Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
Insights into the role of protein molecule
size and structure on interfacial properties
using designed sequences
rsif.royalsocietypublishing.org
Research
Cite this article: Dwyer MD, He L, James M,
Nelson A, Middelberg APJ. 2013 Insights into
the role of protein molecule size and structure
on interfacial properties using designed
sequences. J R Soc Interface 10: 20120987.
http://dx.doi.org/10.1098/rsif.2012.0987
Received: 30 November 2012
Accepted: 19 December 2012
Subject Areas:
bioengineering, biophysics,
chemical engineering
Keywords:
protein, peptide, interface, adsorption,
rheology, neutron reflectometry
Author for correspondence:
Lizhong He
e-mail: [email protected]
Electronic supplementary material is available
at http://dx.doi.org/10.1098/rsif.2012.0987 or
via http://rsif.royalsocietypublishing.org.
Mirjana Dimitrijev Dwyer1, Lizhong He1,2, Michael James3, Andrew Nelson3
and Anton P. J. Middelberg1
1
Centre for Biomolecular Engineering, Australian Institute for Bioengineering and Nanotechnology and
School of Chemical Engineering, The University of Queensland, St Lucia, Queensland 4072, Australia
2
Department of Chemical Engineering, Monash University, Clayton, Victoria 3800, Australia
3
Bragg Institute, Australian Nuclear Science and Technology Organisation, Building 87, Locked Bag 2001,
Kirrawee DC, New South Wales 2232, Australia
Mixtures of a large, structured protein with a smaller, unstructured component are inherently complex and hard to characterize at interfaces,
leading to difficulties in understanding their interfacial behaviours and,
therefore, formulation optimization. Here, we investigated interfacial properties of such a mixed system. Simplicity was achieved using designed
sequences in which chemical differences had been eliminated to isolate the
effect of molecular size and structure, namely a short unstructured peptide
(DAMP1) and its longer structured protein concatamer (DAMP4). Interfacial
tension measurements suggested that the size and bulk structuring of the
larger molecule led to much slower adsorption kinetics. Neutron reflectometry at equilibrium revealed that both molecules adsorbed as a monolayer to
the air –water interface (indicating unfolding of DAMP4 to give a chain of
four connected DAMP1 molecules), with a concentration ratio equal to
that in the bulk. This suggests the overall free energy of adsorption is
equal despite differences in size and bulk structure. At small interfacial
extensional strains, only molecule packing influenced the stress response.
At larger strains, the effect of size became apparent, with DAMP4 registering
a higher stress response and interfacial elasticity. When both components were
present at the interface, most stress-dissipating movement was achieved by
DAMP1. This work thus provides insights into the role of proteins’ molecular
size and structure on their interfacial properties, and the designed sequences
introduced here can serve as effective tools for interfacial studies of proteins
and polymers.
1. Introduction
The behaviour of proteins at air–water interfaces determines their behaviour as
foam stabilizers. In order to improve product formulations, it would be desirable
to be able to predict functional behaviour from an understanding of interfacial
properties, and ideally from knowledge of molecular structure. This link is not
well established [1–3], largely owing to the complexity of protein molecules
and the interplay between many contributing factors including surface coverage/pressure, adsorption rate, electrostatic and steric repulsion, and interfacial
rheology. For example, increasing the ionic strength of solution from which
protein foam is formed may have the effect of increasing the measured interfacial
elasticity owing to increased intermolecular interaction, however changes in surface packing, interfacial charge and bulk aggregation may, and often do, occur.
Discrepancies are, therefore, often observed in the literature as to whether a
given property improves foam functionality; for example, in the case of interfacial
rheology, some studies observe a positive correlation with foam functionality
[4,5], whereas others show negative or no correlation [6–8].
In nature as well as in practical formulations, proteins rarely exist as single
components [9], further obscuring understanding. The addition of a smaller,
& 2013 The Author(s) Published by the Royal Society. All rights reserved.
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
2. Material and methods
2.1. Materials
DAMP4 (11 116.5 Da) was produced biologically and purified as
described previously [15]. DAMP1 (2731 Da) was synthesized by
GenScript Corporation (Piscataway, NJ), and quantified by high
sensitivity amino acid analysis (Australian Proteome Analysis
Facility, Sydney). Purity of both DAMP4 and DAMP1 was
greater than 95 per cent by HPLC. A biologically-deuterated
form of DAMP4 (d-DAMP4) was produced by the National
Deuteration Facility (ANSTO, Lucas Heights), and purified
using the same procedure as for non-deuterated DAMP4. The
extent of deuteration of d-DAMP4 was calculated by mass spectroscopy to be 91.1 per cent on a non-exchangeable hydrogen
basis (see the electronic supplementary material, figure S2).
The buffer condition for all samples was 25 mM sodium 4-(2hydroxyethyl)-1-piperazine ethanesulphonate (HEPES), 200 mM
ethylenediamine tetraacetic acid (EDTA), at pH 7.4. The concentrations tested were 8 mM DAMP4, DAMP1, or a mix of the two
at 25, 50, 75 and 90 per cent DAMP4 on a mass basis, 8 mM total
concentration (table 1). Interfacial tension measurements were
also performed at varying concentrations of the individual components, as described in the text. Milli-Q water was used for all
solutions (Millipore, North Ryde, Australia).
2.2. Interfacial tension
A DSA-10 drop-shape analysis unit (Krüss GmbH, Hamburg,
Germany) was used to measure interfacial tension kinetics. The
sample (8 ml) was held in a quartz cuvette (Hellma GmbH,
Mülheim, Germany). Bubbles were formed through a U-shaped
stainless steel capillary of known diameter fed by a glass syringe
operated manually. Cleanliness and operation of the system
2
J R Soc Interface 10: 20120987
reflectometry [18]). It is expected that DAMP1 will behave
similarly. Like AM1 [19], DAMP1 is largely unstructured in
bulk (see the electronic supplementary material, figure S1),
whereas DAMP4 forms a very stable four-helix bundle [15].
The effect of molecule size and structuring on interfacial
adsorption is probed using DAMP1 and DAMP4 at different
ratios of both components. The use of known components
overcomes aforementioned issues regarding the difficulty
of characterizing mixed systems. The kinetics of interfacial
tension is probed by drop-shape analysis, whereas the
equilibrium adsorption and interfacial structure are studied
by neutron reflectometry. Specifically, the relationship
between the bulk and interfacial concentrations of DAMP1
and DAMP4 is determined using neutron reflectometry,
with deuterated DAMP4 used to contrast it from DAMP1.
This approach overcomes the issues discussed above regarding unknown interfacial composition. In the final section of
this work, the response of the DAMP4 and DAMP1 mixed
system to an applied compression and extensional strain is
studied. As the interfacial composition and structure are
known, interesting insights into the effect of molecule size
on interfacial mobility and stress response are gained.
The approach of using a simple, defined peptide and its
concatamer removes the complexity and factor coupling
which usually complicates studies of protein mixed systems.
As DAMP1 and DAMP4 have the same sequence and, therefore,
differ only in their size and bulk structuring, insights are gained
into the role of molecule size and structuring on interfacial
adsorption and rheology that would not be easily possible
with naturally occurring protein systems and their hydrolysates.
rsif.royalsocietypublishing.org
unstructured component (such as a peptide) to a structured
protein has been seen to improve foam properties, as is
often observed in the case of protein hydrolysates [10,11]. It
is generally accepted that this functional improvement is
related to the smaller size (better adsorption kinetics) and
increased hydrophobicity (greater exposure of hydrophobic
residues usually buried in the structured protein) of the peptide component [10,12,13]. At a fundamental level, however,
the effect of molecule size and structure on interfacial properties is even less clear than in the case of single-component
systems. Improved understanding is important to facilitate
the optimized use of naturally occurring proteins, which is
currently empirical, as well as the design of synthetic
sequences that bring new or enhanced function.
Two issues hinder progress in this area. First, systems
composed of natural components are often complex and
difficult to characterize. For example, heat treatment of
b-lactoglobulin at pH 2 to form fibres results in a mixed
system of fibres, monomers, peptides and deamidated variants of each [14]. The exact extent of cleavage to peptides,
formation of fibres and deamidation is very difficult to identify and quantify, therefore the bulk composition of the system
being studied is at best only qualitatively understood. This type
of issue makes even the simplest of experiments difficult to
interpret, as in the study by Jung et al. [14], where the observed
interfacial tension kinetics could not be clearly explained.
Second, the interfacial composition is even more difficult to
characterize than that in the bulk. This is not only due to the
same issues as discussed for bulk characterization, but also
due to the low accessibility and high complexity of experimental techniques that characterize surface composition (such as
neutron reflectometry). When the concentrations and structures of components at an interface are unknown, observed
properties, such as interfacial tension, viscoelasticity, foam
and thin film stability, cannot be directly interpreted. A meaningful study of the roles of molecular size and structure on
interfacial properties requires a well-defined model system to
allow decoupling of these factors from differences in chemistry.
In this work, we present such a model system, which has
known bulk concentrations, known amino acid sequence and
structuring in bulk. We study a larger, structured protein
(DAMP4), a smaller, unstructured peptide (DAMP1) and
mixtures of the two. DAMP4 is a concatameric repeat of
the DAMP1 sequence, therefore this system represents the
simplest form of a mixed protein –peptide system, where
only size and bulk structuring varies between components,
and differences in chemistry have been eliminated. DAMP4
[15], MD(PSMKQLADS-LHQLARQ-VSRLEHAD)4, is four
repeats of the DAMP1 sequence, PSMKQLADS-LHQLARQVSRLEHAD, with two additional residues at the start, a
methionine (required for bioexpression purposes), and an
aspartate (for aspartyl–proline bond cleavage studies [16]).
Besides these N-terminal residues, DAMP1 and DAMP4 are
chemically identical. While the presence of the additional
aspartate residue (with its negative charge) has the effect of
shifting the theoretical pI slightly from 7.3 for DAMP1 to 6.7
for DAMP4, our earlier study of the foaming and film-forming
properties of DAMP1 and DAMP4 [16] indicated that their
behaviours were not significantly influenced by the presence
of these additional residues. DAMP1 and DAMP4 are based
on the surface-active peptide AM1 [17], which was designed
to adsorb as a single a-helix at the air–water or oil–water
interface (monolayer structure was confirmed by neutron
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
Table 1. Experimental concentrations of DAMP1 and DAMP4.
3
molar basis
(DAMP4),
mg ml21
(DAMP1),
mg ml21
total
mg ml21
%
DAMP4
100
89
—
89
100
8
90
75
62
36
6.6
13
69
49
67
42
5.6
3.2
2.4
4.8
8
8
50
18
18
35
19.5
1.6
6.4
8
25
0
6.7
—
20
22
27
22
7.5
0
0.6
—
7.4
8
8
8
was checked by forming an air bubble in milli-Q water, and confirming a constant interfacial tension of 72.8 mN m21 for 10 min.
To measure the interfacial tension kinetics, the cuvette was filled
with the sample of interest, a bubble of about 10 ml formed, and
interfacial tension as extracted by the software (via images of the
bubble collected by a connected camera) was monitored at a rate
of about 1 measurement per second.
2.3. Calculation of experimental and theoretical
characteristic diffusion times
The experimental diffusion times were estimated from interfacial
tension data, as the approximate time it takes to reach around 90
per cent of the final interfacial tension [20]. Theoretical characteristic
times of diffusion for DAMP1 and DAMP4 were calculated by the
method previously used for similar peptides Lac28 and Lac21 [21].
First, the Polson equation [22], D ¼ 2.85 1025 M21/3 cm2 s21
(where M is molecular mass) was used to calculate the molecular diffusion coefficients, D, in bulk, giving DDAMP4 ¼ 1.28 10210 m2 s21 and DDAMP1 ¼ 2.04 10210 m2 s21. Characteristic
diffusion time constants, td, were then calculated assuming a
linear isotherm [20]: td ¼ 1/D(Gmax/C1)2, where C1 is the bulk
concentration and Gmax is the maximum surface coverage, calculated based on the method for Lac21 and Lac28 [21], where 35 Å2
is the area assumed per exposed hydrophobic residue, and two
hydrophobic residues per heptad are assumed for an a-helix.
This gives a Gmax equal to 1.5 1027 mol m22 for DAMP4 and
6.8 1027 mol m22 for DAMP1.
2.4. Neutron reflectometry
Information on the interfacial thickness and coverage of
DAMP1/DAMP4 systems was collected using the Platypus
time of flight neutron reflectometer (ANSTO, Lucas Heights;
[23,24]), which is fed by a cold neutron beam (2.8 Å l 18.0 Å) from the OPAL 20 MW research reactor. The neutron
reflectivity (the ratio of reflected and incident intensities)
was measured as a function of momentum transfer, Q, where
Q ¼ 4p sinu/l, where l is the neutron wavelength and u is
the angle of incidence of the collimated neutron beam onto the
air – liquid surface. A Q-range of 0.013– 0.33 Å21 with a constant
Q resolution of 8 per cent was used.
To distinguish DAMP4 from DAMP1, deuterated DAMP4
was used as a second contrast condition. Each sample condition
(besides 8 mM DAMP1) was, therefore, measured at two contrast
conditions, with deuterated DAMP4 (d-DAMP4) and with
non-deuterated DAMP4 (h-DAMP4), totalling 11 experiments:
four mix conditions (2 contrasts) þ 1 100% DAMP4 condition
(2 contrasts) þ 1 100% DAMP1 condition (1 contrast). All
samples were prepared in the same buffer condition as for interfacial tension tests (25 mM HEPES, 200 mM EDTA, pH 7.4), but
(DAMP4),
mM
(DAMP1),
mM
total
mM
—
8
in null-reflecting water (8.1% w/w D2O) instead of Milli-Q to
eliminate reflection from the aqueous phase. Thirty-five millilitres of sample was held in a Teflon sample trough, within a
sealed chamber with single-crystal quartz windows to allow passage of the neutron beam. At least 2 h aging was allowed prior to
first measurement of each sample; however, owing to sample
sequencing 3–7 h aging was often achieved.
The MOTOFIT analysis program was used to fit the data [25].
This program describes the interface of interest with an arbitrary
number of layers. Each of these layers is described by four parameters, a thickness, scattering length density (SLD), solvent
penetration and roughness. For these systems, a single-layer
model proved suitable for all datasets. Each dataset was fitted
simultaneously with its corresponding contrast variant (e.g.
the 50% h-DAMP4 data were fitted simultaneously with the
50% d-DAMP4 data). The layer roughness, thickness and solvent
penetration values were linked for the h- and d- contrast conditions, and allowed to vary. The scattering lengths of each
component were calculated based on the molecular formulae of
h-DAMP4, DAMP1 and d-DAMP4 (knowing the extent of deuteration on a non-exchangeable hydrogen basis is 91.1%, see
the electronic supplementary material, figure S2), and using
tabulated neutron scattering lengths for the atoms C, H, D, O,
N and S [26]. The effect of 8.1% w/w D2O in the solution on scattering length was accounted for assuming 8.1 per cent
replacement of exchangeable H atoms by D. Molecular volumes
were estimated at 13919.2 Å3 for DAMP4 (both h- and d-) and
3408.7 Å3 for DAMP1 using the sum of individual amino acid
volumes reported elsewhere [27], as previously done for the
related sequence AM1 [28]. Dividing the scattering lengths
by molecular volume allowed the SLDs to be estimated:
SLDh-DAMP4 ¼ 1.95 1026 Å22, SLDd-DAMP4 ¼ 5.96 1026 Å22,
SLDDAMP1 ¼ 1.95 1026 Å22. As DAMP1 and h-DAMP4 have
equal SLDs, the SLD for all h-conditions was fixed at 1.95 1026 Å22. The SLD for the 100 per cent d-DAMP4 data was
fixed at 5.96 1026 Å22. In the case of mixed d-DAMP4 and
h-DAMP1 layers, the SLD parameter depends on the relative
proportions of d-DAMP4 and h-DAMP1, so was allowed to vary.
2.5. Langmuir-trough compression tests
A commercially available Langmuir trough (9.2 5 cm, Nima
Technology, Coventry, UK) was used to study interfacial
pressure during a compression and expansion cycle. Prepared
samples (40 ml of 8 mM DAMP4, 8 mM DAMP1, or 1.6 mM
DAMP4 þ 6.4 mM DAMP1 in 25 mM HEPES, 200 mM EDTA,
pH 7.4) were poured into the trough, covered and allowed to
age for 3 h. After aging, a compression – expansion cycle was performed by linear movement of the Langmuir barriers (at the
maximum speed of 114 cm2 min21), from the maximum area of
46 cm2 to the minimum of 11 cm2 and back to 46 cm2. Interfacial
J R Soc Interface 10: 20120987
%
DAMP4
rsif.royalsocietypublishing.org
mass basis
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
interfacial tension (mN m–1)
4
2.4 mM DAMP1
4.8 mM DAMP1
6.4 mM DAMP1
7.4 mM DAMP1
8.0 mM DAMP1
70
65
60
55
50
0.6 mM DAMP4
1.6 mM DAMP4
3.2 mM DAMP4
5.6 mM DAMP4
8.0 mM DAMP4
(b) 75
70
65
60
55
3. Results and discussion
(c) 75
3.1. Interfacial tension behaviour
interfacial tension (mN m–1)
50
The air–water interfacial tension (IFT) kinetics to 5 min (300 s)
for DAMP1 solutions at concentrations 2.4, 4.8, 6.4, 7.4 and
8 mM are shown in figure 1a. At 2.4 mM DAMP1, a significant
lag time of about 1 min was observed prior to a rapid decrease
in IFT to a final value of 53.8 + 0.2 mN m21. At all other
concentrations, the lag time was not so significant and
rapid IFT decrease was complete in less than a minute. The
plateau IFT values reached were similar, between 52 and
53 mN m21, and the effect of concentration on rate of IFT
decrease was not substantial over the observed timescale.
The interfacial tension kinetics of DAMP1 resemble closely
those of AM1 under similar conditions [19], with most of the
change occurring in the first 50 s, and reaching statistically
identical final values (51.6 + 0.4 mN m21 for AM1 [19] and
52.0 + 0.4 mN m21 for DAMP1).
Figure 1b shows the interfacial tension kinetics of DAMP4
at concentrations of 0.6, 1.6, 3.2, 5.6 and 8 mM up to a time
of 5 min. It can immediately be seen that DAMP4 adsorption
is much slower than that of DAMP1, and the effect of concentration more pronounced. As concentration increases, the rate
of interfacial tension decrease (adsorption) is also increased.
Unlike DAMP1, which at all concentrations tested completes
interfacial adsorption within 5 min, DAMP4 does not reach
equilibrium within this time frame for the concentrations
tested. Compared with 8mM DAMP1, which takes about
1 min to reach its final value, 8 mM DAMP4 takes more
than 5 min. This concentration is on a per-molecule basis
and would be four times greater on a per-monomer basis.
(i.e. the mass in 8 mM of DAMP4 is four times that in 8 mM
DAMP1, table 1). To shed light on the rates of interfacial
tension decrease, the approach previously described for similar peptides Lac21 and Lac28 [21], was used to calculate
theoretical diffusion time constants, td, for DAMP4 and
DAMP1, and compared with those experimentally observed
(table 2). Adsorption of DAMP1 appears to be diffusion
70
25% DAMP4 (0.6 mM DAMP4 7.4 mM DAMP1)
50% DAMP4 (1.6 mM DAMP4 6.4 mM DAMP1)
75% DAMP4 (3.2 mM DAMP4 4.8 mM DAMP1)
99% DAMP4 (5.6 mM DAMP4 2.4 mM DAMP1)
65
60
55
50
0
50
100
150
200
250
300
time (s)
Figure 1. Interfacial tension versus time for (a) DAMP1, (b) DAMP4, (c) mixtures of DAMP4 and DAMP1 in 25 mM HEPES (200 mM EDTA, pH 7.4).
controlled, as the theoretical td values are close to those
observed experimentally. Adsorption of DAMP4 is much
slower than predicted by the diffusion time constants, indicating a significant energy barrier to adsorption exists for
DAMP4 that is not present for DAMP1. DAMP4 folds into
a very stable four-helix bundle in bulk [15], therefore the
slow adsorption rates observed are likely to be due to the significant energy barrier associated with the unfolding of this
four-helix bundle in order for adsorption to occur, as
observed previously for Lac28 compared with Lac21 [21].
The interfacial tension kinetics at the air –water interface
for a DAMP1 and DAMP4 mixed system are shown in
figure 1c. The mix conditions chosen corresponded to 90,
75, 50 and 25 per cent DAMP4 on a mass basis, as shown
in table 1. The total molar concentration was kept constant
at 8 mM under all conditions. The interfacial tension kinetics
under all mix conditions are similar, dropping rapidly within
the first 30 s to final values of 52– 53 mN m21, which is much
faster than the kinetics of the individual components at each
mix condition. This is seen most clearly by comparing the
kinetics of the 75 per cent DAMP4 condition (3.2 mM
J R Soc Interface 10: 20120987
The extensional rheology of adsorbed films at the air – water
interface was measured using a custom-made apparatus, the
Cambridge interfacial tensiometer, which has been previously
described [29]. Samples of interest (6.5 ml) were pipetted into a
PTFE trough. Located at the sample air –water interface were
two optical fibre T-pieces, one connected to a piezoelectric
motor which imposed a strain on the interface, and the other
which measured the stress developed in response to this strain
via connection to a force transducer. Upon filling of the trough,
the moving T-piece was programmed to automatically run
once per minute from 0 to 1 per cent linear strain and then
back to 0 per cent (at a speed of 10% strain per second). The
stress response was monitored, and from this time course an indication of the time taken for mechanical equilibration of the
interfacial system was obtained, without destruction of the structure. After 3 h aging and monitoring 1 per cent strain response, a
large-strain test to at least 60 per cent strain was performed. The
true data frequency is around 100 points per second; however,
this is greatly reduced during analysis to enhance graphing clarity.
interfacial tension (mN m–1)
2.6. Interfacial extensional rheology
(a) 75
rsif.royalsocietypublishing.org
pressure during the cycle was monitored with a filter-paper
Wilhelmy plate attached to a microbalance, which had been
pre-zeroed on the surface of a 40 ml sample of milli-Q water.
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
Table 2. Theoretical and approximate experimental diffusion time constants (td) for DAMP4 and DAMP1a.
concentration,
mM DAMP4
concentration,
mM monomer
equivalent
theoretical
td (s)
approx.
experimental
td (s)
concentration,
mM DAMP1
theoretical
td (s)
approx.
experimental
td (s)
0.6
2.4
488.3
.1000
2.4
391.2
150
1.6
3.2
6.4
12.8
68.7
17.2
.800
.500
4.8
6.4
97.8
55.0
50
40
5.6
8
22.4
32
5.6
2.8
.300
.300
7.4
8
41.1
35.2
30
25
Experimental diffusion time constants were estimated from interfacial tension data shown in figure 1.
DAMP4 þ 4.8 mM DAMP1) with that of the individual components (3.2 mM DAMP4 and 4.8 mM DAMP1), which are
much slower than when combined. The major reason for
the increased adsorption kinetics is the change of total bulk
mass concentration of the mixed system. Owing to the
larger size of DAMP4 in comparison to DAMP1, an increase
in the molar proportion of DAMP4 in the systems in figure 1c
results in an increased bulk mass concentration (table 1),
which is larger than DAMP1 alone. This increased mass concentration is expected to increase kinetics of protein
adsorption at interfaces [30]. Indeed, control experiments
show that the DAMP1-only system and the mixture of
DAMP1 and DAMP4 have similar dynamic interface tensions
when the total concentration is equivalent or higher than
12.8 mM DAMP1 (see the electronic supplementary material,
figure S4). While interfacial tension measurements give
important information on adsorption kinetics, it should
be noted that measurements of interfacial tension cannot be
directly interpreted as indications of interfacial adsorption
because of nonlinear relationship between interfacial tension
and interfacial coverage, as evidenced by direct measurements of surface coverage such as neutron reflectometry
[31], and measurement of surface excess using radioactive
surfactants [32]. Neutron reflection experiments have thus
been carried out to determine composition of the interfacial
layer as discussed in §3.2.
3.2. Composition of the interfacial layer (neutron
reflectometry with deuterated DAMP4)
To identify the components adsorbed at the interface at
equilibrium, neutron reflectometry was used. This technique
exploits the fact that the neutron reflectivity of an air–water
interface is modulated by adsorbed molecules, and with
enough prior knowledge of the system, can provide information
such as adsorbed amount and layer thickness.
Figure 2a shows the neutron reflectivity profiles of the
same systems as in figure 1c, in the same buffer conditions
(25 mM HEPES, 200 mM EDTA, pH 7.4), but in null-reflecting
water (8.1% w/w D2O) instead of Milli-Q in order to eliminate reflection from the solution subphase. Non-deuterated
DAMP4 (h-DAMP4) was used. As DAMP4 is a repeat of
the DAMP1 sequence, DAMP1 and h-DAMP4 have the
same SLDs (1.95 1026 Å22), and are, therefore, indistinguishable with neutrons. The neutron reflectivity profiles in
figure 2a contain information about the overall interfacial
film structure, and not the individual components. As the
reflectivity curves are almost identical, qualitative observation of figure 2a indicates that the overall structure of the
interfacial film formed does not vary with changes in
DAMP1 : DAMP4. Given that figure 2a shows equivalent
neutron reflectivity profiles at equilibrium for 100 per cent
DAMP1, 100 per cent DAMP4 and for mixtures of these
two molecules, these results indicate that the surface excess
for DAMP1 and DAMP4 are essentially the same despite
the fact that the bulk mass of DAMP1 added to the subphase
was 1/4 that of DAMP4. No Kiessig fringes were observed in
these reflection data owing to the very thin interfacial layers,
and the relatively high background resulting from incoherent
neutron scattering from the largely protonated aqueous
subphase (approx. 92% H2O).
In order to determine the composition of the interfacial
layer, the ‘visibility’ with respect to neutrons of one of the components must differ compared with the other. The conditions in
figure 2a were, therefore, repeated in figure 2b using deuterated
DAMP4 (d-DAMP4), which has a much higher SLD than
DAMP1 (5.96 1026 compared with 1.95 1026 Å22), and,
therefore, provides contrast variation between the two molecules. Immediately it can be seen that the total reflectivity is
greater than that in figure 2a, which indicates that d-DAMP4
is present in the equilibrium interfacial film. Reflectivity
increases with increasing bulk d-DAMP4 concentration,
which indicates that the proportion of DAMP4 at the interface
increases as its bulk concentration is increased. This is a result
that could not be determined by interfacial tension profiles
(figure 1) which suggest that, on the timescale of minutes,
DAMP1 dominates the interface. Here, from neutron reflectometry data, it is evident that both DAMP4 and DAMP1
are co-populating the interface once equilibrium has been
established (on the timescale of hours).
A single-layer model fits the neutron profiles very well
(black solid lines in figure 2a,b). Previous neutron reflectivity studies showed that a related peptide, AM1, adsorbs
as a single monolayer at the air –water interface, about
one a-helix width thick (approx. 15 Å) [18]. AM1, and subsequently DAMP4 and DAMP1, are designed with
hydrophobic and hydrophilic residues placed strategically
to drive helical interfacial structuring [15,17]. DAMP4 and
DAMP1 are, therefore, expected to form a similar monolayer
structure. For thin layers, the layer thickness and overall SLD
of the layer (a volume fraction weighted average of the solvent SLD and protein SLD) can be inversely correlated in
J R Soc Interface 10: 20120987
a
DAMP1
rsif.royalsocietypublishing.org
DAMP4
5
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
(a)
6
reflectivity
1 × 10–4
interfacial DAMP1
1 × 10–5
bulk DAMP1
1 × 10–6
100% d-DAMP4
90% d-DAMP4
75% d-DAMP4
50% d-DAMP4
25% d-DAMP4
0% d-DAMP4
1 × 10–3
1 × 10–4
1 × 10–5
1 × 10–6
0
0.05
0.10
0.15 0.20
Q (Å–1)
0.25
0.30
0.35
Figure 2. Neutron reflectometry profiles of the same systems as in figure 1,
with (a) non-deuterated DAMP4 (h-DAMP4), and (b) with deuterated DAMP4
(d-DAMP4). The buffer conditions (25 mM HEPES, 200 mM EDTA, pH 7.4) are
the same as figure 1, but in null-reflecting water (8.1% w/w D2O). DAMP1
and DAMP4 have almost identical neutron SLDs, so the overlap of profiles in
(a) shows that there is no difference in the interfacial structure (thickness and
surface coverage) for the conditions tested. With increased d-DAMP4 in bulk
(b), the reflectivity increases, indicating the presence of d-DAMP4 at the interface. The black lines show the fits to the data. (Online version in colour.)
the analysis as there are no Kiessig fringes. However the overall excess of SLD (surface excess) at the surface can be
determined accurately. This is the case for these measurements; as only one solvent condition (8.1% w/w D2O) was
tested, and the films are thin, it is not possible to determine
layer thickness with high precision. However, the fitted thickness values for all datasets were in the range of 10– 14 Å,
which is approximately the thickness observed for AM1, indicating that a monolayer structure is being formed (figure 3).
In the case of DAMP4, this means unfolding of the fourhelix bundle existing in bulk (as indicated by CD spectra
previously collected [15]) to expose the hydrophobic core,
and adsorption to the air –water interface as a connected
‘chain’ of four DAMP1 monomers.
The fitting process yielded SLD profiles for the non-deuterated dataset that were integrated to obtain total scattering
length/unit area at each experimental condition. As the scattering length per molecule is known, the area per DAMP1
monomer at each experimental condition was calculated
(table 3). In general, the values obtained are close, but slightly
lower than the area/molecule of 390 Å2 found for the similar
molecule, AM1 [18], under the same buffer conditions. The pI
of AM1 is higher than that of DAMP4 and DAMP1 (pI 8.5
versus 6.7 and 7.4, respectively), therefore the higher area/
monomer is likely to be due to the increased charge and therefore lower surface ‘packing’ of AM1 at pH 7.4 compared with
DAMP4 and DAMP1. Further, the area per monomer for
Table 3. Parameters from fitting of neutron reflectometry data (figure 2).
DAMP4 bulk
mass fraction
area/a-helix (Å2)
(from fits of figure 2a)
0
0.25
350 + 2.2
374 + 1.4
0.50
0.75
362 + 2.0
369 + 2.4
0.90
360 + 2.2
1.00
366 + 0.7
DAMP1 (350 + 2.2 Å2) is slightly lower than that for DAMP4
(366 + 0.7 Å2), which is consistent with their differences in
pI. The area per monomer values are similar (ranging from
360 + 2.2 to 374 + 1.4 Å2) under all of the mixture conditions.
The scattering length/area values extracted from fitting of
the deuterated dataset were multiplied by corresponding area/
monomer values to obtain the average scattering length/helix
(SLDav) at each condition. The fraction of total area accounted
for by DAMP4 at each condition, calculated using the formula DAMP4 interfacial fraction ¼ (SLDav – SLDDAMP1)/
(SLDd-DAMP4 – SLDDAMP1) (i.e. assuming a linear relationship
between area/helix and SLD) is shown in figure 4. Interestingly,
the amount of surface area accounted for by each component is
very similar to the mass fractions in the original bulk solution
(figure 4). The linear relationship DAMP4 interfacial mass
fraction ¼ DAMP4 bulk mass fraction fits well, yielding a R 2
value of 0.985 (figure 4, dotted line).
3.3. Langmuir-trough compression tests
To investigate interfacial adsorption of these two components
further, interfacial pressure isotherms were measured using
a Langmuir trough. Figure 5 shows the compression –
expansion isotherms for 100 per cent DAMP4, 50 per cent
DAMP4/DAMP1 and 100 per cent DAMP1, after each of
these adsorbed films were allowed to age for 3 h to reach
equilibrium. The starting interfacial pressure for DAMP4 is
25.7 mN m21, higher than that of DAMP1 and the mixed
system, which had starting interfacial pressures of 20.9 and
22.8 mN m21, respectively. The higher interfacial pressure
of DAMP4 corresponds to the low interfacial tension values
it reaches after several hours of aging (shown for 3.2 mM
DAMP4, electronic supplementary material, figure S3). The
area under the curve is related to the energy required for
compression. It can be seen that DAMP4 has a greater
J R Soc Interface 10: 20120987
(b)
1 × 10–7
bulk DAMP4
Figure 3. Adsorption at the air – water interface of unstructured DAMP1 and
structured DAMP4. DAMP1 most likely folds into a single helical unit at the
air – water interface, whereas DAMP4 unfolds from a four-helix bundle in bulk
into a chain of four connected DAMP1 monomers. (Online version in colour.)
1 × 10–7
reflectivity
interfacial DAMP4
rsif.royalsocietypublishing.org
100% h-DAMP4
90% h-DAMP4
75% h-DAMP4
50% h-DAMP4
25% h-DAMP4
0% h-DAMP4
1 × 10–3
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
(a)
y=x
R2 = 0.985
0.8
7
unfolding of
four-helix bundle
0.6
DG
transition
bulk
0.4
DGDAMP4
0.2
0.2
0.4
0.6
0.8
bulk DAMP4 mass fraction
J R Soc Interface 10: 20120987
0
interface
1.0
(b)
Figure 4. The relationship between interfacial and bulk DAMP4 as determined by fitting of deuterated neutron reflectometry data in figure 2.
interfacial pressure (mN m–1)
transition
DG
40
20
10
bulk
DGDAMP1
30
8 mM DAMP1
8 mM DAMP4
50% DAMP4 (1.6 mM DAMP4 6.4 mM DAMP1)
10
20
30
area (cm–2)
40
50
Figure 5. Compression –expansion cycle, performed using a Langmuir
trough, for the conditions 100 per cent DAMP4, 100 per cent DAMP1 and
50 per cent DAMP4. Rate of compression was 114 cm2 min21.
energy barrier to desorption than DAMP1, consistent with
DAMP4 being a larger molecule. The compression isotherm
for the mixed system (50% DAMP4) has an area under the
curve, which is intermediate to that of DAMP4 and
DAMP1. Upon expansion, the slope of the DAMP4 isotherm
is much greater than that of DAMP1 and reaches a lower
final interfacial pressure. This is due to the slower adsorption
kinetics of DAMP4 compared with DAMP1 (observed in
figure 1); the slower the interface is repopulated, the more
‘drop’ in interfacial pressure should be observed. In the case
of the mixed system, the slope of the curve in the expansion
region is between that of DAMP4 and DAMP1. Once the area
was returned to the original value, it took the DAMP4 system
about 1 h to reach the original interfacial pressure while the
DAMP1 system re-reached equilibrium in a few seconds.
3.4. Adsorption behaviour
From the information gathered using interfacial tension
measurements (figure 1), neutron reflectometry (figures 2
and 4), and Langmuir-trough tests (figure 5), an energy–
state diagram for DAMP1 and DAMP4 from the bulk to
the interface can be proposed (figure 6). The slower adsorption
of DAMP4 to the interface compared with DAMP1 (figure 1
and expansion region figure 5) indicates that DAMP4 has a
much higher energy barrier to overcome. This is shown
in (figure 6a) as the greater difference between the bulk and
transition states of DAMP4 compared with that of DAMP1
(figure 6b). As previously mentioned, a significant factor
rsif.royalsocietypublishing.org
interfacial DAMP4 mass fraction
1.0
interface
reaction coordinate
Figure 6. Possible reaction coordinate diagrams for (a) DAMP4, and
(b) DAMP1 adsorbing at the air – water interface.
which most likely contributes to this large energy barrier slowing DAMP4 adsorption is the need for unfolding of the
four-helix bundle. This structure was observed to be very
stable as very little helical structure is lost even when the
bulk temperature is increased to 908C [15]. The neutron reflectometry results indicated that the free energy of adsorption of
the two components is similar. This is shown in figure 6 as similar changes in energy between the bulk and final interfacially
adsorbed states for both DAMP1 and DAMP4. This finding
indicates that in this case the free energy of adsorption is a function only of the molecule chemistry, not size. This is interesting
from the perspective of competitive adsorption of surface-active
molecules, which is not well understood even qualitatively [1].
The compression cycle of the Langmuir-trough experiment
(figure 5) confirms that the energy of desorption is higher for
DAMP4 than for DAMP1, as depicted in figure 6.
While DAMP4 has shown a two-state adsorption mechanism
in this work, it should be noted that larger proteins may have
complicated adsorption kinetics. Adsorption of larger protein
often involves multi-state unfolding kinetics, and a true equilibrium is rarely attained on observable timescales [33]. For a
mixed system of larger proteins, the interaction of two components can further complicate their adsorption. For example,
previous work by Le Floch-Fouéré et al. [34] reported that foaming properties of lysozyme was enhanced by the addition of a
small amount of ovalbumin, suggesting possible synergistic
adsorption of a mixture of model proteins.
3.5. Influence of molecule length on
interfacial rheology
Having confirmed the presence of and quantified the interfacial concentrations of both components, the influence of
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
(a) 10
stress (mN m–1)
6
4
100% DAMP4
100% DAMP1
2
(b) 10
stress (mN m–1)
8
6
4
75% DAMP4
50% DAMP4
25% DAMP4
95% DAMP4
90% DAMP4
80% DAMP4
2
0
(c) 10
stress (mN m–1)
8
6
4
2
0
10
20
30
strain (%)
40
50
60
Figure 7. The mechanical response of the DAMP1 and DAMP4 systems to
extensional strain. (a) DAMP1 and DAMP4 only, (b) mixed systems. For the
DAMP1 and DAMP4-only conditions, only one Maxwell element was required
to describe the data to maximum stress (black solid lines, equation (3.1)).
In the case of the mixed systems, two Maxwell elements were required
(black solid lines, equation (3.2)). (c) Decomposition of the four-parameter
Maxwell model for the 90% DAMP4 condition. The dashed lines show the individual contributions of the first Maxwell element (short dashes), and second
Maxwell element (long dashes), respectively. (Online version in colour.)
molecular size on interfacial rheology was then investigated.
Figure 7a shows the stress response of DAMP1 and DAMP4
(after aging for 3 h to reach equilibrium) to an interfacial
extensional strain deformation of 10 per cent per second.
In the initial region of the stress strain curve (less than 5%
strain), DAMP1 responds with a somewhat higher stress
than DAMP4. It is conceivable that this is due to the slightly
greater ‘packing’ of DAMP1 compared with DAMP4 at
pH 7.4, as evidenced by its slightly lower area per molecule
(table 3). It has been previously observed that the lowstrain elasticity of adsorbed proteins at the air –water
interface is highest around the pI [35– 37], which is likely
where VM is the Maxwell model viscosity parameter, EM is the
Maxwell model elasticity parameter, sM is the stress response
and t is time. Fitting of the Maxwell model (equation (3.1)) to
the DAMP4 and DAMP1 data up to the point of maximum
stress (32 and 40% strain, respectively), provided an excellent
fit (black lines in figure 7a). The extracted parameters were
EM ¼ 89 mN m21 and VM ¼ 104 s mN m21 for DAMP4, and
EM ¼ 180 mN m21 and VM ¼ 48 s mN m21 for DAMP1. The
associated time constants (tM ¼ VM/EM) are 1.17 s for
DAMP4 and 0.27 s for DAMP1, indicating DAMP4 is much
slower to respond to extensional strain than DAMP1. This is
expected as DAMP4 is much larger, and will take longer
to reorient.
The stress response of the mixed DAMP1 and DAMP4
systems previously discussed, as well as two additional conditions, 80 and 95 per cent DAMP4, are shown in figure 7b. There
is very little variance of the stress response in the initial strain
(less than 5%) region between conditions. From about 5 per
cent strain, the curves diverge, and in general, the conditions
with higher DAMP4 fraction reach greater maximum stresses.
The two-parameter Maxwell model was not able to
provide a satisfactory fit to the data in figure 7b. However,
two Maxwell elements in parallel (four-parameter model,
equation (3.2)) [38], provided excellent fits to the point of
maximum stress for all data (black solid lines).
s4p ¼ V1 R 1 eE1 t=V1 þ V2 R 1 eE2 t=V2 :
ð3:2Þ
This finding suggests there are two relaxation processes
occurring in the case of the mixed systems, compared with
J R Soc Interface 10: 20120987
0
8
rsif.royalsocietypublishing.org
8
due to increased packing density and, therefore, ‘jammed
system’ behaviour [1]. For peptides similar to the systems
reported here, there is a clear correlation between the pH of
testing and the measured moduli. Lac21E displayed an extensional elastic modulus of approximately 5 mN m21 when
measured 3.7 pH units away from its pI, but a modulus of
approximately 430 mN m21 close to its pI [18]. AM1 shows
a low modulus of less than 30 mN m21 at approximately
1.1 pH units away from its pI [17], and Lac21 behaves similarly [19]. The stress response of DAMP1 reaches a plateau
of 5 mN m21 at around 10 per cent strain, at which point
it is exceeded by that of DAMP4, which continues to increase
to a maximum value of approximately 10 mN m21. This large
difference in maximum stress achieved between DAMP1 and
DAMP4 is likely to be due to the longer molecule length of
DAMP4. The effect of interfacial networking on interfacial
rheology of DAMP4 and DAMP1 has been investigated in
previous work, by deliberately inducing intermolecular
interactions via metal ion–histidine bridging [38]. In the
absence of such intermolecular interactions (i.e. in the absence
of metal ions), it is unlikely that interfacial networking plays
a significant role in the systems reported here. This is supported by the fact that the extensional moduli of DAMP4 and
DAMP1 in the presence of histidine-bridging metal ions, as
reported in the previous work, is much greater than in its
absence, reported in this current work.
The Maxwell model (equation (3.1)) is the simplest of the
spring and dashpot-based mechanical models, which were
previously discussed in detail in relation to their application
to protein interfaces in the extensional strain model [38].
sM ¼ VM R 1 eEM t=VM ;
ð3:1Þ
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
(a) 200
do
(a)
9
E (mN m–1)
100
E1,4par
E2,4par
EM
d >> d
o
50
DAMP4
DAMP1
(b) 100
V (s mN m–1)
80
60
(b)
do
V1,4par
V2,4par
VM
d >> d
40
o
20
0
0.2
0.4
0.6
0.8
interfacial DAMP4 mass fraction
1.0
Figure 8. The parameters extracted from fits of the data in figure 7, shown
with respect to % DAMP4 at the interface (from figure 4) using equation
(3.1) when only one component is present and equation (3.2) when two
components are present. (Online version in colour.)
one when only DAMP1 or DAMP4 are populating the interface. The total response of the four-parameter Maxwell model
can be decoupled into the individual contributions of each
element, as shown in figure 7c, using the 90 per cent
DAMP4 dataset as an example. The short-dotted line represents the first Maxwell element (first term of equation
(3.2), characterized by the parameters E1 and V1), whereas
the long-dashed line represents the second Maxwell element
(second term of equation (3.2), characterized by E2 and V2).
Representing the four-parameter Maxwell model in this
way allows the contributions of the two Maxwell elements
to be visualized. It can be seen that the first Maxwell
element initially contributes almost all of overall stress
response, plateauing at around 15 –20% strain, at which
point the second element begins to dominate.
Figure 8 shows the parameters EM and VM obtained from
the fits in figure 7a, and E1, V1, E2 and V2, obtained from fits
in figure 7b, as a function of DAMP4 interfacial mass fraction
(assuming interfacial DAMP4 mass fraction ¼ bulk DAMP4
mass fraction, as shown in figure 4). Below the 75 per cent
condition, the parameters are almost constant. E2 is almost
negligible compared with E1, and V1 is greater than V2, indicating dominance of the first Maxwell element across all
strains when less than 75 per cent DAMP4 is present in the
interface. Beyond 75 per cent, the influence of the second
Maxwell element increases, as evidenced by the increase in
E2, decrease in E1 and dominance of V2 over V1.
The physical meaning of the first and second Maxwell
elements may be hypothesized as being determined by
movement in the interfacial plane of DAMP1 and DAMP4,
respectively (figure 9). This hypothesis is supported by the
Figure 9. Depiction of interfacially adsorbed DAMP1 and DAMP4 behaviour in
response to interfacial extensional strain. (a) At interfacial proportions of less
than 75% DAMP4, DAMP1 provides most of the stress-reducing reorientation
and movement. (b) Movement and re-orientation of DAMP4 only occurs at
high interfacial DAMP4 proportions of greater than 75%. (Online version
in colour.)
fact that in the presence of only one component ( pure
DAMP1 and pure DAMP4 in figure 7a), only one Maxwell
element, not two, is needed. The easier it is for molecules
to reorient in response to an applied strain, the less stress
will be registered. Such movement will be a function of molecule packing as well as size. This is reflected in figure 7a,
where the higher molecule packing (lower area per monomer) of DAMP1 compared with DAMP4 results in a higher
initial stress response, then subsequently the larger size of
DAMP4 results in a higher maximum stress. With this in
mind, figure 8 shows that in response to extensional strain,
DAMP4 reorients to redistribute stress very little when
DAMP1 is present (figure 9a), except at very high interfacial
concentrations (greater than 75%, figure 9b). The essentially
unvarying parameters in this region indicate that DAMP1,
the smaller, more mobile molecule literally ‘takes the strain’.
Beyond this apparent critical concentration, the proportion of
DAMP4 becomes significant enough to affect the nature of
the response. Transport kinetics from the bulk and subsurface
may be expected to play a role on the measured stress response,
as previously observed [38], washing-out of the bulk solution
caused a higher stress to be registered at strains of greater
than 15 per cent for b-lactoglobulin. For the systems studied
here however, figure 1c showed that the interfacial tension kinetics were almost identical for all mix ratios, indicating that a
factor other than subsurface/bulk diffusion is causing the
observed variation in extensional stress response.
The question of how a smaller additive will influence the
behaviour of a larger molecule is also pertinent in polymer
literature [39]. The finding that short chains lower the stress
response of long chains is often observed in bulk polymer
J R Soc Interface 10: 20120987
0
rsif.royalsocietypublishing.org
150
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
Using designed sequences, the effect of molecular length on
interfacial rheology and structure was investigated. As the
molecule chemistry was unchanged, the contribution of
other factors such as altered charge structure was eliminated,
allowing the role of molecule size to be isolated and studied
in a way which is difficult to do using natural proteins. The
rate of adsorption of the larger molecule, DAMP4, was
much slower than the smaller molecule, DAMP1. Interfacial
tension kinetics of mixed DAMP4/DAMP1 systems were
much faster than those of the individual components, indicating a synergistic interaction. The reason for the much
slower adsorption of DAMP4 compared with DAMP1 is
likely to be due to its bulk structure, a stable four-helix
bundle, which requires a significant energy barrier to unfold
at the air–water interface. DAMP1 adsorption, however, was
diffusion-limited, as it is largely unstructured in bulk.
Using neutron reflectometry, it was found that the interfacial film structure at equilibrium of all conditions tested
was a monolayer of one a-helix in thickness, similar to the
related molecule, AM1 [18]. DAMP4 is, therefore, likely
unfolding into a chain of four DAMP1 monomers at the air
interface. This was the expected structure as the molecules
are designed to form a-helices when adsorbed interfacially.
The area per molecule (related to the inverse of surface coverage) was lower for pure DAMP1 than for pure DAMP4. This
was attributed to the higher expected charge, therefore
greater molecular repulsion, of DAMP4 owing to its slightly
lower pI ( pI 6.7 compared with 7.4 for DAMP1). The area
per DAMP1 monomer did not vary significantly with
changes in composition of the mixed systems. The relative
concentrations of DAMP4 and DAMP1 at the interface were
close to equal to their original relative mass concentrations
in bulk. This finding indicates that the free energy of adsorption is similar for both DAMP1 and DAMP4 and only
depends on the molecule chemistry, not size. However, the
This investigation was conducted with the financial support of the
Australian Research Council (grant no. DP1093056). L.H. acknowledges the receipt of an AINSE Research Fellowship. A.P.J.M.
acknowledges the support of the Queensland Government through
the award of a Smart Futures Premier’s Fellowship. Access to the
Platypus neutron reflectometer at ANSTO was supported by the Australian Institute of Neutron Science and Engineering (AINSE). The
authors acknowledge Karyn Wilde, Anthony Duff and Peter
Holden at the National Deuteration Facility (ANSTO, Lucas Heights)
for preparation of deuterated DAMP4 cell pellet.
References
1.
2.
Wierenga PA, Gruppen H. 2010 New views on foams
from protein solutions. Curr. Opin. Colloid Interface Sci.
15, 365–373. (doi:10.1016/j.cocis.2010.05.017)
Prins A, Bos MA, Boerboom FJG, van Kalsbeek HKAI.
1998 Relation between surface rheology and foaming
3.
behaviour, vol. 7. Amsterdam, The Netherlands:
Elsevier.
Dickinson E, Murray BS, Stainsby G. 1988
Coalescence stability of emulsion-sized droplets at a
planar oil–water interface and the relationship to
4.
protein film surface rheology. J. Chem. Soc. Faraday
Trans. I 84, 871–883. (doi:10.1039/f19888400871)
Miquelim JN, Lannes SCS, Mezzenga R. 2010 pH
Influence on the stability of foams with protein–
polysaccharide complexes at their interfaces. Food
10
J R Soc Interface 10: 20120987
4. Conclusions
energy barrier to adsorption and desorption is much higher
for the larger molecule DAMP4, as evidenced by the interfacial tension results and Langmuir-trough compression tests.
Insights into the mobility of DAMP1 and DAMP4 molecules at the interface were gained by studying their stress
response to a large (60%) extensional strain applied at constant rate. Below 5 per cent strain, DAMP1 displays a
higher stress response than DAMP4 or the mixed systems.
This is attributed to its higher packing at the pH tested, as
evidenced by the area/helix values obtained by neutron
reflectometry. The single-component data (i.e. pure DAMP4
and DAMP1) were successfully fitted by single-element
Maxwell models (one associated time constant). However,
models with two elements (four-parameter Maxwell model)
were required to fit the mixed-system data, as two relaxation
processes (i.e. movement of both DAMP4 and DAMP1 in
response to imposed strain) occur. The relationship between
interfacial composition (% DAMP4) and interfacial extensional parameters was able to be plotted using the finding
from neutron reflectometry results that the interfacial proportions of DAMP4 and DAMP1 are equal to the bulk
mass proportions. It was proposed that the physical reason
underlying the two relaxation process is the movement of
DAMP1 monomers and movement of the DAMP4 fourmonomer chains in the interfacial plane. The contribution
of the second Maxwell element (i.e. movement of DAMP4)
was only significant at very high interfacial concentrations
of DAMP4 (greater than 75%), indicating that DAMP4 stays
essentially immobile when enough DAMP1 is present to
‘take the strain’, as it is easier for the smaller molecule
to reorient and transfer. These results highlight the qualitative analogy between interfacial protein rheology and bulk
polymer rheology.
The approach applied in this work, namely using designed
sequences with well-defined size, chemistry and bulk structure
combined with contrast-varied neutron reflectometry, accessed
information which is difficult to gain with naturally occurring
proteins. Designed peptide/protein molecules offer a customizable ‘tool kit’ to enable study into the fundamentals of
protein interfacial structure and behaviour which is difficult
to access with naturally existing systems. In this work, we
demonstrated this for the case of a mixed small, unstructured
peptide with a large, structured protein.
rsif.royalsocietypublishing.org
systems. For example, Kornfield et al. [40] observed that short
chains relax first followed by long chains in a polymer bimodal melt. Separate work by Shausberger et al. [41] found that
addition of the short-chain polymer to a long-chain polymer
in a binary blend had the effect of lowering the plateau modulus and relaxation time of the long-chain molecule alone. As
observed in previous work [38], these results further show
parallels between polymer bulk mechanical behaviour and
protein interfacial mechanical behaviour, and indicate the
field of protein interfacial rheology could benefit from application of theoretical frameworks existing in the mature
field of polymer rheology. Designed sequences such as
those introduced here would serve as effective tools in
validating such approaches.
Downloaded from http://rsif.royalsocietypublishing.org/ on June 16, 2017
6.
8.
9.
10.
11.
12.
13.
14.
15.
16.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
interfacially adsorbed protein film. Chem. Eng. Sci.
57, 1711 –1722. (doi:10.1016/S00092509(02)00057-X)
Miller R, Aksenenko EV, Fainerman VB, Pison U.
2001 Kinetics of adsorption of globular proteins at
liquid/fluid interfaces. Colloid Surf. A Physicochem.
Eng. Asp. 183, 381 –390. (doi:10.1016/s09277757(01)00523-4)
Thomas RK. 2004 Neutron reflection from liquid
interfaces. Annu. Rev. Phys. Chem. 55, 391–426.
(doi:10.1146/annurev.physchem.54.011002.103830)
Menger FM, Rizvi SAA. 2011 Relationship between
surface tension and surface coverage. Langmuir 27,
13 975 –13 977. (doi:10.1021/la203009m)
Graham DE, Phillips MC. 1979 Proteins at liquid
interfaces. 1. Kinetics of adsorption and surface
denaturation. J. Colloid Interface Sci. 70, 403 –414.
(doi:10.1016/0021-9797(79)90048-1)
Le Floch-Fouere C, Pezennec S, Lechevalier V,
Beaufils S, Desbat B, Pezolet M, Renault A. 2009
Synergy between ovalbumin and lysozyme leads to
non-additive interfacial and foaming properties of
mixtures. Food Hydrocolloids 23, 352 –365.
(doi:10.1016/j.foodhyd.2008.01.007)
Pezennec S, Gauthier F, Alonso C, Graner F,
Croguennec T, Brule G, Renault A. 2000 The protein
net electric charge determines the surface
rheological properties of ovalbumin adsorbed at
the air– water interface. Food Hydrocolloids 14,
463–472. (doi:10.1016/S0268-005X(00)00026-6)
Roberts SA, Kellaway IW, Taylor KMG, Warburton B,
Peters K. 2005 Combined surface pressure-interfacial
shear rheology study of the effect of pH on the
adsorption of proteins at the air–water interface.
Langmuir 21, 7342 –7348. (doi:10.1021/la050272l)
Malcolm AS, Dexter AF, Middelberg APJ. 2006
Mechanical properties of interfacial films formed by
lysozyme self-assembly at the air–water interface.
Langmuir 22, 8897 –8905. (doi:10.1021/la060565u)
Dimitrijev-Dwyer M, Middelberg APJ. 2011 The
extensional viscoelasticity of protein-coated
interfaces. Soft Matter 7, 7772– 7781. (doi:10.1039/
c1sm05253e)
Knoglinger H, Schausberger A, Janeschitzkriegl H.
1987 The role of short chain molecules for the
rheology of polystyrene melts. I. Molar mass
dependent shift factors. Rheol. Acta 26, 460–467.
(doi:10.1007/bf01333847)
Kornfield JA, Fuller GG, Pearson DS. 1989 Infrared
dichroism measurement of molecular relaxation in
binary blend melt rheology. Macromolecules 22,
1334– 1345. (doi:10.1021/ma00193a055)
Schausberger A, Knoglinger H, Janeschitzkriegl H.
1987 The role of short chain molecules for the
rheology of polystyrene melts. II. Linear viscoelastic
properties. Rheol. Acta 26, 468–473. (doi:10.1007/
bf01333848)
11
J R Soc Interface 10: 20120987
7.
17.
protein hydrolysates. Soft Matter 8, 5131–5139.
(doi:10.1039/c2sm25082a)
Dexter AF, Malcolm AS, Middelberg APJ. 2006
Reversible active switching of the mechanical
properties of a peptide film at a fluid–fluid interface.
Nat. Mater. 5, 502–506. (doi:10.1038/nmat1653)
Middelberg A, He L, Dexter A, Shen H, Holt S,
Thomas R. 2008 The interfacial structure and
Young’s modulus of peptide films having switchable
mechanical properties. J. R. Soc. Interface 5, 47 –54.
(doi:10.1098/rsif.2007.1063)
Dexter AF, Middelberg APJ. 2007 Switchable
peptide surfactants with designed metal binding
capacity. J. Phys. Chem. C 111, 10 484–10 492.
(doi:10.1021/jp071554s)
Makievski AV, Fainerman VB, Miller R, Bree M,
Liggieri L, Ravera F. 1997 Determination of
equilibrium surface tension values by extrapolation
via long time approximations. Colloid Surf. A
Physicochem. Eng. Asp. 122, 269–273. (doi:10.
1016/s0927-7757(96)03857-5)
Middelberg APJ, Radke CJ, Blanch HW. 2000 Peptide
interfacial adsorption is kinetically limited by the
thermodynamic stability of self association. Proc.
Natl Acad. Sci. USA 97, 5054–5059. (doi:10.1073/
pnas.080042597)
Tyn MT, Gusek TW. 1990 Prediction of diffusion
coefficients of proteins. Biotechnol. Bioeng. 35,
327 –338. (doi:10.1002/bit.260350402)
Nelson A. 2010 Motofit—integrating neutron
reflectometry acquisition, reduction and analysis into
one, easy to use, package. J. Phys.: Conf. Ser. 251,
012094. (doi:10.1088/1742-6596/251/1/012094)
James M, Nelson A, Holt SA, Saerbeck T, Hamilton
WA, Klose F. 2011 The multipurpose time-of-flight
neutron reflectometer ‘Platypus’ at Australia’s OPAL
reactor. Nuclear Instrum. Methods Phys. Res. Sec. A,
Accel. Spectrom. Detect. Assoc. Equip. 632, 112–
123. (doi:10.1016/j.nima.2010.12.075)
Nelson A. 2006 Co-refinement of multiple-contrast
neutron/X-ray reflectivity data using MOTOFIT.
J. Appl. Crystallogr. 39, 273 –276. (doi:10.1107/
S0021889806005073)
Sears VF. 1992 Neutron scattering lengths and cross
sections. Neutron News 3, 26 –37. (doi:10.1080/
10448639208218770)
Jacrot B, Zaccai G. 1981 Determination of molecular
weight by neutron scattering. Biopolymers 20,
2413 –2426. (doi:10.1002/bip.1981.360201110)
He LZ, Malcolm AS, Dimitrijev M, Onaizi SA, Shen HH,
Holt SA, Dexter AF, Thomas RK, Middelberg APJ. 2009
Cooperative tuneable interactions between a designed
peptide biosurfactant and positional isomers of SDOBS at
the air–water interface. Langmuir 25, 4021–4026.
(doi:10.1021/la802825c)
Jones DB, Middelberg APJ. 2002 Direct
determination of the mechanical properties of an
rsif.royalsocietypublishing.org
5.
Hydrocolloids 24, 398–405. (doi:10.1016/j.foodhyd.
2009.11.006)
Croguennec T, Renault A, Bouhallab S, Pezennec S.
2006 Interfacial and foaming properties of sulfydrylmodified bovine beta-lactoglobulin. J. Colloid
Interface Sci. 302, 32 –39. (doi:10.1016/j.jcis.
2006.06.061)
Partanen R, Paananen A, Forssell P, Linder MB, Lille
M, Buchert J, Lantto R. 2009 Effect of
transglutaminase-induced cross-linking of sodium
caseinate on the properties of equilibrated
interfaces and foams. Colloid Surf. A Physicochem.
Eng. Asp. 344, 79– 85. (doi:10.1016/j.colsurfa.
2009.02.004)
Martin AH, Grolle K, Bos MA, Stuart MA, van Vliet T.
2002 Network forming properties of various
proteins adsorbed at the air/water interface in
relation to foam stability. J. Colloid Interface Sci.
254, 175–183. (doi:10.1006/jcis.2002.8592)
Davis JP, Foegeding EA. 2007 Comparisons of the
foaming and interfacial properties of whey protein
isolate and egg white proteins. Colloid Surf. B
Biointerfaces 54, 200–210. (doi:10.1016/j.colsurfb.
2006.10.017)
Murray BS. 2011 Rheological properties of protein
films. Curr. Opin. Colloid Interface Sci. 16, 27 –35.
(doi:10.1016/j.cocis.2010.06.005)
Hamada JS, Swanson B. 1994 Deamidation of food
proteins to improve functionality. Crit. Rev. Food Sci.
Nutr. 34, 283 –292. (doi:10.1080/
10408399409527664)
Davis JP, Doucet D, Foegeding EA. 2005 Foaming
and interfacial properties of hydrolyzed
b-lactoglobulin. J. Colloid Interface Sci. 288,
412–422. (doi:10.1016/j.jcis.2005.03.002)
Panyam D, Kilara A. 1996 Enhancing the
functionality of food proteins by enzymatic
modification. Trends Food Sci. Technol. 7, 120–125.
(doi:10.1016/0924-2244(96)10012-1)
Foegeding EA, Davis JP, Doucet D, McGuffey MK.
2002 Advances in modifying and understanding
whey protein functionality. Trends Food Sci.
Technol. 13, 151 –159. (doi:10.1016/s0924-2244
(02)00111-5)
Jung JM, Gunes DZ, Mezzenga R. 2010 Interfacial
activity and interfacial shear rheology of native
beta-lactoglobulin monomers and their heatinduced fibers. Langmuir 26, 15 366– 15 375.
(doi:10.1021/la102721m)
Middelberg APJ, Dimitrijev-Dwyer M. 2011 A
designed biosurfactant protein for switchable foam
control. ChemPhysChem 12, 1426 –1429. (doi:10.
1002/cphc.201100082)
Dimitrijev-Dwyer M, He LZ, James M, Nelson A,
Wang LG, Middelberg APJ. 2012 The effects of acid
hydrolysis on protein biosurfactant molecular,
interfacial, and foam properties: pH responsive