Sol and gel states in peptide hydrogels visualized by Gd(III)-enhanced magnetic resonance imaging

Article abstract of DOI:10.1002/bip.21612

The hydrogels assembled from a pair of self-repulsive but mutually attractive decapeptides are visualized by magnetic resonance imaging (MRI). It is found that in the absence of Gd(III)-chelate, gelation has little effect on MRI signal intensity. In the presence of Gd(III)-chelate, gelation leads to significant changes in water relaxation and MR signal intensity. The sol to gel transition is best visualized by T2-weighted imaging using large echo time with the sol producing a bright spot and the gel producing a dark spot. MRI studies point to high local Gd(III)-chelate concentration. Small-angle X-ray scattering study indicates that this local enrichment of Gd(III)-chelate has two contributing processes: first, the aggregation of peptides into fibers; second, within peptide fibers, Gd(III)- chelate further aggregate into clusters. This work demonstrates that the status of peptide-based hydrogels can be visualized by MRI with the aid of covalently linked Gd(III)-chelates. This result has implications for monitoring peptide scaffolds in vivo.

Full text of DOI:10.1002/bip.21612

Sol and Gel States in Peptide Hydrogels Visualized by
Sol and Gel States in Peptide Hydrogels Visualized by
Gd(III)-enhanced Magnetic Resonance Imaging
Gd(III)-enhanced Magnetic Resonance Imaging
1
2
3
4
5
Mahika Weerasekare, Marc B. Taraban, Xianfeng Shi, Eun-Kee Jeong, Jill Trewhella,
Yihua Bruce Yu
2
,6
1
Department of Pharmaceutics and Pharmaceutical Chemistry, University of Utah, Salt Lake City, UT 84112
Fischell Department of Bioengineering, University of Maryland, College Park, MD 20742
Department of Physics, University of Utah, Salt Lake City, UT 84112
2
3
4
5
6
Department of Radiology, University of Utah, Salt Lake City, UT 84112
School of Molecular Bioscience, The University of Sydney, Sydney, NSW 2006, Australia
Department of Pharmaceutical Sciences, University of Maryland, Baltimore, MD 21201
Received 31 December 2010; revised 21 February 2011; accepted 22 February 2011
Published online 7 March 2011 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/bip.21612
demonstrates that the status of peptide-based hydrogels
ABSTRACT:
can be visualized by MRI with the aid of covalently linked
Gd(III)-chelates. This result has implications for
monitoring peptide scaffolds in vivo. # 2011 Wiley
The hydrogels assembled from a pair of self-repulsive but
mutually attractive decapeptides are visualized by
magnetic resonance imaging (MRI). It is found that in the
absence of Gd(III)-chelate, gelation has little effect on MRI
signal intensity. In the presence of Gd(III)-chelate, gelation
leads to significant changes in water relaxation and MR
signal intensity. The sol to gel transition is best visualized
Periodicals, Inc. Biopolymers (Pept Sci) 96: 734-743, 2011.
Keywords: hydrogels; biomaterials; gadolinium chelate; contrast
agents; magnetic resonance imaging; small-angle X-ray scatter-
ing
by T -weighted imaging using large echo time with the sol
2
This article was originally published online as an accepted
preprint. The ‘‘Published Online’’ date corresponds to the
preprint version. You can request a copy of the preprint by
emailing the Biopolymers editorial office at biopolymers@wiley.
com
producing a bright spot and the gel producing a dark spot.
MRI studies point to high local Gd(III)-chelate
concentration. Small-angle X-ray scattering study
indicates that this local enrichment of Gd(III)-chelate has
two contributing processes: first, the aggregation of
peptides into fibers; second, within peptide fibers, Gd(III)-
chelate further aggregate into clusters. This work
INTRODUCTION
d(III)-chelates have a long history of broad applica-
tions in medical magnetic resonance imaging (MRI)
as effective contrast agents by virtue of their ability
to change water proton relaxation rates (T and
1
1
). Gd(III)-chelates can help visualize variations
G
T
2
Additional Supporting Information may be found in the online version of this
article.
in tissue pH, oxidative state, oxygen content, and the concen-
tration of certain metabolites. Recently, it has been demon-
strated that Gd(III)-macromolecular or peptide conjugates
are capable to sense the receptor environment and/or tissue
Correspondence to: Yihua Bruce Yu; e-mail: byu@rx.umaryland.edu
Contract grant sponsor: National Institutes of Health
Contract grant number: EB004416
Contract grant sponsors: Australian Federation Fellowship
Contract grant sponsors: U. S. Department of Energy
Contract grant number: DE-FG02-05ER64026
2
pH. Such sensitivity with respect to environment opens the
possibilities of using Gd(III)-chelates to trace the fate of
VV C 2011 Wiley Periodicals, Inc.
engineered biomaterials by means of MRI. Indeed, one of the
734
PeptideScience Volume 96 / Number 6

Sol and Gel States in Peptide Hydrogels
735
13
challenging tasks in the development of tissue repair scaffolds particles. Favorable results are also obtained when Gd(III)-
is how to monitoring them over time noninvasively.
chelates are conjugated to rigid dendrimers where the relax-
1
The status of tissue repair scaffolds can be monitored ivity is less dependent on water exchange rates. Water
4
3
through the relaxivities of water protons. For example, exchange rate is another factor that affects relaxivities of
15
changes in T1q images in the presence of iron oxide particles Gd(III)-conjugates.
4
encapsulated into hydrogel scaffolds could be used to moni-
These results suggest that the nuclear magnetic relaxation
tor the integrity of regenerated articular cartilage. However, properties of Gd(III)-chelates are controlled not only by the
T imaging, which appears to be a common technique to an- size or molecular weight of Gd(III)-conjugates, but also by
2
5
alyze soft tissue and cartilage, could not be used in this case the structural characteristics of the assembly, such as the
for monitoring the status of tissue repair due to the high sen- accessibility of the Gd(III) ion to water molecules as well as
sitivity of T to iron oxide. This is because in the presence of local distribution of Gd(III)-chelates. Indeed, T -relaxivity is
2
1
iron oxide, T becomes insensitive towards changes in the known to be controlled by the accessibility of Gd(III)-che-
2
repair process. However, Gd(III)-enhanced MRI is increas- lates by water molecules, while T -relaxivity strongly depends
2
1
ingly used to trace the structural changes in articular cartilage on the local concentration of Gd(III)-chelates. Most
6
implants. The dGEMRIC (delayed Gd(DTPA) -enhanced Gd(III)-conjugates that are currently employed as MRI con-
2ꢀ
7
MRI of cartilage) allows one to visualize biomacromolecular trast agents demonstrate T -dominated relaxation enhance-
1
networks for cartilage formation. However, the delayed and ment. However, in cases where local gadolinium concentra-
slow penetration of Gd(III)-chelates into the biopolymer tion is high, such as incorporating 9-12 Gd(III)-chelate into
16
scaffold significantly decreases the applicability of the tech- one dextran aggregate, or as in gadolinium oxide nanopar-
1
7
nique in tissue repair monitoring. These prior results point ticles, T -relaxation enhancement is more than T -relaxa-
2
1
to the evident need to incorporate Gd(III)-chelates into tion enhancement, and the resulting material becomes a T2-
tissue repair scaffolds for their noninvasive monitoring contrast agent. These prior studies suggest that Gd(III)-che-
using MRI.
lates in sol and gel states might behave differently, thereby
The most straightforward approach for incorporating generating a MR contrast that can be used to monitor sol to
Gd(III) into tissue repair scaffolds is to covalently conjugate gel transition. We explore this possibility in this work.
Gd(III)-chelates to the constituent molecules of the scaffolds.
In an effort to create peptide-based biomaterials, we
The resulting bulky assemblies would have increased rota- designed mutually attractive but self-repulsive oligopeptide
8
tional correlation time, s , and much better relaxivities. It pairs. Dissolved in physiological buffers, each peptide exists
r
has been shown that T -relaxivity of the hydrogel formed by as a stable solution. Upon mixing, the two peptides assemble
1
18,19
self-assembling amphiphilic peptides conjugated to Gd(III)- into fibrous networks.
In light of the potential in vivo
chelates is significantly higher as compared with the control applications of this type of mixing-induced hydrogels, it
hydrogel assembled from the same peptide without Gd(III)- would be very valuable if the two states of hydrogel, sol and
9
chelate. Significant enhancement in T -weighted MR imag- gel, show a sharp contrast in MR images so that the status of
1
ing has been attained when a Gd(III)-chelate was conjugated the hydrogel can be monitored by MRI. In this work, we con-
to a protein of 264 amino acid residues. The resulting assem- ducted MR imaging studies of the sol and gel states with and
bly was further cross-linked to form high molecular weight without Gd(III)-chelate. It was found that only in the pres-
hydrogel scaffold with controlled monomer content, ence of Gd(III)-chelate can a sharp contrast between sol and
1
0
sequence, and chain length.
gel be produced. Structure of the hydrogel, with and without
However, size or molecular weight of Gd(III)-conjugates Gd(III)-chelate, is probed by small-angle X-ray scattering
is only one of several factors affecting their relaxivities. For (SAXS) technique.
example, in the case of Gd(III)-chelates conjugated to linear
polymers, relaxivities do not improve significantly compared
to free Gd(III)-chelates. This is partly due to the flexibility of MATERIALS AND METHODS
linear polymers, and partly due to slower water exchange
rates in polymer-conjugated Gd(III)-chelates as compared Peptide Synthesis and Purification
1
1
The chemical structures of the two free peptides (1 and 2) and the
two paramagnetic chelate-peptide conjugates (3 and 4) used in this
study are shown in Table I. The negatively charged peptide module,
with free Gd(III)-chelates. Aside from molecular weight,
the relaxivity of Gd(III)-chelates attached to proteins is
known to also depend on the amino acid sequence of
1, contains neutral Trp and Ala and negatively charged Glu. The
positively charged peptide module, 2, contains neutral Trp and Ala
1
2
proteins. Very high relaxivities were detected when Gd(III)-
chelates are conjugated to the surface of nano-diamond and positively charged Orn (ornithine). Ornithine is used instead of
Biopolymers (Peptide Science)

7
36
Weerasekare et al.
a
Table I Sequences of Oligopeptides and the Structures of Paramagnetic Chelates
Compound No.
Peptide Sequences
w
M (Da)
1
2
3
4
formyl-EFEAEAEAEW-amide
formyl-OFOAOAOAOW-amide
1235.7
1162.7
1848.0
1842.7
[Gd(do3a)]GGGOFOAOAOAOW-amide
[Eu(do3a)]GGGOFOAOAOAOW-amide
Structures of paramagnetic peptide chelates
a
Modular material assembly is achieved by pairing a positively charged module (2, 3, and 4) with a negatively
charged module (1). A, alanine; E, glutamic acid; F, phenylalanine; G, glycine; O, ornithine; W, tryptophan. do3a
refers to the macrocyclic chelator 1,4,7,10-tetraazacyclododecane-1,4,7-triacetic acid. The C-terminals of 1, 2, 3,
and 4 are amidated (-amide); The N-terminal of 1 and 2 are formylated (formyl-).
lysine to increase aqueous solubility (ornithine has one fewer alytical runs. The temperature was 258C in all runs. In all runs, a
ꢀCH ꢀꢀ group in its side chain compared to lysine). For the same linear gradient was used (typically 0.25%B/min for preparative
reason, N-terminals of the two gelator peptides, 1 and 2, were formy- runs and 1 or 2%B/min for analytical runs).
ꢀ
2
2
0
lated instead of acetylated. To obtain the formylation agent, 2,4,5-
The mass value for each oligopeptide was verified by MALDI-
trichloro formic acid, dicyclohexyl carbodiimide, and 2,4,5-trichloro MS using Voyager DE-STR (PerSeptive Biosystems/ABI) with 337
phenol were added to a solution of 98% formic acid in ethylacetate nm nitrogen laser, 3 ns pulse, maximum 20 Hz firing rate, operated
precooled to 08C. The reaction mixture was then filtered and sequen- in a linear detector mode. (See additional Supporting Information.)
tially washed with 150 mL of saturated NaHCO
and 150 mL of water. The end product was obtained after drying a known procedure with slight modification: 6.0 g of GGG (31.7
over Na SO for 1 hr followed by rotary evaporation (*96% yield). mmol) was added to a solution of 8.4 g of sodium carbonate (79.25
Oligopeptides were synthesized using Fmoc solid-phase peptide mmol) in water (80 mL) followed by 45 mL of tetrahydrofuran, and
3
aqueous solution
Fmoc-Gly-Gly-Gly-COOH (Fmoc-GGG) was synthesized using
2
2
2
4
2
1
synthesis method and formylation of the N-terminal was carried the mixture was cooled to 08C. Then 8.2 g (31.7 mmol) of Fmoc-Cl
out using five-fold excess of 2,4,5-trichloro formic acid under ice- was added in small portions to the cooled mixture and the resulting
cold condition for 1 hr. A ninhydrin test for free N-terminal was mixture was equilibrated to room temperature for 4 hr followed by
used to monitor the completion of the formylation reaction. After overnight stirring. The resulting precipitate was dissolved in 800 mL
cleavage from the resin, oligopeptides were purified by ion-exchange of water and extracted with ethylacetate (3 3 200 mL). The aqueous
and reversed-phase HPLC, using a HP100 preparative chromatogra- fraction was cooled in an ice-water bath, acidified, and refrigerated
phy system. The purity of each oligopeptide was verified by analyti- overnight, then centrifuged to separate a fine precipitate that was
cal HPLC, using a HP100 analytical chromatography system. (See then lyophilized to obtain Fmoc-GGG as a white solid (13 g, 31.6
additional Supporting Information.)
For ion-exchange preparative HPLC, a Zorbax strong-anion
mmol, *100% yield).
[(do3a)]GGGOFOAOAOAOW-amide was synthesized by adding
exchange column (22.5 mm 3 250 mm) was used for purifying the four-fold excess of Fmoc-GGG to protected and resin-bound 2. Af-
negatively charged 1; and a Zorbax strong-cation exchange column ter removing the Fmoc-group by 20% piperidine in dimethyl-form-
(
22.5 mm 3 250 mm) was used for purifying the positively charged amide, 2.5-fold excess of [(do3a)]-tris-(t-butyl ester) was added. A
, 3, and 4. For reversed-phase preparative HPLC, a ZORBAX C18 longer reaction time (48 hrs) was necessary for conjugating
2
column (22.5 mm 3 250 mm) was used. For analytical HPLC, a [(do3a)]-tris-(t-butyl ester) to the N-terminus of the resin-bound
XDB-C18 column (4.6 mm 3 250 mm) was used. The eluents for peptide. After deprotection and cleavage from the resin with 90%
anion-exchange chromatography were A: 10 mM NH
water, pH 7; B: 2M NH Cl, 10 mM NH HCO in 80%MeCN/ silane, water, and dichloromethane, the product was concentrated
0%H O, pH 7. The eluents for cation-exchange chromatography by rotary evaporation, washed with ether, dissolved in water/aceto-
4 3
HCO in trifluoroacetic acid and 2.5% each of ethylene dithiol, tri-isopropyl
4
4
3
2
2
were A: 0.05% TFA in water, pH 2; B: 2M NH
0%H O with 0.05% TFA as the pH modifier. The eluents for ion-exchange and reversed-phase preparative HPLC to obtain
reversed-phase HPLC were the same as the ones for ion-exchange [(do3a)]GGG-OFOAOAOAOW-amide.
chromatography, except that there is no NH Cl in eluent B. The To obtain 3, [(do3a)]GGGOFOAOAOAOW-amide (86 mg, 0.05
flow rate was 5 mL/min for preparative runs and 1 mL/min for an- mmol) was dissolved in 1M NH OAc buffer at pH 7. To this, GdCl
4
Cl in 80%MeCN/ nitrile, and lyophilized. The resulting solid was purified using cat-
2
2
4
4
3
Biopolymers (Peptide Science)

Sol and Gel States in Peptide Hydrogels
737
(
26 mg, 0.1016 mM) was added and pH of the resulting solution Hz/pixel receiver bandwidth, and 3.0 s repetition time (TR). The
was adjusted to pH 7. After stirring at room temperature for 1 hr, experiment was repeated for variable echo times (TE) of 12, 24, 35,
the pH of the solution was increased to 9 to precipitate out excess 47, 59, 71, 83, 94, and 106 ms.
Gd(III) ions. After filtration the pH was adjusted back to 7. The
resulting compound 3 was purified by reversed-phase preparative
HPLC at pH 7. Compound 4 was synthesized following the above
procedure for compound 3, and EuCl was used instead of GdCl
Small-Angle X-ray Scattering (SAXS)
Sample Preparation. 10 mM solutions of 1, 2, and 3 were pre-
3
3
.
2 2
pared in PBS buffer (5% H O + 95% D O) and equilibrated on ice.
Mass values for 3 and 4 were verified using MALDI-MS and their
purity was assessed by analytical HPLC.
For hydrogel formation, 1 was mixed with equal amount of 2 or 3
gravimetrically so the concentration of each gelator in the mixture
is 5 mM. The mixture was vortexed in a 1.5 mL centrifuge tube
kept on ice to slow down gelation initiation. The mixed sample was
immediately centrifuged (30 s at < 2500 rpm) into a 1 mm thin-
walled (0.001 mm) glass capillary (Charles Supper Company)
fixed in a brass holder (which was also pre-equilibrated on ice)
and placed into the sample holder of SAXS instrument, which
was thermostated at 128C (the temperature at which the gelation
was monitored).
Peptides were dissolved in 100 mM NaCl and 50 mM sodium
phosphate buffer (phosphate-buffered saline, PBS), pH 7. Samples
for MRI experiments were prepared using 70% H
while samples for SAXS experiments were prepared using 5% H O
2 2
O + 30% D O
2
+
mined on the basis of the molar absorptivity of tryptophan at 280
95% D
2
O. Concentrations of the peptide solutions were deter-
ꢀ
1
ꢀ1 23
)
nm (e280 ¼ 5690M cm
and ICP-OES of Gd(III) or Eu(III)
(
Perkin-Elmer, Optima 3100XL).
Three hydrogels were prepared from 1, 2, 3, and 4, including: gel
+ 2, which contains no Gd(III)-chelate; gel 1 + 3, which contains
1
Gd(III)-chelate; and gel 1 + 3/4, which contains Gd(III)-chelate SAXS Data Measurements and Analysis. SAXS data were
diluted by Eu(III)-chelate.
acquired at 128C and reduced to I(q) versus q using the SAXS
27
instrument and procedures described previously. The parameter q
is the amplitude of the scattering vector (q-range for our experi-
ꢀ1
˚
Relaxivity Measurements and MRI
ments was 0.006–0.32 A ), and is equal to 4p(sinh)/k, where 2h is
Sample Preparation. For all MRI experiments, each compound the scattering angle and k is the wavelength of the incident X-ray (k
˚
(
1, 2, 3, and 4) was dissolved in PBS buffer (70% H
pH 7 with no correction for D O). For the relaxivity measurements, tance or vector length distribution function, P(r), for the scattering
one phantom was prepared with solutions containing mixtures of 3 particle and this function is calculated using indirect Fourier trans-
2
O + 30% D
2
O, ¼ 1.54 A). The inverse Fourier transform of I(q) gives the pair dis-
2
2
8–30
and 4; the concentration of 3 was varied from 0.1 to 1.0 mM, while form methods.
The r value at which P(r) goes to 0 gives the
the concentration of 4 was varied correspondingly from 4.9 to 4.0 maximum linear dimension for the scattering particle, dmax. The
mM to keep the total concentration of 3 and 4 equal to 5 mM. In second moment of P(r) is the radius of gyration of the scattering
the relaxivity measurements of gel 1 + 3/4, the concentration of 1 particle, Rg.
was always 5 mM while the concentrations of 3 and 4 were the same
as in sol 3/4. Also, to measure the relaxivity of 4, its concentration gated fibrils formed as a result of hydrogelation. Guinier showed
was varied from 0.1 to 2.5 mM.
that when one dimension of the particle is much larger than the
To collect MR images, two phantoms were prepared. The first other two, the scattering in the small q region (from 0.01 to 0.02
Guinier analysis for rod-shaped objects was used to analyze elon-
ꢀ
˚
A ) can be approximated as
1
31
phantom contained sol 1 (5 mM), sol 2 (5 mM), gel 1 + 2 (5 mM +
mM), sol 3 (5 mM), and gel 1 + 3 (5 mM + 5 mM). The second
phantom contained sol 3/4 (1 mM/4 mM) and gel 1 + 3/4 (5 mM +
mM/4 mM).
5
2
2
qIðqÞ ¼ I ð0Þeꢀ
q R =2
c
ð1Þ
c
1
Relaxation Measurements and MR Imaging. Relaxivity meas- where R
urements and T - and T -weighted images were performed using a particle and I (0) is proportional to the average mass per unit length
2
c
is the average radius of gyration of the cross-section of the
1
c
Siemens Trio 3 Tesla clinical MRI system (Siemens Medical Solu- along the fiber.
tions, Erlangen, Germany) with Avanto gradients (45 mT/m
strength and 150 T/m/s slew rate). The transmit/receive wrist coil used a simulated annealing algorithm implemented in a purpose-
For more detailed modeling of the cross-sectional shape, we
3
was used for the transmission of the RF wave and signal detection written program that is described elsewhere. This algorithm is a
2
3
3
to and from the sample. Net magnetization amplitude for each sam- 2D implementation of the 3D dummy atom modeling of Svergun.
ple was determined from the appropriate region of interest using In all calculations, dummy atoms were arranged in a grid with
2
4
MRIcro software.
dimensions of 30 3 50 of dummy atoms. Based on the estimates of
-weighted images were obtained using the inversion-recovery the resolution of SAXS technique, the reasonable size of each
3
4
T
1
2
5
˚
spin echo pulse sequence with a 192 3 256 imaging matrix, 135 3 dummy atom was 3 A in diameter [in order not to affect the results
2
2
˚
1
80 mm field of view, 0.7 3 0.7 mm in-plane resolution with 2.5 of final model fit, this value should be < p/(2q ), which is 4.9 A
max
1
ꢀ
˚
m slice thickness, 100 Hz/pixel receiver bandwidth, 5.0 s TR, and 16 for our qmax * 0.32 A ]. The program evaluates pair distance dis-
ms TE for variable inversion recovery times (TI) of 25, 35, 50, 75, tribution function [P (r)] for the cross-section of the fibers and cal-
00, 200, 400, 800, and 1600 ms. culates the scattering profile of the fibers. From P (r), the maximum
For T -weighted imaging, a turbo-spin echo was employed cross-sectional dimension (D
using a 144 3 256 imaging matrix, 100 3 180 mm field of view, the radius of gyration of the cross-section (R
c
1
c
2
6
2
c
), the cross-sectional area (S
), were determined.
.7 3 0.7 mm in-plane resolution with 2.5 m slice thickness, 130 Corrections for the instrument slit geometry, scaling, and incoher-
c
), and
2
c
2
0
Biopolymers (Peptide Science)

7
38
Weerasekare et al.
the hope that covalently bound Gd(III)-chelate would result
in significant MR signal change upon gelation.
Hydrogelation in the Presence of Gd(III)-chelate (Gel
1
+ 3). Figure 2 shows that in the presence of covalently
bound Gd(III)-chelate, gelation indeed produced drastic
change in MR signal intensity, to the point that the gel is
completely dark in both T - and T -weighted images. This is
1
2
due to complete signal saturation, which makes it impossible
to extract T and T values. Such signal saturation suggests
1
2
high Gd(III) concentration. But the overall Gd(III) concen-
tration in gel 1 + 3 is the same as in sol 3, which is 5 mM in
both cases. The fact that the MR signal is not saturated in sol
3
but saturated in gel 1 + 3 points to high local Gd(III) con-
centration as a result of peptide gelation.
The sharp contrast between sol 3 and gel 1 + 3 means that
the gelation process can be visualized using either T -weighted
1
imaging at long TI values (Figure 2A) or T -weighted imaging
2
FIGURE 1 Images of hydrogelation with no Gd(III)-chelate. (A)
T1-weighted images with different inversion recovery time TI; (B)
2
T -weighted images with different echo time TE; (C) MRI intensity
vs. TI; (D) MRI intensity vs. TE.
ent background are applied to compare the simulated model profile
3
with experimental scattering data.
2
RESULTS AND DISCUSSION
Relaxivity Determination and MRI
Hydrogelation in the Absence of Gd(III)-chelate (Gel
1
+ 2). Figure 1 shows that in the absence of Gd(III)-che-
late, gelation produced negligible change in MR signal inten-
sity and MR images of gel 1 + 2 are almost identical to those
of sol 1 and sol 2. These results indicate that in the absence
of Gd(III)-chelate, there is negligible contrast between the sol
and the gel states. This is hardly surprising considering that
FIGURE 2 Images of hydrogelation with Gd(III)-chelate. (A)
T
1
-weighted images with different inversion recovery time TI; (B)
T2-weighted images with different echo time TE; (C) MRI intensity
jugated Gd(III)-chelate to one of the gelator peptides, with vs. TI; (D) MRI intensity vs. TE.
*
99% w/w of the hydrogel is water. In light of this, we con-
Biopolymers (Peptide Science)

Sol and Gel States in Peptide Hydrogels
739
FIGURE 3 T (A) and T (B) in sol 4. The concentration of 4 was varied from 0.1 to 2.5 mM. T
1
1
2
(
2
C) and T (D) in sol 3/4 (orange) and in gel 1 + 3/4 (wine). In sol 3/4, the total concentration of 3
and 4 is 5 mM with that of 3 varied from 0.1 to 1.0 mM, and that of 4 was varied correspondingly
from 4.9 mM to 4.0 mM. In gel 1 + 3/4, the concentration of 1 is 5 mM while the concentrations of
3
and 4 are the same as in sol 3/4. The T
from the slope of linear fit; the goodness of fit for (A) and (B) is 0.985 and that for (C) and (D) is
.9995. R , was not calculated for gel 1 + 3/4 due to the nonlinearity of the data, which is indicative
1 1 2 2
-relaxivity, R , and the T -relaxivity, R , are calculated
0
2
of clustering of Gd(III)-chelates in the hydrogel matrix.
at short TE values (Figure 2B), with the sol producing bright ing the concentrations of both 1 and 3 from 5 to 1 mM will
spots and the gel producing dark spots.
abolish gelation. To solve this difficulty, we opted to ‘‘dilute’’
Gd(III)-chelate by the structurally and chemically similar
Hydrogelation in the Presence of Diluted Gd(III)-chelate Eu(III)-chelate. It is well-known that Eu(III)-chelates, the
Gel 1 + 3/4). If the saturation of MR signal in gel 1 + 3 is closest analog of Gd(III)-chelates, are much weaker MRI
(
indeed caused by high local Gd(III)-chelate concentration, contrast agents as compared to Gd(III)-chelates, and present
1
then reducing the overall Gd(III)-chelate concentration has little influence on the relaxation rates of H O. The dilution
2
the potential to eliminate saturation. However, simply lower- of Gd(III)-chelate with Eu(III)-chelate is realized by making
Biopolymers (Peptide Science)

7
40
Weerasekare et al.
gel 1 + 3/4 (Figure 3). However, gelation shortened T signifi-
2
cantly. Further, unlike T , T in gel 1 + 3/4 does not have a lin-
1
2
ear dependency on Gd(III) concentration, once again suggest-
ing high local concentration of Gd(III)-chelates in the hydrogel.
With T1 slightly elevated in the gel, one might have
expected gel 1 + 3/4 to have higher signal intensity and
FIGURE 4 Images of hydrogelation with Gd(III)-chelate diluted
by Eu(III)-chelate. (A) T -weighted images with different inversion
recovery time TI; (B) T -weighted images with different echo time
1
2
TE; (C) MRI intensity vs. TI; (D) MRI intensity vs. TE.
a new compound, 4, in which Eu(III)-chelate is conjugated
to the gelator 2 (Table I). Diluting 3 with 4 is far better than
diluting 3 with 2 because the structural dissimilarity between
3
and 2 may cause significant structural variation in the
resulting hydrogel.
To confirm that 4 indeed has little effect on water relaxa-
1
tion, T and T of H O at various concentrations of 4 are
1
2
2
determined. As expected, the T and T relaxivities of 4 are
1
2
ꢀ
1
ꢀ1
ꢀ1 ꢀ1
0
.016 mM
s
and 0.072 mM
s
, respectively, much
ꢀ
1
ꢀ1
lower than those of 3, which are 8.46 mM
ꢀ1
s and 10.75
ꢀ1
mM
s , respectively (Figure 3).
With 1 mM overall Gd(III)-chelate, MR signal from gel is no
longer saturated. Indeed, gel 1 + 3/4 produced bright spots in
MR images at certain TI and TE values (Figure 4), attesting to
the effectiveness of diluting 3 with 4. As for water relaxivity,
FIGURE 5 Scattering profiles of I(q) vs. q (A) and the corre-
sponding Guinier plots for rod-like particles (B). Scattering data
were collected after 24 hr of gelation; red circles, gel 1 + 2; green
circles, gel 1 + 3 pair. Arrows in the Guinier plots mark the roll-over
for points characteristic for rod-shaped particles.
gelation shortened T only slightly, with the T relaxivity
1
1
ꢀ
1
ꢀ1
ꢀ1 ꢀ1
increased from 8.46 mM
s
for sol 3/4 to 9.94 mM
s
Biopolymers (Peptide Science)

Sol and Gel States in Peptide Hydrogels
741
Combining the imaging results at 5 and 1 mM overall
Gd(III) concentration, the sol to gel transition is best
visualized by T -weighted imaging at higher TE values with
2
the sol producing a bright spot and the gel producing a dark
spot.
Small-Angle X-Ray Scattering (SAXS)
MRI studies point to high local concentration of Gd(III)-
chelate as a result of hydrogelation. Considering that these
1
8
hydrogels are made of peptide nanofibers, it is hardly sur-
prising that gelation leads to inhomogeneous distribution of
Gd(III) and thereby high local Gd(III) concentration. What
is surprising is the extent of Gd(III) aggregation in gel 3 as
the gel is completely dark in both T - and T -weighted
1
2
images. One possibility is that Gd(III)-chelate further aggre-
gates inside the peptide fibers, forming clusters. To investi-
gate this possibility, we employed SAXS to compare the
structure of gel 1 + 2 with that of gel 1 + 3.
FIGURE 6 Representative 2D dummy atom models of the cross-
sections of the fibers in gel 1 + 2 (red) and gel 1 + 3 (green).
Both gels, 1 + 2 and 1 + 3, form networks made of fibrils
with high scattering intensity [I(q)] (Figure 5A). The co-as-
sembly of the peptides results in long-range ordered struc-
tures with the X-ray scattering intensity reaching a plateau
thereby brighter spots in T -weighted images than sol 3/4.
1
But the opposite was observed: gel 1 + 3/4 produces lower
signal intensity than sol 3/4 (Figure 4C) and less bright spots
2
within 24 hr after mixing. Guinier plots of lnqI(q) vs q for
2
the two gels after 24 hr show a distinctive roll-over at q < 2
in T -weighted images (Figure 4A). The explanation of this
1
ꢀ4 ꢀ2
˚
A (Figure 5B), which is characteristic of highly
3
10
observation lies in the fact that T is shortened to a much
2
asymmetric particles with one dimension much larger than
31
the other two. Indirect Fourier transform of I(q) to P(r)
greater extent than T upon gelation (Figures 3C and 3D).
1
Keep in mind that MR signal intensity SI is determined not
(not shown) suggests that the length of the fibrils in both gels
only by T but also by T with the following functional form:
1
2
exceed the limits of precise measurement for our instrument,
ꢂ
ꢀ
ꢁꢃ
ꢀ
ꢁ
TI
T1
TE
T2
SI / 1 ꢀ 2 exp ꢀ
3 exp ꢀ
ð2Þ
where TI and TE are the recovery and echo times, respec-
5,26
2
tively.
It is seen from Eq. (2) that T shortening enhances
1
SI while T shortening reduces SI. When T -relaxivity domi-
2
2
nates, that is, when T is shortened to a much greater extent
2
than T , SI can decrease even as T is shortened. At 1 mM
1
1
Gd(III) upon gelation, T has a 15% reduction, from 104.8
1
ms in sol 3/4 to 89.8 ms in gel 1 + 3/4 while T has a 70%
2
reduction, from 81.2 ms in sol 3/4 to 25.0 ms in gel 1 + 3/4.
Clearly, as a result of aggregation of Gd(III) in the gel, T2
relaxation dominates, which is reflected in MR signal inten-
sity and images (Figure 4).
As for visualizing the sol and gel states, the contrast
between sol 3/4 and gel 1 + 3/4 is not as stark as that between
sol 3 and gel 1 + 3 in T -weighted images. Unlike gel 1 + 3,
1
c
FIGURE 7 Pair-wise distance distribution functions [P (r)]
gel 1 + 3/4 is not completely dark at all TI values. However,
just like gel 1 + 3, gel 1 + 3/4 is still completely dark at higher
TE values.
obtained from 2D dummy atom model of the cross-sections of rod-
like fibers in gel 1 + 2 (red) and gel 1 + 3 (green). Arrows mark the
shift in the prevailing vector lengths.
Biopolymers (Peptide Science)

7
42
Weerasekare et al.
Table II Structural Parameters of Oligopeptide Hydrogels
a
Derived from Two-Dimensional Analysis of SAXS Data
spots in both T and T -weighed images. At 1 mM overall
1 2
Gd(III) concentration, gelation has much stronger effect on
T relaxation than on T relaxation. The MRI results point to
2
2
2
v
1
˚
D (A)
c
˚
S (A )
c
˚
R (A)
c
Gel
local enrichment of Gd(III)-chelate, which has two contrib-
uting processes: first, the aggregation of oligopeptides into
1
1
+ 2
+ 3
170
170
3762
6309
44.3
41.7
0.81
0.79 fibers; second, within the peptide fibers, Gd(III)-chelate fur-
ther aggregate into clusters, as indicated by SAXS studies. In the
a
˚
is the maximum dimension of the peptide fiber cross-section in A;
D
c
˚
2
S is the area of the peptide fiber cross-section in A ; R is the radius of gyra-
c c
presence of covalently bound Gd(III)-chelate, there is s sharp
˚
2
tion of the peptide fiber cross-section in A; v reflects the quality of fit contrast between the sol and gel states of peptide hydrogels.
2
between the model and experimental data (v < 1 indicates excellent fit).
which are limited to a minimum measurable q value of 0.008 REFERENCES
ꢀ
1
˚
˚
, corresponding to a maximum length scale of 400 A
1. Caravan, P.; Ellison, J. J.; McMurry, T. J.; Lauffer, R. B. Chem
Rev 1999, 99, 2293–22352.
A
3
4
˚
(
dmax < p/qmin; dmax–400 A). Therefore, we turned to the
analysis of fiber cross-section in these two gels.
2
. De Leon-Rodriguez, L. M.; Lubag, A. J. M.; Malloy, C. R.; Mar-
tinez, G. V.; Gillies, R. J.; Sherry, A. D. Acc Chem Res 2009, 42,
The two-dimensional analysis used a 2D dummy atom
modeling approach that provides the representative shape of
the cross-sections of the fibers underlying the hydrogel
9
48–957.
3. Marcos, M.; Cano, P.; Fantazzini, P.; Garavaglia, C.; Gomez, S.;
Garrido, L. Magn Reson Imaging 2006, 24, 89–95.
4
. Saldenha, K. J.; Doan, R. P.; Ainslie, K. M.; Desai, T. A.; Majumdar,
S. Magn Reson Imaging 2011, 29, 40–49.
network (Figure 6) and the associated P (r) functions
c
(
Figure 7). From P (r), the maximum dimension of fiber
c
5
. Taylor, C.; Carballido-Gamio, J.; Majumdar, S.; Li, X. Magn
Reson Imaging 2009, 27, 779–784.
cross-section, D , the area of fiber cross-section, S , and the
c
c
radius of gyration of fiber cross-section, R , can be calculated
c
6. Plank, C. M.; Kubin, K.; Weber, M.; Marlovits, S.; Trattnig, S. C.
Magn Reson Imaging 2005, 23, 739–744.
3
2
(
Table II). The fibers in both gels have identical D values.
c
Nonetheless, fibers in gel 1 + 3 are thicker than fibers in gel
7. Burstein, D.; Velyvis, J.; Scott, K. T.; Stock, K. W.; Kim, Y. J.; Jar-
amillo, D.; Boutin, R. D.; Gray, M. L. Magn Reson Imaging
˚
˚
1
+ 2 (60 A vs. 45 A). One might have expected thicker fibers
would lead to higher S and R values. However, S and R
c
2
. Brasch, R. C. Magn Reson Med 1991, 22, 282–287.
001, 45, 36–41.
c
c
c
8
9
show the opposite trend of variation among the two hydro-
gels: while S increases from gel 1 + 2 to gel 1 + 3 (nearly
. Bul,l S. R.; Guler, M. O.; Bras, R. E.; Venkatasubramanian, P. N.;
Stupp, S. I.; Meade, T. J. Bioconjug Chem 2005, 16, 1343–1348.
c
two-fold), R decreases from gel 1 + 2 to gel 1 + 3 by almost 10. Karfield, L. S.; Bull, S. R.; Davis, N. E.; Meade, T. J.; Barron, A.
c
˚
A. R is the root mean square contrast-weighted distance
c
E. Bioconjug Chem 2007, 18, 1697–1700.
1. Toth, E.; Helm, L.; Kellar, K. E.; Merbach, A. E. Chem Eur J
3
1
1
of area elements from the center of scattering density, where
the contrast is the difference in electron density between the
1
999, 5, 1202–1211.
2. Karfield-Sulzer, L. S.; Waters, E. A.; Davis, N. E.; Meade, T. J.;
Barron, A. E. Biomacromolecules 2010, 11, 1429–1436.
the larger cross-sectional area of gel 1 + 3 thus implies that 13. Manus, L. M.; Mastarone, D. J.; Waters, E. A.; Zhang, X. Q.;
scattering particle and the solvent. The smaller R value for
c
the electron dense Gd(III) is on average located more toward
the center of the cross sectional area. It therefore appears that
the incorporation of Gd(III)-chelate leads to thicker fibers
with Gd(III)-chelate clustered toward the center of fiber
cross-section. This further clustering of Gd(III)-chelates
within peptide fibers provides an explanation as to why the
MR signal is completely saturated in gel 1 + 3.
Schultz-Sikma, E. A.; MacRenaris, K. W.; Ho, D.; Meade, T. J.
Nano Lett 2010, 10, 484–489.
1
4. Wiener, E. C.; Brechbiel, M. W.; Brothers, B.; Magin, R. L.; Gan-
sow, O. A.; Tomalia, D. A.; Lauterbur, P. C. Magn Reson Med
1
994, 31, 1–8.
1
1
1
5. Toth, E.; Pubanz, D.; Vauthey, S.; Helm, L.; Merbach, A. E.
Chem Eur J 1996, 2, 1607–1615.
6. Modo, M.; Mellowdew, K.; Cash, D.; Fraser, S. E.; Meade, T. J.;
Williams, S. C. Neuroimage 2004, 21, 311–317.
7. McDonald, M. A.; Watkin, K. L. Acad Radiol 2006, 13, 421–
4
27.
CONCLUSION
18. Ramachandran, S.; Tseng, Y.; Yu, Y. B. Biomacromolecules 2005,
6, 1316–1321.
In the absence of Gd(III)-chelate, gelation of the oligopeptide
pair has little effect on MR signal intensity. In the presence of
covalently linked Gd(III)-chelate, gelation of the oligopeptide
pair produces significant change in water relaxation and MR
signal intensity. At 5 mM overall Gd(III) concentration, gela-
tion results in complete signal saturation and thereby dark
1
9. Ramachandran, S.; Trewhella, J.; Tseng, Y.; Yu, Y. B. Chem Mater
006, 18, 6157–6162.
2
2
2
0. Martinez, J.; Laur, J. Synthesis 1982, 11, 979–981.
1. Chan, W. C.; White, P. D. Fmoc Solid Phase Peptide Synthesis:
A Practical Approach; OxfordUniversity Press: New York, 2000;
p 1–75.
Biopolymers (Peptide Science)

Sol and Gel States in Peptide Hydrogels
743
2
2
2. Bodanszky, M.; Bodanszky, A. The Practice of Peptide Synthesis; 28. Svergun, D. I.; Semenyuk, A. V.; Feigin, L. A. Acta Crystallogr
1988, A44, 244–250.
Springer-Verlag: Berlin, 1994. p 21–22.
3. Gill, S. C.; von Hippel, P. H. Anal Biochem 1989, 182,
2
3
3
3
9. Svergun, D. I. J Appl Crystallogr 1991, 24, 485–492.
0. Svergun, D. I. J Appl Crystallogr 1992, 25, 495–503.
1. Guinier, A. Ann Phys (Paris) 1939, 12, 161–236.
3
19–326.
2
2
4. Rorden, C.; Brett, M. Behav Neurol 2000, 12, 191–200.
5. Smith R. C.; Reinhold C.; McCauley, T. R.; Lange, R. C.;
Constable, R. T.; Kier, R.; McCarthy, S. Radiology 1992, 184,
2. Whitten, A. E.; Jeffries, C. M.; Harris, S. P.; Trewhella, J. Proc
Natl Acad Sci USA 2008, 105, 18360–18365.
6
71–675.
6. Hennig, J.; Naureth, A.; Friedburg, H. Magn Reson Med 1986,
, 823–833.
7. Heidorn, D. B.; Trewhella, J. Biochemistry 1988, 27, 909–915.
3
3. Svergun, D. I. Biophys J 1999, 76, 2879–2886.
2
2
3
34. Glatter, O.; Kratky, O., Eds. Small Angle X-Ray Scattering; Aca-
demic Press; London, 1982.
Biopolymers (Peptide Science)