-
LETTERS
Designed biomaterials to mimic the mechanicalproperties of
musclesShanshan Lv1, Daniel M. Dudek2{, Yi Cao1, M. M. Balamurali1,
John Gosline2 & Hongbin Li1
The passive elasticity of muscle is largely governed by the
I-bandpart of the giant muscle protein titin1–4, a complex
molecular springcomposed of a series of individually folded
immunoglobulin-likedomains as well as largely unstructured unique
sequences5. Thesemechanical elements have distinct mechanical
properties, and whencombined, they provide the desired passive
elastic properties ofmuscle6–11, which are a unique combination of
strength, exten-sibility and resilience. Single-molecule atomic
force microscopy(AFM) studies demonstrated that the macroscopic
behaviour oftitin in intact myofibrils can be reconstituted by
combining themechanical properties of these mechanical elements
measured atthe single-molecule level8. Here we report artificial
elastomericproteins that mimic the molecular architecture of titin
throughthe combination of well-characterized protein domains GB112
andresilin13. We show that these artificial elastomeric proteins
can bephotochemically crosslinked and cast into solid biomaterials.
Thesebiomaterials behave as rubber-like materials showing high
resili-ence at low strain and as shock-absorber-like materials at
highstrain by effectively dissipating energy. These properties are
com-parable to the passive elastic properties of muscles within
thephysiological range of sarcomere length14 and so these
materialsrepresent a new muscle-mimetic biomaterial. The
mechanicalproperties of these biomaterials can be fine-tuned by
adjustingthe composition of the elastomeric proteins, providing the
oppor-tunity to develop biomaterials that are mimetic of different
types ofmuscles. We anticipate that these biomaterials will find
applica-tions in tissue engineering15 as scaffold and matrix for
artificialmuscles.
The string of folded immunoglobulin domains and
unstructuredunique sequences constitute two distinct types of
entropic springs intitin7,8. The string of folded immunoglobulin
domains has higher per-sistence length than the unstructured
sequences and extend first duringstretching. Only under high
stretching forces at the high end of thephysiological range of
sarcomere length, when the string of immuno-globulin domains are
straightened, can a small number of foldedimmunoglobulin domains
unfold to extend the length of titin anddissipate energy,
effectively preventing damage due to overstretching16.These
features are combined to give rise to the passive
mechanicalproperties of muscles at the macroscopic level1,2,7,17,
which manifestas a Young’s modulus close to 100 kPa, increasing
energy dissipation athigher sarcomere length, and stress relaxation
at a constant strain4,14,16.To design biomaterials mimicking the
fine-tuned passive elastic pro-perties of muscle, it is critical to
incorporate these mechanical featuresat both single-molecule and
macroscopic biomaterial levels. Towardsthis goal, here we first
engineered artificial elastomeric proteins thatmimic the molecular
architecture and nanomechanical properties ofindividual titin
molecules. Then we used these proteins to constructbiomaterials
that mimic the passive elastic properties of muscle.
To engineer titin-mimicking artificial elastomeric proteins
(seeMethods and Supplementary Fig. 1), we used the
well-characterizedGB1 domains12 to mimic folded titin
immunoglobulin domains,because GB1 domains exhibit mechanical
properties comparable tothose of titin immunoglobulin domains12,
and we used a consensusrepeat of the random-coil-like protein
resilin13 to mimic unstruc-tured sequences (such as the N2B
sequence in titin), because resilinis a highly elastic and
resilient protein13,18,19 and is also largelyunstructured20 (see
Supplementary Fig. 2). With these two buildingblocks, we
constructed artificial elastomeric proteins (G–R)4 andGRG5RG4R,
where G represents individual GB1 domain, and Rrepresents
individual resilin repeat.
We first used AFM techniques6,21 to characterize their
nanomecha-nical properties at the single-molecule level. Stretching
(G–R)4results in characteristic sawtooth-like force–extension
relationships12
(Fig. 1a), where individual force peaks correspond to the
mechanicalunfolding of GB1 domains and are characterized by an
unfoldingforce of ,180 pN and a contour length increment DLc of ,18
nm.Owing to the dimerization of (G–R)4 via carboxy-terminal
cysteineresidues (Supplementary Information), force–extension
curves canshow as many as eight GB1 unfolding events. The
featureless ‘spacer’,which is of length L0 and occurs before the
GB1 unfolding forcepeaks, corresponds to the stretching of
random-coil-like resilinsand folded GB1 domains (Fig. 1a),
confirming the entropic spring-nature of resilin repeats. Because
the persistence length of GB1 ismuch larger than that of
unstructured resilin, fitting the Worm-like-chain model of polymer
elasticity to the spacer yielded a persist-ence length of 0.49 6
0.09 nm (average 6 s.d., n 5 188) for resilin,comparable to that of
the random-coil-like sequence N2B in titin8,9,22
and unfolded polyprotein chains6,21. Stretching
polyproteinGRG5RG4R yielded force–extension curves with similar
sawtoothpatterns but with shorter spacers owing to the fewer
resilin domainsin GRG5RG4R (Fig. 1b and Supplementary Fig. 3).
Moreover, themechanical unfolding of GB1 domains is reversible,
becauseunfolded GB1 domains can refold to regain mechanical
resistanceupon relaxation12. These nanomechanical properties of
(G–R)4 andGRG5RG4R largely mimic those of individual titin
molecules.
We then used these miniature-titin-like elastomeric proteins
toconstruct biomaterials to mimic the passive mechanical
propertiesof muscle. Individual titin molecules are well-aligned
and organizedin the filament lattice of muscle3. However, it
remains challenging tomimic such ordered structures in synthetic
biomaterials. As analternative, we created chemically crosslinked
GB1–resilin networksto exploit the nanomechanical properties
engineered into individualGB1–resilin molecules. We used the
well-developed [Ru(bpy)3]
21-mediated photochemical crosslinking strategy23, which allows
thecrosslinking of two tyrosine residues in close proximity into
dityrosineadducts (Supplementary Fig. 4). This method was used
successfully to
1Department of Chemistry, University of British Columbia,
Vancouver, British Columbia V6T 1Z1, Canada. 2Department of
Zoology, University of British Columbia, Vancouver, BritishColumbia
V6T 1Z1, Canada. {Present address: Department of Engineering
Science and Mechanics, Virginia Polytechnic Institute and State
University, Blacksburg, Virginia 24061, USA.
Vol 465 | 6 May 2010 | doi:10.1038/nature09024
69Macmillan Publishers Limited. All rights reserved©2010
www.nature.com/doifinder/10.1038/nature09024www.nature.com/naturewww.nature.com/nature
-
crosslink recombinant resilins into solid biomaterials13. The
use ofresilin repeats, which provide the majority of crosslinking
sites inGB1–resilin polyproteins, enables an efficient approach
with whichto prepare GB1–resilin-based biomaterials. We found that
on illu-mination with white light, GB1–resilin polyproteins can
readily becrosslinked into solid and transparent biomaterials at
room temper-ature from their concentrated (.150 mg ml21) solutions
(Sup-plementary Information). The middle panel of Fig. 2a shows
opticalphotographs of moulded rings of both polyproteins. The
formation ofdityrosine crosslinks was indicated by their
characteristic blue fluor-escence upon ultraviolet irradiation13
(Fig. 2a, top panel).
Protein-based biomaterials, such as those based on
elastin24–28,resilin13 and abductin29, are engineered from
non-globular elasto-meric proteins that behave like entropic
springs. To our knowledge,the GB1–resilin-based biomaterial is the
first chemically crosslinkedbiomaterial that incorporates folded,
mechanically resistant globulardomains in its constituent
elastomeric proteins, enabling us to examinetheir macroscopic
mechanical properties and investigate how themicroscopic properties
of individual proteins are translated into mac-roscopic ones in
biomaterials.
Here we carried out tensile measurements to characterize the
mech-anical properties of GB1–resilin-based biomaterials in PBS at
roomtemperature. For technical reasons, we used ring-shaped samples
fortensile testing30 (see Supplementary Information). Typical
stress–strain curves of (G–R)4 and GRG5RG4R-based biomaterials are
shownin Fig. 2b-c. It is evident that GB1–resilin-based
biomaterials areelastic. GRG5RG4R can be stretched to a strain as
high as 135%without breaking. The Young’s modulus is ,70 kPa for
(G–R)4 and,50 kPa for GRG5RG4R (at 15% strain), both close to the
Young’smodulus measured for myofibrils/myocytes, which is in the
range60–100 kPa within the physiological range of sarcomere
length4,14.These biomaterials are isotropic (Supplementary
Information), sothe measured Young’s modulus reflects the overall
isotropic propertyof the biomaterials.
Resilience is a measure of a material’s ability to deform
reversiblywithout loss of energy18. Resilin is known for its superb
resilience19,and resilin-based biomaterials constructed using the
same photoche-mical crosslinking method did not show appreciable
hysteresis evenat 250% strain13,31 (Fig. 2d). To examine the
influence of folded GB1domains on the resilience of
GB1–resilin-based biomaterials, wemeasured the resilience of these
biomaterials. The stretching andrelaxation curves of both (G–R)4
and GRG5RG4R at low strain(,15%) were superimposable and no
hysteresis was observed(Fig. 2b, c), suggesting high resilience for
both materials at lowstrains. However, the stretching and
relaxation curves were no longersuperposable at higher strains and
hysteresis started to develop, indi-cating that some of the work
done during stretching was dissipatedand cannot be recovered upon
relaxation. The hysteresis increaseswith the increase of strain
(Fig. 2b, c), indicating that the resilience ofGB1–resilin-based
materials decreases with the increase of strain(Fig. 2d). This
behaviour is similar to that of myofibrils or myo-cytes14,16,32,
which showed increasing hysteresis between stretchingand relaxation
at increasing sarcomere lengths, indicating thatGB1–resilin-based
biomaterials, just like muscles16, behave likeshock-absorbers at
higher strains by effectively dissipating energy.
The observed hysteresis, that is, energy dissipation, during
cyclicexperiments indicates that stretching GB1–resilin-based
biomaterialsto higher strains involved the breakage of weak
non-covalent bonds33 inthe crosslinked network. And the breaking of
such bonds is reversible,as the hysteresis observed in
GB1–resilin-based biomaterials can befully recovered upon
relaxation. As shown in the insets of Fig. 2b andc, during
subsequent stretching, stress–strain curves superpose on oneanother
regardless of the final strain, suggesting a full recovery of
thehysteresis. Moreover, the recovery of hysteresis is very fast:
during cyclicstretching–relaxation experiments, the
stretching–relaxation loopswere identical (Supplementary Fig. 6)
even when there was no waitingtime between consecutive cycles,
suggesting that the recovery of hys-teresis occurs at a timescale
significantly shorter than the dead time ofour Instron, which is
estimated to be ,1 s. Again, this reversible hys-teresis behaviour
is similar to that for myofibrils and myocytes16,32. It isalso
interesting to note that the recovery of hysteresis can occur
evenunder residual stress. In partial relaxation experiments (Fig.
2e), whenthe biomaterial was partially relaxed to a strain above
35%, no recoveryof hysteresis was observed. When the biomaterial
was relaxed to below35% strain, partial recovery started to occur.
The degree of recoverydepends on the residual stress: the lower the
residual stress, the higherthe percentage of recovery (Fig.
2e).
It is evident that the mechanical behaviours of
GB1–resilin-basedbiomaterials differ significantly from those of
resilin-based bioma-terials, highlighting the significant roles of
folded GB1 domains indetermining the mechanical properties of the
resultant biomaterials.Given that photochemically crosslinked
resilin-based biomaterialsshow only negligible hysteresis during
stretching (Fig. 2d), theobserved hysteresis in GB1–resilin-based
biomaterials probablyresulted from folded GB1 domains and the
associated structuralchanges of the crosslinked network. The
hysteresis observed in tensileexperiments indicates that the
stretching of GB1–resilin-based bio-materials involved breaking of
weak non-covalent bonds33. It is wellknown from single-molecule AFM
experiments that, on stretching,force-induced rupture of
non-covalent bonds can lead to the unfold-ing of GB1 domains and
dissipation of energy12. Therefore, theunfolding of some GB1
domains during stretching could provide aplausible molecular
mechanism to explain the hysteresis observed inGB1–resilin-based
biomaterials at high strains. The fast recovery rateof hysteresis
and the ability to recover hysteresis under residual stressare
consistent with the fast folding kinetics of GB1 domains and
theability of GB1 domains to refold under residual force observed
insingle-molecule AFM experiments12, providing qualitative
evidencethat the hysteresis observed in biomaterials probably
originates fromthe unfolding of a small number of GB1 domains.
Extension (nm)
Forc
e (p
N)
L0
0 50 100 200 250150
ΔLc = 18.3 nm
200
pN
GB1
GB1 GB1R
R R R R
R RGB1 GB1 GB1 GB1 GB1 GB1 GB1 GB1
GB1 GB1 GB1a
b
Figure 1 | Force–extension curves of two polyproteins. a,
(G–R)4.b, GRG5RG4R. The force peaks, characterized by a DLc of ,18
nm and anunfolding force of ,180 pN, result from the mechanical
unfolding of GB1domains. Stretching resilins does not result in any
unfolding force peaks;instead we see a featureless spacer of length
L0. The notable difference betweenthe force–extension curves of
(G–R)4 and GRG5RG4R is the shorter featurelessspacer of GRG5RG4R,
which is due to fewer resilin repeats in GRG5RG4R. Greylines
correspond to the worm-like chain model fits to the experimental
data.
LETTERS NATURE | Vol 465 | 6 May 2010
70Macmillan Publishers Limited. All rights reserved©2010
-
To further compare the energy dissipation behaviours of
thedesigned biomaterials with those of myofibrils/myocytes, we
carriedout stress-relaxation experiments at constant strains.
WhenGRG5RG4R was stretched rapidly to a given strain that was
heldconstant afterwards, we observed clear stress relaxation (Fig.
3a),again suggestive of the existence of energy dissipation
processes.The larger the initial strain, the greater the amplitude
of stress relaxa-tion. We found that the stress-relaxation
behaviours can be describedreasonably well by double-exponential
fits. The relaxation rates (k1and k2) were observed to increase
with the increase of strain, but theincrease of fast-phase rate k1
mainly occurs at higher strain while theincrease of slow-phase rate
k2 occurs at lower strain (Fig. 3b).
The stress-relaxation behaviours of GB1–resilin-based
biomaterialsare qualitatively similar to those of myofibrils16,17,
but they are verydifferent from the behaviour of biomaterials made
of resilin, in whichnegligible stress-relaxation was observed in
similar experiments31. Theunfolding of a few immunoglobulin domains
was proposed as apossible molecular mechanism to explain the
stress-relaxation beha-viours of myofibrils31. Similarly, Monte
Carlo simulations6 on force-relaxation behaviour of GRG5RG4R at
constant extension revealedthat the unfolding of some GB1 domains
can lead to force-relaxationbehaviours similar to those seen in our
experiments and similardependence of the fast-phase relaxation rate
on extension(Supplementary Fig. 7). However, the simulated
behaviour of theslow-phase relaxation rate differed from the
experimental data. It isclear that the relaxation behaviours
simulated at the single-molecule
level cannot be directly compared with the stress-relaxation
beha-viours of GB1–resilin-based biomaterials quantitatively,
becauseGB1–resilin molecules are not well-aligned in the
photochemicallycrosslinked three-dimensional network and the force
experienced byindividual molecules cannot be measured directly. A
more detailedmodel combining the possibility of GB1 unfolding with
a three-dimensional network is required to describe the
stress-relaxationbehaviour at the macroscopic level. Moreover, it
is important to notethat stress relaxation is considered to be a
viscoelastic property macro-scopically. Although domain unfolding
can lead to stress relaxation,the direct demonstration of domain
unfolding in macroscopicmaterials is yet to be achieved. Therefore,
it is possible that othermicroscopic processes or mechanisms, such
as friction experiencedby folded domains during stretching, may
also contribute to thestress-relaxation behaviours of muscles as
well as GB1–resilin-basedbiomaterials.
Our results demonstrate that the incorporation of folded,
mech-anically resistant globular domains into elastomeric proteins
pro-vides a novel approach with which to construct biomaterials
thathave unusual macroscopic mechanical properties. Such a
bottom-up approach offers the opportunity to tailor the macroscopic
pro-perties of biomaterials by fine-tuning the nanomechanical
propertiesof their molecular building blocks at the single-molecule
level. Todemonstrate such possibilities, we used chemical
denaturant to affectthe nanomechanical properties of individual GB1
domains in orderto modulate the mechanical properties of
macroscopic biomaterials.
Str
ess
(kP
a)
Str
ess
(kP
a)
Str
ess
(kP
a)
Strain (%)
Strain (%)
Str
ess
(kP
a)
Strain (%)
Strain (%)
a b c
d e
30
25
20
15
100
95
90
40
20
0
0 40 6020
85
80
75
70
65
20 40 60 80 100 120
5
0 0
10
10 20
20
30
40
40
50
60
0 10
10
20
20
30
0 0 20 40 60 80 100 120 14010 20 30
0 20 40 60
40 6050
10%
10%
20%
20%
30%
30%
40%
50%
60%
70%
35%
Res
ilien
ce (%
)
Str
ess
(kP
a)
Strain (%)Strain (%)
(G–R)4GRG5RG4RGRG5R
G8GRG9R
Resilin
Time
Str
ain
Figure 2 | Mechanical properties of (G–R)4 and
GRG5RG4R-basedbiomaterials. a, Photographs of moulded rings built
from (G–R)4 (left,intact) and GRG5RG4R (right, after being loaded
to failure in tensile test)under white light (middle panel) and
ultraviolet illumination (top panel).b, c, Representative
stress–strain curves of (G–R)4 (b) and GRG5RG4R(c) measured in PBS.
For clarity, stress–strain curves are offset relative toone
another. Final strains are shown on the curves. Insets show
thesuperposition of the stress–strain curves at different strains.
d, Resilience ofGB1–resilin-based biomaterials decreases with the
increase of strain. Incontrast, biomaterials constructed from
resilin do not show any appreciable
hysteresis (data taken from ref. 13). e, GRG5RG4R-based
biomaterials canrecover hysteresis under residual stress. During
stretching–relaxationexperiments, when the biomaterial is partially
relaxed to a strain above 35%,no recovery of hysteresis was
observed. When the biomaterial was relaxed tobelow 35% strain, we
started to observe partial recovery. The degree ofrecovery
increased with the decrease of residual stress. For clarity, the
initialstretching trace is coloured blue. The inset shows the
experimental protocolof the partial relaxation experiments. The
pulling speed used in theexperiments was 25 mm min21. Error bars
indicate standard deviation of thedata.
NATURE | Vol 465 | 6 May 2010 LETTERS
71Macmillan Publishers Limited. All rights reserved©2010
-
Folded globular domains are mechanically more resistant than
theirunfolded conformations, but less extensible. Because chemical
dena-turants can affect folded states of globular proteins, we used
urea tomodulate the nanomechanical properties of
GB1–resilin-basedelastomeric proteins. Figure 4a shows such an
example forGRG5RG4R. In the presence of 4 M urea, about half of GB1
domainsare unfolded, resulting in the loss of their mechanical
resistance. Sucha change is clearly evident in the force–extension
relationships ofGRG5RG4R (Fig. 4a), which are characterized by long
featurelessspacers before the unfolding events of the remaining
folded GB1domains. Such long featureless spacers correspond to the
stretchingof predominantly unfolded GB1 domains. The conversion of
foldedGB1 into unfolded sequences leads to a dramatic decrease in
Young’smodulus of the biomaterials in a
urea-concentration-dependentfashion: the Young’s modulus reduced
from ,60 kPa in PBS to,10 kPa in 8 M urea. We note that this change
is fully reversible atboth molecular and macroscopic levels.
Replacing urea with PBSallowed GB1 domains to refold and thereby
regain their mechanicalresistance. Macroscopically, the biomaterial
can recover its originalYoung’s modulus when replacing urea with
PBS. This macroscopicchange in mechanical properties of
biomaterials can readily beexplained with information from the
single-molecule level: the con-version of folded GB1 domains into
mechanically labile and moreextensible sequences effectively
increased the length between cross-linking points, leading to the
decrease in Young’s modulus of thematerial. Similarly, it is also
possible to modulate the mechanicalproperties of these biomaterials
(Supplementary Fig. 8) in otherways, such as adjusting the relative
GB1/resilin content, just as thepassive elastic properties of
different muscles are mediated by differ-ent isoforms of titin.
To fulfill their biological functions, different biological
tissuespossess distinct mechanical properties. For example,
mammalian ten-don is highly resilient (resilience .90%) but
relatively inextensible(breaking strain of ,13%), whereas elastin
is resilient (90%) andextensible (breaking strain of ,150%) but
lacks toughness18.Mimicking the biomechanical properties of
different tissues has beenan important challenge in biomaterials
research. Here we havedesigned a muscle-mimetic biomaterial, which
is highly resilient atlow strains, but also extensible and tough at
high strain, to mimic thepassive elastic properties of muscles.
Titin is largely responsible for thepassive elastic properties of
myofibrils. A hallmark of titin-like elasto-meric proteins is their
ability to unfold under a stretching force todissipate energy
effectively and prevent damage to tissues by
over-stretching6,8,10,11,16. The hysteresis and stress-relaxation
observed instretching of myofibrils have been explained by
force-induced unfold-ing of a small number of immunoglobulin
domains16.
All these properties have been well reproduced in
biomaterialsconstructed from GB1–resilin-based artificial
elastomeric proteins.Therefore, GB1–resilin-based polyproteins
mimic the architectureand mechanical properties of titin at the
single-molecule level, andbiomaterials based on GB1–resilin
polyproteins mimic the titin-mediated passive elastic properties of
muscles (Supplementary
100
80
60
40
20
00
0 20 40 60 80
20 40 60 80 100 120
Str
ess
(kP
a)
Time (s)
80%70%60%50%40%30%20%10%
0.2
0.02
0.04
0.06
0.08
0.10
0.3
0.4
0.5
0.1
Strain (%)
Rel
axat
ion
rate
, k1
(s–1
)
Rel
axat
ion
rate
, k2
(s–1
)
a
b
Strain
Figure 3 | GB1–resilin-based biomaterials exhibit pronounced
stressrelaxation behaviours. a, Representative stress-relaxation
curves ofGRG5RG4R at varying strains. b, Relaxation rates of
GRG5RG4R-basedbiomaterials depend upon the initial stress. The
relaxation rates wereobtained by fitting the stress-relaxation to a
double-exponential equation:s(t) 5 s0 1 A1exp(2k1t) 1 A2exp(2k2t),
where s(t) is the stress at time t, s0is the offset, A1 and A2 are
decay amplitudes and k1 (filled squares) and k2(open triangles) are
relaxation rates. Error bars indicate fitting errors.
Forc
e (p
N)
Youn
g’s
mod
ulus
(kP
a)
0
0 2 4 6 8
50 100 150 250200Extension (nm)
60
50
40
30
20
Urea concentration (M)
200
pN
4 M urea
PBSa
b
Figure 4 | The macroscopic mechanical properties of
GB1–resilin-basedbiomaterials can be fine-tuned by controlling the
nanomechanicalproperties of the constituting elastomeric proteins
at the single-moleculelevel. a, Force–extension curves of single
GRG5RG4R molecules in PBS andin 4 M urea. The long featureless
spacers observed in force-extension curvesof GRG5RG4R in 4 M urea
largely correspond to the stretching ofmechanically labile,
unfolded GB1 domains. The unfolding force of GB1domains that remain
folded in 4 M urea is also significantly reduced. Greylines are WLC
fits. b, Young’s modulus of GRG5RG4R-based biomaterial canbe
modulated by chemical denaturant urea. The conversion of folded
GB1domains into unfolded sequence leads to the dramatic decrease in
Young’smodulus of the biomaterials in a
urea-concentration-dependent manner.Error bars indicate standard
deviation of the data.
LETTERS NATURE | Vol 465 | 6 May 2010
72Macmillan Publishers Limited. All rights reserved©2010
-
Information). These designed biomaterials represent a new type
ofmuscle-mimic, which is fully hydrated and biodegradable, and
weanticipate that they will find applications in material sciences
as wellas in tissue engineering by serving as scaffold and matrix
for artificialmuscles. Moreover, our results indicate that
nanomechanical pro-perties engineered into individual polyproteins
can be translated intomacroscopic properties in materials, a new
example of obtainingnovel macroscopic mechanical features by
designing in such featuresat the single-molecule level. This method
represents a new avenuetowards tailoring the macroscopic mechanical
properties of bio-materials and can be applied to the design of a
wide range of materials.
METHODS SUMMARYPreparation of GB1–resilin-based polyproteins
were performed using previously
published protocols12,21. We used the 15-amino-acid consensus
resilin repetitive
sequence (GGRPSDSYGAPGGGN) from the first exon of the Drosophila
mela-
nogaster CG15920 gene to construct GB1–resilin-based elastomeric
proteins13.
Single-molecule AFM experiments were performed on a
custom-designed
atomic force microscope as described12. Hydrogel-like
biomaterials of GB1–
resilin was constructed using a photochemical crosslinking
strategy as
described13,23. Tensile tests were performed on an Instron-5500R
tensometer
with a custom-made force gauge in PBS at constant temperature
(22 uC). Fortechnical reasons, ring-shaped biomaterial specimens
were used30 (Supplemen-tary Information). Resilience was calculated
from the ratio of the area under the
relaxation curve to the area under the extension curve at a
given strain using
custom-written software in Matlab. The local slope at 15% strain
on the exten-
sion curve was taken as the modulus at 15% reported in the
paper.
Received 22 September 2009; accepted 9 March 2010.
1. Maruyama, K. Connectin/titin, giant elastic protein of
muscle. FASEB J. 11,341–345 (1997).
2. Wang, K. Titin/connectin and nebulin: giant protein rulers of
muscle structure andfunction. Adv. Biophys. 33, 123–134 (1996).
3. Tskhovrebova, L. & Trinick, J. Titin: properties and
family relationships. NatureRev. Mol. Cell Biol. 4, 679–689
(2003).
4. Granzier, H. L. & Irving, T. C. Passive tension in
cardiac muscle: contribution ofcollagen, titin, microtubules, and
intermediate filaments. Biophys. J. 68,1027–1044 (1995).
5. Labeit, S. & Kolmerer, B. Titins: giant proteins in
charge of muscle ultrastructureand elasticity. Science 270, 293–296
(1995).
6. Rief, M., Gautel, M., Oesterhelt, F., Fernandez, J. M. &
Gaub, H. E. Reversibleunfolding of individual titin immunoglobulin
domains by AFM. Science 276,1109–1112 (1997).
7. Linke, W. A. et al. I-band titin in cardiac muscle is a
three-element molecularspring and is critical for maintaining thin
filament structure. J. Cell Biol. 146,631–644 (1999).
8. Li, H. et al. Reverse engineering of the giant muscle protein
titin. Nature 418,998–1002 (2002).
9. Watanabe, K. et al. Molecular mechanics of cardiac titin’s
PEVK and N2B springelements. J. Biol. Chem. 277, 11549–11558
(2002).
10. Kellermayer, M. S., Smith, S. B., Granzier, H. L. &
Bustamante, C. Folding-unfoldingtransitions in single titin
molecules characterized with laser tweezers. Science276, 1112–1116
(1997).
11. Tskhovrebova, L., Trinick, J., Sleep, J. A. & Simmons,
R. M. Elasticity and unfoldingof single molecules of the giant
muscle protein titin. Nature 387, 308–312 (1997).
12. Cao, Y. & Li, H. Polyprotein of GB1 is an ideal
artificial elastomeric protein. NatureMater. 6, 109–114 (2007).
13. Elvin, C. M. et al. Synthesis and properties of crosslinked
recombinant pro-resilin.Nature 437, 999–1002 (2005).
14. Linke, W. A., Popov, V. I. & Pollack, G. H. Passive and
active tension in singlecardiac myofibrils. Biophys. J. 67, 782–792
(1994).
15. Langer, R. & Tirrell, D. A. Designing materials for
biology and medicine. Nature428, 487–492 (2004).
16. Minajeva, A., Kulke, M., Fernandez, J. M. & Linke, W. A.
Unfolding of titin domainsexplains the viscoelastic behavior of
skeletal myofibrils. Biophys. J. 80, 1442–1451(2001).
17. Higuchi, H. Viscoelasticity and function of connectin/titin
filaments in skinnedmuscle fibers. Adv. Biophys. 33, 159–171
(1996).
18. Gosline, J. et al. Elastic proteins: biological roles and
mechanical properties. Phil.Trans. R. Soc. Lond. B 357, 121–132
(2002).
19. Andersen, S. O. The cross-links in resilin identified as
dityrosine and trityrosine.Biochim. Biophys. Acta 93, 213–215
(1964).
20. Nairn, K. M. et al. A synthetic resilin is largely
unstructured. Biophys. J. 95,3358–3365 (2008).
21. Carrion-Vazquez, M. et al. Mechanical and chemical unfolding
of a single protein:a comparison. Proc. Natl Acad. Sci. USA 96,
3694–3699 (1999).
22. Li, H. et al. Multiple conformations of PEVK proteins
detected by single-moleculetechniques. Proc. Natl Acad. Sci. USA
98, 10682–10686 (2001).
23. Fancy, D. A. & Kodadek, T. Chemistry for the analysis of
protein-proteininteractions: rapid and efficient cross-linking
triggered by long wavelength light.Proc. Natl Acad. Sci. USA 96,
6020–6024 (1999).
24. Urry, D. W. et al. Elastin: a representative ideal protein
elastomer. Phil. Trans. R.Soc. Lond. B 357, 169–184 (2002).
25. Vrhovski, B., Jensen, S. & Weiss, A. S. Coacervation
characteristics ofrecombinant human tropoelastin. Eur. J. Biochem.
250, 92–98 (1997).
26. Bellingham, C. M. et al. Recombinant human elastin
polypeptides self-assembleinto biomaterials with elastin-like
properties. Biopolymers 70, 445–455 (2003).
27. Welsh, E. R. & Tirrell, D. A. Engineering the
extracellular matrix: a novel approachto polymeric biomaterials. I.
Control of the physical properties of artificial proteinmatrices
designed to support adhesion of vascular endothelial
cells.Biomacromolecules 1, 23–30 (2000).
28. Nagapudi, K. et al. Protein-based thermoplastic elastomers.
Macromolecules 38,345–354 (2005).
29. Bochicchio, B. et al. Synthesis of and structural studies on
repeating sequences ofabductin. Macromol. Biosci. 5, 502–511
(2005).
30. Lillie, M. A., Chalmers, G. W. & Gosline, J. M. The
effects of heating on themechanical properties of arterial elastin.
Connect. Tissue Res. 31, 23–35 (1994).
31. Dudek, D. M. et al. Dynamic Mechanical Properties of
Synthetic Resilin. in AnnualMeeting of the Society for Integrative
and Comparative Biology abstr. S7.7,
e50,http://icb.oxfordjournals.org/cgi/reprint/49/suppl_1/e1.
(OxfordUniversity Press, 2009).
32. Helmes, M. et al. Mechanically driven contour-length
adjustment in rat cardiactitin’s unique N2B sequence: titin is an
adjustable spring. Circ. Res. 84, 1339–1352(1999).
33. Amin, A. F. M. S., Lion, A., Sekita, S. & Okui, Y.
Nonlinear dependence of viscosityin modeling the rate-dependent
response of natural and high damping rubbers incompression and
shear: experimental identification and numerical verification.Int.
J. Plast. 22, 1610–1657 (2006).
Supplementary Information is linked to the online version of the
paper atwww.nature.com/nature.
Acknowledgements We thank M. Lillie and R. Shadwick for
discussions. This workis supported by the Canadian Institutes of
Health Research, the Canada ResearchChairs program, the Canada
Foundation for Innovation, the Michael SmithFoundation for Health
Research, and the Natural Sciences and EngineeringResearch Council
of Canada. H.L. is a Michael Smith Foundation for HealthResearch
Career Investigator.
Author Contributions H.L. conceived the project. H.L. and J.G.
designed the overallexperiments. S.L., D.M.D., Y.C., M.M.B. and
J.G. designed, performed individualexperiments and analysed data.
H.L. wrote the manuscript and all authors editedthe manuscript.
Author Information Reprints and permissions information is
available atwww.nature.com/reprints. The authors declare no
competing financial interests.Correspondence and requests for
materials should be addressed to H.L.([email protected]) or J.G.
([email protected]).
NATURE | Vol 465 | 6 May 2010 LETTERS
73Macmillan Publishers Limited. All rights reserved©2010
http://icb.oxfordjournals.org/cgi/reprint/49/suppl_1/e1www.nature.com/naturewww.nature.com/reprintsmailto:[email protected]:[email protected]
-
SUPPLEMENTARY INFORMATION
1www.nature.com/nature
doi: 10.1038/nature09024
1
Supplementary Information Protein Engineering GB1-resilin-based
polyprotein genes were constructed using standard molecular biology
techniques following a well-established stepwise construction
scheme1. The gene encoding protein GB1 was a generous gift from
David Baker of University of Washington. One 15 amino acid
consensus resilin repetitive sequence (GGRPSDSYGAPGGGN) from the
first exon of the D. melanogaster gene CG15920 (Ref. 2) was used in
this study to construct GB1-resilin based polyproteins. The DNA
sequence of resilin, flanked with a 5’ BamHI restriction site and
3’ BglII and KpnI restriction sites, was synthesized by PCR
(polymerase chain reaction) based oligonucleotide assembly. The
expression vector of pQE80L-(GR)4 was constructed by iterative
cloning of G and R genes into empty pQE80L vector, on the basis of
the identity of the sticky ends generated by BamHI and BglII
restriction enzymes. GRG5RG4R, GRG5R, GRG9R and G8 were constructed
in the same way. The resulted polyproteins carry two additional
cysteines at their C-termini, which can lead to the dimerization of
some polyproteins. The expression of polyproteins was carried out
in Escherichia coli strain DH5α. Cultures were grown at 37 °C in
2.5% LB containing 100mg/L ampicillin, and induced with 0.8mM
isopropyl-1-β-D-thiogalactoside (IPTG) when the optical density
reached ~1. Protein expression continued for 5 hours. The cells
were harvested by centrifugation at 15,000g for 15min and cell
lysis was done using lysozyme from egg white (100 mM,
SigmaAldrich). The soluble fraction was purified using Ni2+
affinity chromatography. The yield of the polyproteins is in the
range of 40mg to 50mg per liter of culture. Supplementary Fig. 1
shows Coomassie blue stained SDS-PAGE picture for the constructed
proteins (G-R)4, GRG5RG4R, GRG5R and GRG9R. The purity of the
purified polyproteins is around 90%, as estimated from the SDS-PAGE
using AlphaEaseFC software (Version 4.0.0, Alpha Innotech
Corporation, San Leandro, CA). The 10% “impurity” is likely to be
truncated fragments of polyproteins, which are frequently observed
in the expression of polyproteins, such as G8 as well as other
polyproteins.
Figure S1. Coomassie blue stained PAGE gel for polyproteins
(GR)4, GRG5RG4R, GRG5R, GRG9R and G8 (left to right). The first
lane is the broad range molecular weight marker (New England
Biolabs).
-
2www.nature.com/nature
doi: 10.1038/nature09024 SUPPLEMENTARY INFORMATION
2
The purified proteins were then used in single-molecule atomic
force microscopy (AFM) experiments. To prepare GB1-resilin-based
biomaterials, the purified proteins were then dialyzed against
deionized water for 3 days to remove all the salt from elution
buffer. During dialysis, the water was changed every 5 hours. The
protein was lyophilized following dialysis.
-8000
-6000
-4000
-2000
0
260250240230220210200
Wavelength (nm)
[θ]M
RE(deg•cm
2 •dm
ol-1
)
Figure S2. Far ultraviolet circular dichroism (CD) spectrum of
R12 indicates that R12 is largely unstructured. To confirm that the
resilin sequence used in our study is largely unstructured,
polyprotein R12, which is composed of 12 identical tandem repeats
of the 15 amino acid consensus sequence (GGRPSDSYGAPGGGN), was
constructed following the strategy described for the engineering of
(G-R)4. There is a two-amino acid linker Arg-Ser between R repeats
in the engineered R12, which results from the restriction sites
used in the construction of the gene of R12. The CD measurements
were carried out on a Jasco-J810 spectropolarimeter. Our CD data
showed a minimum in ellipticity at ~200 nm, indicating that the R12
is largely unstructured. Our result contradicts the study by Lyons
and et al3. They showed that the CD spectrum of R16, which is
composed of 16 identical repeats of the same resilin consensus
sequence, is consistent with the existence of β-strand structure3.
The origin of the discrepancy between these two studies is unknown.
One noted difference between R12 and R16 is that there exists an
Arg-Ser linker between R repeats in R12 while there is none in R16.
Despite the discrepancy on the secondary structure of this resilin
repeat sequence, single-molecule AFM results showed that resilin
behaves largely as an entropic spring, and does not show measurable
mechanical stability (see Fig. 2 in the main text). This result is
consistent with our observation that R is largely unstructured.
Single-Molecule AFM Measurements
-
3www.nature.com/nature
SUPPLEMENTARY INFORMATIONdoi: 10.1038/nature09024
3
Single molecule AFM experiments were performed on a
custom-designed AFM following procedures described previously4. All
the force-extension measurements were carried out in PBS, unless
noted otherwise. In a typical experiment, ~1 μl polyprotein sample
was deposited onto a clean glass cover slip covered by PBS buffer
(50 mL) and was allowed to adsorb for approximately 5 min before
force-extension measurements. The spring constant of each
individual cantilever (Si3N4 cantilevers from Vecco, with a typical
spring constant of 40 pN/nm) was calibrated in solution using the
Equipartition Theorem before and after each experiment.
Figure S3. Mechanical properties of (G-R)4 and GRG5RG4R at the
single-molecule level. A, B) Histogram of spacer length L0 of
(G-R)4 and GRG5RG4R. For force-extension curves with similar number
of GB1 domains (six or higher) for both polyproteins, the spacer
length of (G-R)4 is longer than that for GRG5RG4R, due to the more
resilin sequences presented in (G-R)4. C,D) Persistence length of
resilins measured in both polyproteins yield similar values of 0.49
nm. E, F) Unfolding force histogram of GB1 domains in both
proteins. The unfolding force for GB1 is 191±42 pN in (G-R)4 and
180±41 pN in GRG5RG4R
Preparation of Biomaterials We used a well-developed
[Ru(bpy)3]2+-mediated photochemical crosslinking strategy5 to
prepare GB1-resilin-based biomaterials. This photochemical strategy
allows the crosslinking of two tyrosine residues that are in close
proximity into a dityrosine adduct, and leads to rapid and
quantitative formation of dityrosine crosslinks between soluble
proteins. Supplementary Fig. S4 shows the photocrosslinking scheme
and the schematic structure of the resultant biomaterials. It is of
note that the majority of crosslinking sites are located in the
resilin sequences, and a small fraction of the crosslinking sites
originate from the exposed tyrosine residues in GB1. To prepare
GB1-resilin-based biomaterials, lyophilized proteins were
redissolved in phosphate saline buffer (PBS). In a typical
experiment, 18 mg of the protein was weighted using analytical
balance and added to a microcentrifuge tube containing 84.4μl
-
4www.nature.com/nature
doi: 10.1038/nature09024 SUPPLEMENTARY INFORMATION
4
of PBS (100mM, pH 7.4) and 4.5μl of APS (1M). The trapped air
bubbles can be removed by centrifugation at 1000 g for 5 minutes.
0.9μl of [Ru(bpy)3]2+ (20mM) solution was then added to the
microcentrifuge tube and quickly mixed with the protein solution by
tapping the bottom of the tube. The final solution contains
200mg/ml of GB1-resilin polyprotein, ~200μM [Ru(bpy)3]2+ and 50 mM
ammonium persulfate (APS) in PBS buffer. The solution was cast into
a custom-made plexiglass mold with inner diameter of 8 mm, outer
diameter of 10 mm and height of 3mm. The sample was then irradiated
for 30 seconds using a 200 W fiber optical white light source. The
irradiation was 10 cm away from the mold. The ring was then taken
out from the mold and stored in PBS buffer (10mM with 0.05% (w/v)
sodium azide). These chemically crosslinked biomaterials show
superior long term stability, and no noticeable erosion was
observed in PBS buffer (in the presence of 0.5% azide) over a
period of one year.
Figure S4. Photocrosslinking scheme and the schematic structure
of hydrogel based on GRG5RG4R. The mini-titin-mimic molecules are
mainly crosslinked via tyrosine residues in the R sequences. Some
of the solvent exposed tyrosine in the folded GB1 domains may also
contribute to the photocrosslinking, as shown in the schematic.
The resultant rings show different degrees of swelling in PBS
buffer, resulting in different initial dimensions of the rings
(Figure S5, also see Fig. 2a in the main text). The swelling ratio
of the rings was found to be consistent within the same batch, but
varies across different batches. The Poisson ratio of these
biomaterials was measured to be ~0.5. To determine the isotropy of
the constructed biomaterials, we used polarized light microscopy to
measure the birefringence of the biomaterials. Cubic samples of the
biomaterials (5mm×5mm×1mm, width×length×height) were viewed between
crossed polarizers on a Leitz Orthoplan polarizing microscope with
a 10× achromatic-pol lens, and the retardation was determined at
546 nm using a Leitz Senarmont compensator and a 546 nm
interference filter. The rotational angle between the maximum
brightness and darkness was determined and the retardation caused
by the sample is calculated as 3.03 nm per degree of the analyzer
angle. Birefringence was then calculated by dividing the
retardation of the cube by its thickness. The smallest
birefringence we can resolve is ~3×10-6. Some areas of
GB1-resilin-biomaterials showed unmeasurable birefringence,
-
5www.nature.com/nature
SUPPLEMENTARY INFORMATIONdoi: 10.1038/nature09024
5
while other areas showed maximal birefringence of 5×10-6 to
1×10-5. This birefringence is about 20 to 50 times smaller than
that of optically isotropic elastin samples (2×10-4)6, and
vanishingly smaller than that of anisotropic major ampullate silks
from A.diadematus (birefringence is 2.5×10-2 in the dry state and
6.1×10-3 in the hydrated, supercontracted state)7. This result
strongly indicates that GB1-resilin-based biomaterials are
optically isotropic.
Figure S5. Swelling ratio of different types of
GB1-resiline-based crosslinked biomaterials samples in PBS (10mM).
Error bar indicates standard deviation.
Tensile Testing The tensile tests were performed using an
Instron-5500R tensometer with a custom-made force gage as described
elsewhere.8 Unless otherwise noted, these tests were done at a
constant temperature of 22 oC in PBS buffer (10mM). The fastest
strain rate of this tensometer is 20 mm/s.
For technical consideration, the tensile testing did not follow
an ASTM (American Society for Testing and Materials) standard
through the use of dogbone-shaped specimens. Instead, we used
ring-shaped specimens. Tensile testing of rings of materials was
conducted to minimize difficulties that arise from gripping soft
materials. Because the test strains are large in these experiments,
gripped material would thin substantially upon stretching, so the
material would need to be clamped so tightly that it would fail at
the grips. Self adjusting pneumatic grips that automatically adjust
for material thinning are designed for materials much stiffer than
our hydrogel and would have the same problem of material failure or
slippage. We followed previously published methods for testing
arterial elastin rings9 to avoid these problems. An additional
benefit of the ring method is that we can likely perform tests at
significantly greater strains using rings. Lille found testing had
to be limited to 40% strain when testing dogbone shaped purified
elastin strips (M. Lillie, personal communication). Non-Gaussian
behavior of this material does not begin until at least 60% strain
and breaking strain is over 100%. Therefore, approved dogbone
shapes could actually limit our ability to adequately describe the
mechanical properties of this material.
-
6www.nature.com/nature
doi: 10.1038/nature09024 SUPPLEMENTARY INFORMATION
6
Figure S6. Consecutive stretching-relaxation curves of GRG5RG4R
at a pulling speed of 200 mm/min. The waiting time between
consecutive cycles was zero. These curves are superimposable,
indicating that GB1-resilin-based biomaterials can fully recover
their hysteresis. This behavior is consistent with the fast folding
rate of GB1 domains measured in single-molecule AFM
experiments4.
Monte Carlo simulations on the force-relaxation of GRG5RG4R
Monte Carlo simulations on the force-relaxation of a single
GRG5RG4R protein were carried out according to published
protocols10. (GRG5RG4R)6 was used to mimic the polyprotein in a
three dimensional network. During the simulations, the polyprotein
was quickly stretched to a given extension, and then the stretching
force was monitored as a function of time while the extension was
kept constant. The unfolding of GB1 domains was described using the
Bell-Evans model11, 12, )/exp()( 0 TkxFF BuΔ⋅= αα , where α(F) is
the unfolding rate constant at a stretching force F, α0 is the
intrinsic unfolding rate constant at zero force, Δxu is the
unfolding distance and kBT is the thermal energy. The
force-extension relationship of polyproteins was described using
the WLC model of polymer elasticity. The persistence length of
resilin sequence was taken as 0.5 nm, the same as that of the
unfolded polypeptide chain, and the persistence length for the
folded GB1 domains were taken as 10 nm. Using an unfolding rate
constant of 1x10-4 s-1, an unfolding distance of 0.2 nm and a
folding rate constant of 300 s-1, we simulated the force relaxation
of GRG5RG4R at constant extensions and measured the
extension-dependence of the relaxation rate (Fig. S7). It is clear
that the relaxation rate observed in the force-relaxation of
GRG5RG4R polyproteins showed similar extension-dependent behaviors
and the force-relaxation can be well described using double
exponentials. The fast-phase relaxation rate was observed to
increase with the increase of strain, suggesting that the fast
relaxation process in the stress-relaxation behaviors of
GB1-resilin-based biomaterials is qualitatively consistent with the
forced-unfolding of a few GB1 domains. However, the
extension-dependent behaviors of the slow-phase relaxation rate are
different from experimental data. These results suggest that other
microscopic processes, which were not accounted for in Monte Carlo
simulations, may also contribute to the stress-relaxation behaviors
of GB1-resilin-based biomaterials. Moreover, the three
-
7www.nature.com/nature
SUPPLEMENTARY INFORMATIONdoi: 10.1038/nature09024
7
dimensional networks as well as the random orientation of
GB1-resilin-based elastomeric chains in the GB1-resilin-based
biomaterials are not considered in Monte Carlo simulations. A more
detailed model that combines the unfolding of individual GB1
domains, the three dimensional network architecture and possibly
chain friction during stretching is required to adequately describe
the stress-relaxation behavior at the macroscopic level.
Figure S7. Monte Carlo simulation on the force-relaxation
behaviors of GRG5RG4R under constant extensions. A)
Force-relaxation behavior of GRG5RG4R under constant extension. The
force-relaxation behavior of GRG5RG4R is resulted from the forced
unfolding of GB1 domains in the polyprotein, and the non-single
exponential relaxation kinetics is due to the constant change of
the stretching force during the experiment at a constant extension.
B) Relaxation rate constant increases as a function of the initial
force. The initial force is determined by the constant extension
applied to the polyprotein during force-relaxation experiments. The
bars indicate standard deviation of the fitted rate constants.
-
8www.nature.com/nature
doi: 10.1038/nature09024 SUPPLEMENTARY INFORMATION
8
Mimicking the passive elasticity of muscle in the full range of
sarcomere length Titin is largely responsible for the passive
elasticity of muscles. However, at
longer sarcomere length, which is often beyond the physiological
range of sarcomere length, collagen molecules are recruited and
contribute to the passive elasticity of muscles. The
GB1-resilin-based biomaterials we designed largely mimic the
passive elastic properties of muscles within the physiological
range of sarcomere length. However, the lack of collagen-like
molecules as well as structural organization of titin-mimetic
proteins in the designed biomaterials make it impossible to mimic
the passive mechanical properties of muscle in the full range of
sarcomere length.
Figure S8. Mechanical properties of GB1-resilin-based
biomaterials can be modulated by adjusting the content of GB1 and
resilin. Experimental data is presented as average±standard
deviation. To demonstrate such
-
9www.nature.com/nature
SUPPLEMENTARY INFORMATIONdoi: 10.1038/nature09024
9
feasibility, we constructed elastomeric proteins GRG5R, GRG9R
and G8, and their corresponding biomaterials. Preliminary results
indicate that the modulus declines significantly (P ≤0.05) as GB1
domains increase when rings are tested to a strain of ~0.15. The
significant trend continues when calculating the modulus from
breaking tests. Rings with blocks of G-repeats break at
significantly higher strains (P = 0.0001). However, there is no
clear trend for breaking stress and ring type, although GRG5RG4R
breaks at significantly higher stress than the other ring types.
These results indicated that it is feasible to modulate the
macroscopic properties of these biomaterials in a wide range by
controlling the relative GB1/resilin content as well as
crosslinking density. However, systematic work is needed to fully
explore such approaches, including varying the modular sequences
with the same GB1/resilin content, to modulate the mechanical
properties of these novel biomaterials, just like how the passive
elastic properties of different muscles are mediated by different
isoforms of titin.
References 1. Carrion-Vazquez, M., Oberhauser, A. F., Fowler, S.
B., Marszalek, P. E., Broedel,
S. E., Clarke, J. & Fernandez, J. M. Mechanical and chemical
unfolding of a single protein: a comparison. Proc Natl Acad Sci U S
A 96, 3694-3699 (1999).
2. Elvin, C. M., Carr, A. G., Huson, M. G., Maxwell, J. M.,
Pearson, R. D., Vuocolo, T., Liyou, N. E., Wong, D. C., Merritt, D.
J. & Dixon, N. E. Synthesis and properties of crosslinked
recombinant pro-resilin. Nature 437, 999-1002 (2005).
3. Lyons, R. E., Nairn, K. M., Huson, M. G., Kim, M., Dumsday,
G. & Elvin, C. M. Comparisons of recombinant resilin-like
proteins: repetitive domains are sufficient to confer resilin-like
properties. Biomacromolecules 10, 3009-3014 (2009).
4. Cao, Y. & Li, H. Polyprotein of GB1 is an ideal
artificial elastomeric protein. NatMater 6, 109-114 (2007).
5. Fancy, D. A. & Kodadek, T. Chemistry for the analysis of
protein-protein interactions: rapid and efficient cross-linking
triggered by long wavelength light. Proc Natl Acad Sci U S A 96,
6020-6024 (1999).
6. Aaron, B. B. & Gosline, J. M. Optical properties of
single elastin fibres indicate random protein conformation. Nature
287, 865-867 (1980).
7. Savage, K. N. & Gosline, J. M. The role of proline in the
elastic mechanism of hydrated spider silks. J Exp Biol 211,
1948-1957 (2008).
8. Bell, E. & Gosline, J. Mechanical design of mussel
byssus: material yield enhances attachment strength. J Exp Biol
199, 1005-1017 (1996).
9. Lillie, M. A., Chalmers, G. W. & Gosline, J. M. The
effects of heating on the mechanical properties of arterial
elastin. Connect Tissue Res 31, 23-35 (1994).
10. Rief, M., Fernandez, J. M. & Gaub, H. E. Elastically
coupled two-level systems as a model for biopolymer extensibility.
Physical Review Letters 81, 4764-4767 (1998).
-
10www.nature.com/nature
doi: 10.1038/nature09024 SUPPLEMENTARY INFORMATION
10
11. Bell, G. I. Models for the specific adhesion of cells to
cells. Science 200, 618-627 (1978).
12. Evans, E. Probing the relation between force--lifetime--and
chemistry in single molecular bonds. Annu Rev Biophys Biomol Struct
30, 105-128 (2001).
9
feasibility, we constructed elastomeric proteins GRG5R, GRG9R
and G8, and their corresponding biomaterials. Preliminary results
indicate that the modulus declines significantly (P ≤0.05) as GB1
domains increase when rings are tested to a strain of ~0.15. The
significant trend continues when calculating the modulus from
breaking tests. Rings with blocks of G-repeats break at
significantly higher strains (P = 0.0001). However, there is no
clear trend for breaking stress and ring type, although GRG5RG4R
breaks at significantly higher stress than the other ring types.
These results indicated that it is feasible to modulate the
macroscopic properties of these biomaterials in a wide range by
controlling the relative GB1/resilin content as well as
crosslinking density. However, systematic work is needed to fully
explore such approaches, including varying the modular sequences
with the same GB1/resilin content, to modulate the mechanical
properties of these novel biomaterials, just like how the passive
elastic properties of different muscles are mediated by different
isoforms of titin.
References 1. Carrion-Vazquez, M., Oberhauser, A. F., Fowler, S.
B., Marszalek, P. E., Broedel,
S. E., Clarke, J. & Fernandez, J. M. Mechanical and chemical
unfolding of a single protein: a comparison. Proc Natl Acad Sci U S
A 96, 3694-3699 (1999).
2. Elvin, C. M., Carr, A. G., Huson, M. G., Maxwell, J. M.,
Pearson, R. D., Vuocolo, T., Liyou, N. E., Wong, D. C., Merritt, D.
J. & Dixon, N. E. Synthesis and properties of crosslinked
recombinant pro-resilin. Nature 437, 999-1002 (2005).
3. Lyons, R. E., Nairn, K. M., Huson, M. G., Kim, M., Dumsday,
G. & Elvin, C. M. Comparisons of recombinant resilin-like
proteins: repetitive domains are sufficient to confer resilin-like
properties. Biomacromolecules 10, 3009-3014 (2009).
4. Cao, Y. & Li, H. Polyprotein of GB1 is an ideal
artificial elastomeric protein. NatMater 6, 109-114 (2007).
5. Fancy, D. A. & Kodadek, T. Chemistry for the analysis of
protein-protein interactions: rapid and efficient cross-linking
triggered by long wavelength light. Proc Natl Acad Sci U S A 96,
6020-6024 (1999).
6. Aaron, B. B. & Gosline, J. M. Optical properties of
single elastin fibres indicate random protein conformation. Nature
287, 865-867 (1980).
7. Savage, K. N. & Gosline, J. M. The role of proline in the
elastic mechanism of hydrated spider silks. J Exp Biol 211,
1948-1957 (2008).
8. Bell, E. & Gosline, J. Mechanical design of mussel
byssus: material yield enhances attachment strength. J Exp Biol
199, 1005-1017 (1996).
9. Lillie, M. A., Chalmers, G. W. & Gosline, J. M. The
effects of heating on the mechanical properties of arterial
elastin. Connect Tissue Res 31, 23-35 (1994).
10. Rief, M., Fernandez, J. M. & Gaub, H. E. Elastically
coupled two-level systems as a model for biopolymer extensibility.
Physical Review Letters 81, 4764-4767 (1998).
TitleAuthorsAbstractMethods SummaryReferencesFigure 1
Force-extension curves of two polyproteins.Figure 2 Mechanical
properties of (G-R)4 and GRG5RG4R-based biomaterials.Figure 3
GB1-resilin-based biomaterials exhibit pronounced stress relaxation
behaviours.Figure 4 The macroscopic mechanical properties of
GB1-resilin-based biomaterials can be fine-tuned by controlling the
nanomechanicalproperties of the constituting elastomeric proteins
at the single-moleculelevel.