Microphase Behavior and Enhanced Wet-Cohesion of Synthetic Copolyampholytes Inspired by a Mussel Foot Protein

Article abstract of DOI:10.1021/jacs.5b03827

Numerous attempts have been made to translate mussel adhesion to diverse synthetic platforms. However, the translation remains largely limited to the Dopa (3,4-dihydroxyphenylalanine) or catechol functionality, which continues to raise concerns about Dopa

Full text of DOI:10.1021/jacs.5b03827

Communication
pubs.acs.org/JACS
Microphase Behavior and Enhanced Wet-Cohesion of Synthetic
Copolyampholytes Inspired by a Mussel Foot Protein
Sungbaek Seo,†,# Saurabh Das,†,‡,# Piotr J. Zalicki,†,§ Razieh Mirshafian,† Claus D. Eisenbach,∥,⊥
Jacob N. Israelachvili,‡ J. Herbert Waite,†,§ and B. Kollbe Ahn*,†
‡
§
∥
†Marine Science Institute, Chemical Engineering, Molecular, Cellular and Developmental Biology, and Materials Research
Laboratory, University of California, Santa Barbara, California 93106, United States
⊥Institute for Polymer Chemistry, University of Stuttgart, D-70569 Stuttgart, Germany
S
* Supporting Information
ABSTRACT: Numerous attempts have been made to
translate mussel adhesion to diverse synthetic platforms.
However, the translation remains largely limited to the
Dopa (3,4-dihydroxyphenylalanine) or catechol function-
ality, which continues to raise concerns about Dopa’s
inherent susceptibility to oxidation. Mussels have evolved
adaptations to stabilize Dopa against oxidation. For
example, in mussel foot protein 3 slow (mfp-3s, one of
two electrophoretically distinct interfacial adhesive pro-
teins in mussel plaques), the high proportion of hydro-
phobic amino acid residues in the flanking sequence
around Dopa increases Dopa’s oxidation potential. In this
study, copolyampholytes, which combine the catechol
functionality with amphiphilic and ionic features of mfp-3s,
were synthesized and formulated as coacervates for
adhesive deposition on surfaces. The ratio of hydro-
philic/hydrophobic as well as cationic/anionic units was
varied in order to enhance coacervate formation and wet
adhesion properties. Aqueous solutions of two of the four
mfp-3s-inspired copolymers showed coacervate-like spher-
ical microdroplets (ϕ ≈ 1−5 μm at pH ∼4 (salt
concentration ∼15 mM). The mfp-3s-mimetic copolymer
was stable to oxidation, formed coacervates that spread
evenly over mica, and strongly bonded to mica surfaces
(pull-off strength: ∼17.0 mJ/m2). Increasing pH to 7 after
coacervate deposition at pH 4 doubled the bonding
strength to ∼32.9 mJ/m2 without oxidative cross-linking
and is about 9 times higher than native mfp-3s cohesion.
This study expands the scope of translating mussel
adhesion from simple Dopa-functionalization to mimick-
ing the context of the local environment around Dopa.
Figure 1. Marine mussel and the approximate location of mussel foot
proteins in the byssus. (a) Mytilus californianus attached to poly(methyl
methacrylate) surface. (b) Schematic of the types and distribution of
mfps.
Here, we studied the microphase behavior and wet-adhesion of
copolyampholytes with fixed catechol content and varied other
key functionalities. Potential effects of aromatic moieties (Tyr,
Trp) besides Dopa in mfp-3s have not been specifically tested in
the present structural design of the model copolyampholytes.
Conditions for the experiments were adjusted according to the
microenvironmental conditions of adhesive protein deposition
under the mussel’s foot including acidic to neutral pH and ionic
strength of ≤100 mM.3,4
In mussel adhesion, polyelectrolyte adhesive proteins or mfps
are presented to target surfaces after being condensed as a dense
fluid by complex coacervation, a critical step in the formation of
protein-based underwater adhesives.5 The first synthetic adhesive
to be studied as a complex coacervate was modeled after
sandcastle worm cement and consisted of two oppositely charged
dopamine-functionalized polyelectrolytes.6,7 Coacervation bene-
fits adhesion in several important ways: (1) high polymer
concentration increases density, (2) the low interfacial energy
improves wetting, (3) high diffusivity maintains good mixing, and
(4) reduced viscosity eases delivery.1
Mfp-3s is localized to the plaque−substratum interface (Figure
1b) together with mfp-3f and mfp-5. Due to its amphiphilic and
ampholytic structural characteristics (Figure 2a), mfp-3s is
capable of self-coacervation8 and is more stable at oxidation
than other mfps.9 Intrigued by mfp-3s’s (molecular weight ∼5
kDa) unique property to self-coacervate, we synthesized a series
of ampholytic copolymers consisting of randomly arranged
catechol (M1)-functionalized, cationic (M2), anionic (M3),
arine mussels (Figure 1a) attach to hard surfaces, e.g.,
Mmineral and metal, in the intertidal zone where waves with
and without suspended sand often exceed 25 m/sec velocities.
3,4-Dihydroxyphenylalanine (Dopa), a main constituent in
mussel foot proteins (mfps) and substantially contributing to
wet adhesion, has been incorporated in synthetic polymers to
mimic the bio wet-adhesion.1−5 However, other constitutional
features of mfps, e.g., cationic residues (lysine, K), anionic
residues (aspartic acid, D), nonionic polar residues (asparagine,
N), and nonpolar residues (alanine, A), have not typically been
included in mussel-inspired synthetic wet-adhesion systems.1,2
Received: April 20, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b03827
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
copolymers differ from those in mfp-3s due to the different nature
of the polymer backbones and side chain chemistry; however,
overall structural tendencies are preserved. The primary objective
here is not to mimic specific functionalities in mfp-3s but to
produce a relatively simple and inexpensive model polymer for
fundamental structure−properties studies. Inspired by the
nonrepetitive amino acid sequence of mfp-3s (Figure 2a),9
random copolymers were synthesized via free radical copoly-
merization. Random monomer distribution of the copolymers
was confirmed by monitoring the copolymerization from time
zero through 15, 30, 45, 60, and 720 min: samples taken from the
reaction mixture were analyzed by gas chromatography (GC), gel
permeation chromatography (GPC), and nuclear magnetic
resonance spectroscopy (NMR). Details of the analytical
protocols regarding monomer conversion (or consumption)
determined by GC, polymer yield and total nonvolatile monomer
conversion determined by GPC, and copolymer composition
determined by NMR are given in the Supporting Information
(SI). The composition of the copolymer was consistent during
the entire reaction, demonstrating the random comonomer
distribution along the copolymer chains. The copolyampholytes
were obtained after removal of the protective groups from these
precursor copolymers (see SI for details).
Figure 2. Key features of mfps and synthetic homologs. (a) Primary
sequence of mfp-3s. (b) Pie chart of key functionalities in mfp-3s and
synthetic analogues: copolymer 1 (P1) to copolymer 4 (P4). (c)
Chemical composition of a copolyacrylate with randomly distributed
mfp-3s-mimetic functionalities.
The microphase behavior of aqueous solutions of the
deprotected copolyampholyte, i.e., occurrence of a single-phase
solution, liquid−liquid (coacervation), or liquid−solid (precip-
itation) two-phase separation was investigated and related to the
wet adhesive properties of the materials. Increasing the mole ratio
P/NP of nonionic polar (hydrophilic; M4 based) and nonpolar
(hydrophobic; M5 based) comonomers from 20/40 to 60/0 in
the copolymer increased its water solubility gradually from 0 to
0.5 wt % at pH > 8, where the tertiary-amino group of DEAEA-
based repeat units start to be deprotonated. Hence the polymer
concentration used in this study was 0.01 wt %. Whereas both P2
and P4 showed dynamic microphase behavior, both P1 and P3
were insoluble. P2 and P4, however, were soluble only at pH ≥ 8
due to charge screening, hence a slight concentration of Na+ and
Cl− inevitably was accumulated after adjusting the pH of the
solution by addition of 0.1 M NaOH or HCl (see SI for more
details). The insolubility of P1 in aqueous medium is attributed to
its high content of hydrophobic M5 comonomer. This is
supported by the solubility of P2, which differs from P1 by the
distinct increase in the content of the hydrophilic M4-based
comonomer and corresponding decrease of M5 comonomer. In
comparison to P2, increasing charge densities (doubling of both
positive and negative charges) to the expense of nonionic
hydrophilicM4-basedcomonomerinP3resultedinprecipitation.
This is attributed to an increased number of short-range
electrostatic interactions between acid and base side groups.13
At pH ≈ 4 (salt concentration ≈ 15 mM), both P2 and P4
showed spherical microdroplets (coacervates: ϕ ≈ 1−5 μm)
(Figure 3a). However, at pH ≈ 3 (salt concentration ∼20 mM),
the copolymers precipitated (Figure 3a). This is a consequence of
protonation of carboxylate side-chains, as pH was decreased
below the pKa (∼4) of carboxylic groups in the copolymer chain.
AtpH≥7(saltconcentration≤5mM),particularlyatpHcloseto
the pKa (∼8) of DEAEA, a single phase solution of the copolymer
was obtained as inferred from both the micrograph (Figure 3a)
and turbidity data (Figure 3b). The phase behavior at pH 2 to 9
(salt concentration ≤ 20 mM) (Figure 3a,b) described above was
identical to that measured at a fixed ionic strength of ∼10 mM.
Similarly, mfp-3s showed dynamic microphase behavior at pH 3
to 7.5 at ionic strengths ≤ 100 mM.8 We also studied phase
nonionic hydrophilic (M4), and hydrophobic (M5) comonomer
units (Figure 2b,c). These synthetic copolyampholytes of fixed
catechol content and varied ratio of hydrophilic/hydrophobic as
well as acid/base units (Figure 2b) carry key constituent features
of mfp-3s. Copolymer 1 (P1) was synthesized following the ratio
of catecholic, cationic, anionic, nonionic polar, and nonpolar
amino acid units of mfp-3s. In copolymer 2 (P2), hydrophobicity
was decreased by increasing the fraction of nonionic polar (P;
hydrophilic) groups (comonomer M4) at the expense of
nonpolar (NP; hydrophobic) groups (comonomer M5). In
copolymer 3 (P3), the overall charge density was increased by
doubling the levels of both cationic and anionic residues by
keeping the total fraction of comonomers M2, M3, and M4 the
same as in P2. In copolymer 4 (P4), in comparison to P1 the
nonpolarhydrophobic residue (comonomerM5)was completely
omitted, i.e., its intrinsic hydrophobicity originates from the
polymer backbone, which is hydrophilic (peptide linkage) in
mfps.
The literature on synthetic ampholytic copolymers in aqueous
solution contains only one example where liquid−liquid phase
separation was observed: Copolymers with 60/40 (or 35/65)
mole ratios in their acid/base repeat unit exhibited a two-phase
liquid−liquid system at low salinity, whereas all other copolymer
compositions gave either solid−liquid two-phase systems or were
completely soluble, irrespective of the ion strength.10 In other
words, the coacervation-reminiscent phase behavior occurred
only in a relatively narrow nonstoichiometric acid/base
comonomer ratio range. In the light of these findings and also
considering the unequal content of anionic and cationic residues
inmfp-3s(4vs6mol%),11 theacid/basecomonomerratiosofthe
copolyampholytes synthesized in this study were adjusted
accordingly: 6 and 4 mol % cationic and anionic units,
respectively, as in mfp-3s, or twice as high at 12 and 8 mol %,
respectively.
Building upon the thoroughly investigated complex coacerva-
tion between poly(acrylic acid) and poly(2-dimethyl aminoethyl
methacrylate) or poly(allyl amine),11,12 acrylic acid (AA), and 2-
(diethylamino)ethyl acrylate (DEAEA) were chosen as oppo-
sitely charged comonomers in the synthesis of the copolyam-
pholytes (Figure 2c). The specific functional groups in the model
B
DOI: 10.1021/jacs.5b03827
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 3. Microphase behavior and cyclic voltammograms (CV) of mfp-
3s-mimetic copolyampholytes. (a) Light microscopic images of P2 at pH
3, 3.5, 4, 4.5, 5, and 5.5 (images for P4 not shown, followed same trend).
(b) Turbidity of 0.01 wt % P2 in H2O at pH 2−9 (data for P4 not shown,
followed same trend). (c) CV of 0.01 wt % P2 (upper panel) and P4
(lower panel), respectively, in H2O at pH 3, 4, and 7.
Figure 4. Representative force vs distance plots between P2 films
absorbed onto mica in SFA. (a) Interaction energy (cohesion) between
P2 films on mica, adsorbed and measured at pH 3 and 7, respectively. (b)
Interaction energy between P2 films on mica, measured at pH 4 (black),
pH 3 (red), and pH 7 (green), respectively, after adsorption at pH 4.
behavior of the coacervate formed at pH 4 as a function of salt
concentration. The copolymer started to precipitate above 100
mMsaltwhereprecipitationisattributedtodecreasingthesolvent
quality (aq. NaCl solution) for the copolyampholyte with
increasing salt concentration.11
distance curves between two mica surfaces coated with P2 at
different pH are shown in Figure 4. In the SFA, adhesion refers to
the attraction between two different surfaces, whereas cohesion is
between two similar surfaces. Although P2 and P4 showed
identical microphase behaviors, only P2 exhibited cohesion in the
SFA at pH 4 (optimized coacervation condition for both P2 and
P4). Hence our further studies focused on P2. The cohesive
strength between the adsorbed films (of P2) as measured in the
SFA in this study was independent of the contact time (tc) and
equilibration time (te) except at pH 7, where the cohesion
increased with te. As shown in Figure 4, the polymer films,
adsorbed and measured at pH 3, exhibited a relatively thick hard-
wall (∼20 nm) with evident repulsion beginning at surface
separations of 80 nm and very low cohesion (Wc < 0.1 mJ/m2, tc
and te independent). This is understood when considering that
the copolymer precipitated at pH 3, and the measured data reflect
thehardcontactwithnonstickydepositsofbulkcopolymeronthe
mica surfaces (see AFM data in the SI). However, the polymer
films adsorbed and measured at pH 4or 7, i.e., from the two-phase
liquid−liquid system, resulted in very thin hard-walls (<1 nm),
reminiscent of monomolecular films between the mica surfaces.
However, there was a difference between pH 4 and 7 during
approach. Although polymer layers adsorbed at both pH 4 and 7
exhibited repulsion starting at ∼20 nm distance indicating the
thickness of the Langmuir adsorption layer, the films adsorbed at
pH 7 show a stronger repulsion starting at D ≈ 20 nm on
approach, whereas the films adsorbed at pH 4 exhibited jump-in
instability (due to strong attractive interaction between the
surface films during approach of the surfaces) at D ≈ 7 nm
distance from the final hard-wall (black solid circles in Figure 4b).
The cohesion of the films adsorbed and measured at pH 4 (17.0
2.5 mJ/m2, tc independent) surpassed the maxima of mfp-5 (13.7
0.5 mJ/m2, tc ≥ 1 h), the most adhesive protein tested.15
Cyclic voltammetry (CV) of P2 (more hydrophobic) showed
anoxidationpotential(E0)ofcatechol, ∼0.50VatpH3and∼0.31
V at pH 4 and 7 (Figure 3c), similar to E0 of Dopa in mfp-3s, ∼0.5
V at pH 2 and ∼0.35 V at pH 7.5.9 These are significantly higher
than E0 of catechol in P4 (less hydrophobic), ∼0.45 V at pH 3,
∼0.22VatpH4, and∼0.18VatpH7(Figure3c). IntheCVofP2,
the amount of catecholic and vinyl-catecholic (an α,β-dehydro
derivative of Dopa oxidation in mfps that arises from quinone
tautomerization)14 functionalities at pH 4 was similar to that at
pH 7, whereas P4 did not exhibit the vinyl-catechol at pH 4 and 7.
Comparable experiments carried out between P2 and P4 showed
an oxidation inhibiting effect of hydrophobic comonomer units.
UV−vis absorption spectra (see Figure S13) at pH 3, 4, and 7 also
show similar amounts of catechol (∼280 nm) and vinyl-catechol
(320−330 nm), whereas no quinone (∼390 nm) was observed
evenatpH7, whereDopaoxidation occursgenerally. TheCVand
UV−vis results suggest the high proportion of hydrophobic
residues in the copolymer provides stability against catechol
oxidation by shielding from the solvent just as mfp-3s showed its
ability to maintain adhesion at neutral pH with comparable
oxidative stability.9
The surface forces apparatus (SFA) was used to investigate the
cohesion of the copolyampholytes (P2 and P4, respectively)
adsorbedontomolecularlysmoothmineralsurfaces(e.g., mica)at
pH 3, 4, and 7, respectively, (please see SI for more details about
SFA). Hard-wall thickness of the copolyampholytes was also
measured as the limiting distance between the mica surfaces
during the approach run in the SFA. Representative force−
C
DOI: 10.1021/jacs.5b03827
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
However, thecohesion(tc independent)ofthefilmsadsorbedand
measured at pH 7 increased monotonically with te from no
cohesion (te ≈ 2−10 min) to 1.6 0.4 mJ/m2 (te ≈ 60 min), 15.2
0.4 mJ/m2 (te ≈ 360 min), and 24.3 1.2 mJ/m2 (te ≥ 1080
min). This te-dependent cohesion at pH 7 suggests that the
solution pH not only affects adsorption but also the cohesion of
the polymer films due to changing conformation of the residues
responsible for the bridging of the surface films or the adhesion to
mica surface. As shown by the SFA, AFM measurements (see
Figure S12) confirmed that the copolymer coated the mica
surface more effectively at pH 4 (coacervate) than at pH 7 (a
single phase solution) or at pH 3 (precipitate). Coating patches
(<3nm)thatpersistonmicaafterthoroughrinsingwereobserved
only when the mica was coated with the coacervate but not with
the soluble phase or precipitates (please see SI for more details).
To further investigate pH effects on the cohesion of the films
adsorbed onto mica at pH 4 (coacervate), the cohesion was
measured under a new pH condition (no salt pH 3 or 7) after the
adsorption (Figure 4b). For the films absorbed at pH 4, the
cohesion was tc independent regardless of the changed pH
condition. The cohesive interaction between the polymer films
disappeared at pH 3, whereas cohesion increased to 32.9 3.5
mJ/m2 at pH 7. The cohesion between the films, adsorbed
(coacervate) at pH 4 and measured at pH 7 (32.9 3.5 mJ/m2),
was∼1.4timesgreaterthanthatadsorbedatpH7(asinglesoluble
phase) and measured at pH 7 with te ≥ 18 h (24.3 1.2 mJ/m2).
The SFA and AFM results demonstrate that coacervation is
essential for uniform coating of the mica surface and for obtaining
higher cohesion. When adsorbed at pH 4 and re-equilibrated to
pH 7, the stronger cohesion between the polymer films at the
higher pH 7 may be due to the decreased Coulombic repulsion
since the isoelectric point of the polymer is near neutral pH (pI
6.7). The weak repulsive forces between the polymer films can be
translated into a surface charge density of 0.03 C/m2 (surface
potential, ψ = 120 mV) and 0.018 C/m2 (ψ = 40 mV) at pH 4 and
pH 7, respectively.16 Also, the strong cohesion at pH 7 was
reversible (i.e., similar cohesion forces were measured during
subsequent approach−separation force runs at the same contact
point with no material transfer across surfaces), suggesting
Dopa−Dopaquinone induced cross-linking between the films
was not the operative mechanism of cohesion. CV and UV−vis
measurements confirm the absence of Dopa−Dopaquinone
induced cross-linking to the strong cohesion at neutral pH. The
levels of catechol and vinyl-catechol in the polymer at pH 4 were
similar to those at pH 7, and there were no oxidative products of
catechol (e.g., quinone) at pH 4 nor even at pH 7, where catechol
(or Dopa) oxidation typically occurs. This oxidation stability
could be due to hydrophobic or electrophilic shielding of the
Dopa moieties as proposed for mfp-3s.9
ASSOCIATED CONTENT
* Supporting Information
Additional information and experimental methods. The Support-
ingInformationisavailablefreeofchargeontheACSPublications
website at DOI: 10.1021/jacs.5b03827.
■
S
AUTHOR INFORMATION
Corresponding Author
*ahn@lifesci.ucsb.edu
■
Author Contributions
#S.S. and S.D. contributed equally to this work.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors gratefully acknowledge financial support from Office
of Naval Research N000141310867 and National Science
Foundation MRSEC DMR-1121053. Authors wish to thank
Mary Raven in NRI/MCDB Microscopy Facility (NIH 1 S10
OD010610-01A1), Rachel Berhrens in MRL, and Bruce Lipshutz
for the use of his autocolumn.
REFERENCES
■
(1) Stewart, R. J.; Ransom, T. C.; Hlady, V. J. Polym. Sci., Part B: Polym.
Phys. 2011, 49, 757.
́ ́ ̂
(2) Faure, E.; Falentin-Daudre, C.; Jerome, C.; Lyskawa, J.; Fournier,
D.; Woisel, P.; Detrembleur, C. Prog. Polym. Sci. 2013, 38, 236.
(3) Martinez Rodriguez, N. R.; Das, S.; Kaufman, Y.; Israelachvili, J. N.;
Waite, J. H. Biofouling 2015, 31, 221.
(4)Yu, J.;Wei, W.;Danner, E.; Ashley, R. K.;Israelachvili, J. N.;Waite, J.
H. Nat. Chem. Biol. 2011, 7, 588.
(5) Ortony, J. H.; Hwang, D. S.; Franck, J. M.; Waite, J. H.; Han, S.
Biomacromolecules 2013, 14, 1395.
(6) Shao, H.; Stewart, R. J. Adv. Mater. 2010, 22, 729.
(7) Shao, H.; Bachus, K. N.; Stewart, R. J. Macromol. Biosci. 2009, 9, 464.
(8) Wei, W.; Tan, Y.; Martinez Rodriguez, N. R.; Yu, J.; Israelachvili, J.
N.; Waite, J. H. Acta Biomater. 2014, 10, 1663.
(9) Wei, W.; Yu, J.; Broomell, C.; Israelachvili, J. N.; Waite, J. H. J. Am.
Chem. Soc. 2013, 135, 377.
(10) Corpart, J. M.; Candau, F. Macromolecules 1993, 26, 1333.
(11) Spruijt, E.; Westphal, A. H.; Borst, J. W.; Cohen Stuart, M. A.; van
der Gucht, J. Macromolecules 2010, 43, 6476.
(12)Chollakup, R.;Beck, J. B.;Dirnberger, K.;Tirrell, M.;Eisenbach, C.
D. Macromolecules 2013, 46, 2376.
(13) Wang, Y.; Kimura, K.; Huang, Q.; Dubin, P. L.; Jaeger, W.
Macromolecules 1999, 32, 7128.
(14) Lee, B. P.;Dalsin, J. L.;Messersmith, P. B. Biomacromolecules 2002,
3, 1038.
(15) Danner, E. W.; Kan, Y.; Hammer, M. U.; Israelachvili, J. N.; Waite,
J. H. Biochemistry 2012, 51, 6511.
(16) Das, S.; Donaldson, S. H., Jr.; Kaufman, Y.; Israelachvili, J. N. RSC
Adv. 2013, 3, 20405.
(17) Liu, B.; Burdine, L.; Kodadek, T. J. Am. Chem. Soc. 2006, 128,
15228.
(18)Matos-Perez, C. R.;White, J. D.;Wilker, J. J. J. Am. Chem. Soc.2012,
134, 9498.
(19) Harrington, M. J.; Masic, A.; Holten-Andersen, N.; Waite, J. H.;
This study emphasizes the importance of the balance between
electrostatic and hydrophobic interactions for coacervation and
wet-adhesion in addition to catecholic interactions, e.g., oxidative
cross-linking,17,18 metal coordination,19,20 and intermolecular
hydrogen bonding.21 The mfp-3s-mimetic copolyampholyte has
potentialasahighperformance wetadhesive/coating withitsvery
strong wet-cohesion (∼8.8, ∼2.4, and ∼1.6 times greater than
mfp-3s,9 mfp-5,15 the most adhesive natural protein, and the
recently engineered mfp-amyloid protein,22 respectively) and
stablecoacervation. Thisstudyexpands thescope oftranslation of
the biological organism by combining Dopa with hydrophobic
and electrostatic functionalities.
Fratzl, P. Science 2010, 328, 216.
(20) Das, S.; Miller, D. R.; Kaufman, Y.; Martinez Rodriguez, N. R.;
Pallaoro, A.; Harrington, M. J.; Gylys, M.; Israelachvili, J. N.; Waite, J. H.
Biomacromolecules 2015, 16, 1002.
(21) Ahn, B. K.; Lee, D. W.; Israelachvili, J. N.; Waite, J. H. Nat. Mater.
2014, 13, 867.
(22)Zhong, C.;Gurry, T.;Cheng, A. A.;Downey, J.;Deng, Z.;Stultz, C.
M.; Lu, T. K. Nat. Nanotechnol. 2014, 9, 858.
D
DOI: 10.1021/jacs.5b03827
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX