S.; Yagihara, S.; Higasi, K. J. Phys.
Chem. 1987, 91, 6337−6338.
(30) Liu, A. J.; Nagel, S. R. Nature 1998, 396, 21−22.
(31) Trappe, V.; Prasad, V.; Cipelletti, L.; Segre, P.; Weitz, D. Nature
2001, 411, 772−775.
(32) Halle, B. Philos. Trans. R. Soc., B 2004, 359, 1207−1223.
(33) Zhao, J. K.; Gao, C. Y.; Liu, D. J. Appl. Crystallogr. 2010, 43,
1068−1077.
(34) Arnold, O.; Bilheux, J.-C.; Borreguero, J.; Buts, A.; Campbell, S.
I.; Chapon, L.; Doucet, M.; Draper, N.; Ferraz Leal, R.; Gigg, M.; et al.
Nucl. Instrum. Methods Phys. Res., Sect. A 2014, 764, 156−166.
(35) SasView for Small Angle Scattering Analysis; see: http://www.
sasview.org/.
(36) Mamontov, E.; Herwig, K. W. Rev. Sci. Instrum. 2011, 82,
085109.
(37) Ehlers, G.; Podlesnyak, A. A.; Niedziela, J. L.; Iverson, E. B.;
Sokol, P. E. Rev. Sci. Instrum. 2011, 82, 085108.
(38) Azuah, R. T.; Kneller, L. R.; Qiu, Y.; Tregenna-Piggott, P. L.;
Brown, C. M.; Copley, J. R.; Dimeo, R. M. J. Res. Natl. Inst. Stand.
Technol. 2009, 114, 341−358.
(39) Petrache, H. I.; Feller, S. E.; Nagle, J. F. Biophys. J. 1997, 72,
2237−2242.
(40) Pedersen, J. S. J. Appl. Crystallogr. 1994, 27, 595−608.
(41) Percus, J. K.; Yevick, G. J. Phys. Rev. 1958, 110, 1−13.
(42) Sears, V. F. Neutron News 1992, 3, 26−37.
(43) Nickels, J. D.; O’Neill, H.; Hong, L.; Tyagi, M.; Ehlers, G.;
Weiss, K. L.; Zhang, Q.; Yi, Z.; Mamontov, E.; Smith, J. C.; Sokolov, A.
P. Biophys. J. 2012, 103, 1566−1575.
(44) Toppozini, L.; Armstrong, C. L.; Kaye, M. D.; Tyagi, M.;
Jenkins, T.; Rheinstädter, M. C. ISRN Biophys. 2012, 2012, 1−7.
(45) Weik, M.; Lehnert, U.; Zaccai, G. Biophys. J. 2005, 89, 3639−
3646.
(46) Wood, K.; Plazanet, M.; Gabel, F.; Kessler, B.; Oesterhelt, D.;
Tobias, D.; Zaccai, G.; Weik, M. Proc. Natl. Acad. Sci. U. S. A. 2007,
104, 18049−18054.
(47) Nickels, J. D.; Katsaras, J. In Membrane Hydration; Springer
International Publishing: Cham, Switzerland, 2015; pp 45−67.
(48) Lupi, L.; Comez, L.; Paolantoni, M.; Perticaroli, S.; Sassi, P.;
Morresi, A.; Ladanyi, B. M.; Fioretto, D. J. Phys. Chem. B 2012, 116,
14760−14676.
(49) Perticaroli, S.; Nakanishi, M.; Pashkovski, E.; Sokolov, A. P. J.
Phys. Chem. B 2013, 117, 7729−7736.
(50) Vineyard, G. H. Phys. Rev. 1958, 110, 999−1010.
(51) Chiavazzo, E.; Fasano, M.; Asinari, P.; Decuzzi, P. Nat. Commun.
2014, 5, 1−11.
(52) Toppozini, L.; Roosen-Runge, F.; Bewley, R. I.; Dalgliesh, R. M.;
Perring, T.; Seydel, T.; Glyde, H. R.; García Sakai, V. G.; Rheinstädter,
M. C. Soft Matter 2015, 11, 8354−8371.
(53) Ben Ishai, P.; Mamontov, E.; Nickels, J. D.; Sokolov, A. P. J.
Phys. Chem. B 2013, 117, 7724−7728.
(54) Santucci, S.; Fioretto, D.; Comez, L.; Gessini, A.; Masciovecchio,
C. Phys. Rev. Lett. 2006, 97, 225701−225704.
(55) Torre, R.; Bartolini, P.; Righini, R. Nature 2004, 428, 296−299.
(56) Zanotti, J. M.; Bellissent-Funel, M. C.; Chen, S. H. Phys. Rev. E:
Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 1999, 59, 3084−
3093.
(57) Perticaroli, S.; Comez, L.; Paolantoni, M.; Sassi, P.; Lupi, L.;
Fioretto, D.; Paciaroni, A.; Morresi, A. J. Phys. Chem. B 2010, 114,
8262−8269.
(58) Suzuki, T.; Sota, T. J. Chem. Phys. 2003, 119, 10133−10137.
(59) Denisov, V. P.; Halle, B. Faraday Discuss. 1996, 103, 227−244.
(60) Bellissent-Funel, M.-C.; Teixeira, J.; Bradley, K.; Chen, S. J. Phys.
I 1992, 2, 995−1001.
(61) Paolantoni, M.; Comez, L.; Gallina, M.; Sassi, P.; Scarponi, F.;
Fioretto, D.; Morresi, A. J. Phys. Chem. B 2009, 113, 7874−7878.
(62) Paolantoni, M.; Sassi, P.; Morresi, A.; Santini, S. J. Chem. Phys.
2007, 127, 024504−024509.
743
DOI: 10.1021/acs.biomac.5b01393
Biomacromolecules 2016, 17, 735−743