Supramolecular hydrogels for enzymatically triggered self-immolative drug delivery

Article abstract of DOI:10.1016/j.tet.2010.02.033

For the first time the combination of self-immolative spacers and supramolecular hydrogels has been tested in enzyme triggered drug release. Low-molecular weight drug-gelator conjugates have been prepared, which contain a gel forming lysine moiety linked to model drugs (benzylamine and phenethylamine) through a self-immolating spacer (p-aminobenzyloxycarbonyl). In the presence of trypsin the amide linkage between the gelator moiety and the spacer is hydrolyzed leading to the release of the model drug. This approach provides with distinct advantages, such as sustained release or versatility associated to the use of supramolecular hydrogels and self-immolative spacers, respectively.

Full text of DOI:10.1016/j.tet.2010.02.033

Tetrahedron 66 (2010) 2614–2618
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Supramolecular hydrogels for enzymatically triggered self-immolative
drug delivery
*
*
´
´
Jose A. Saez, Beatriu Escuder , Juan F. Miravet
`
`
´
Departament de Qu´ımica Inorganica i Organica, Universitat Jaume I, 12071 Castello, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 7 January 2010
Received in revised form 4 February 2010
Accepted 5 February 2010
Available online 10 February 2010
For the first time the combination of self-immolative spacers and supramolecular hydrogels has been
tested in enzyme triggered drug release. Low-molecular weight drug-gelator conjugates have been
prepared, which contain a gel forming lysine moiety linked to model drugs (benzylamine and phene-
thylamine) through a self-immolating spacer (p-aminobenzyloxycarbonyl). In the presence of trypsin the
amide linkage between the gelator moiety and the spacer is hydrolyzed leading to the release of the
model drug. This approach provides with distinct advantages, such as sustained release or versatility
associated to the use of supramolecular hydrogels and self-immolative spacers, respectively.
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Prodrugs
Supramolecular hydrogels
Enzyme catalysis
Self-assembly
Self-immolation
Drug delivery
1. Introduction
effect and that self-assemble into a hydrogel (Scheme 1B). In that
case, sol-gel equilibrium and uptake of the drug will determine the
Polymeric hydrogels have been extensively studied in bio-
medical applications, for example, in regenerative medicine and as
drug delivery systems.1 However, in the recent years much atten-
tion has been paid to supramolecular hydrogels, formed by low-
molecular weight compounds that self-assemble in water by
rate of release.6,7 In that later case, serendipity plays an important
role as not all the desired drugs will form hydrogels. Another option
is to conjugate a drug fragment with a reliable assembling moiety
by covalent attachment. In that case, the hydrogel can act as a res-
ervoir that slowly disassembles and releases the active fragment. If
this fragment is connected by an enzyme sensitive group, in the
presence of that, it will be consumed and the gel-sol equilibrium
will be shifted to sol, provoking a controlled release of the active
fragment.6,8,9 Enzymes are one of the most efficient stimuli that can
weak non-covalent interactions (hydrophobic, van der Waals, p–p
stacking, etc.). The use of those systems may present some ad-
vantages as, for example, synthetic economy due to the self-
instructed bottom-up assembly, biocompatibility, lower toxicity,
increased biodegradability, and, importantly, thermoreversible
formation-dissociation processes, which occur faster than in poly-
mers. Additionally, the formation of supramolecular gels can be
controlled by light or chemical stimuli, such as pH by appropriate
design of the gelators.2,3 Supramolecular hydrogels have been
particularly investigated as drug delivery systems.4 In this sense,
different approaches have been undertaken. First, a simple system
can be constructed by entrapment of a hydrophobic drug in the
voids of a hydrogel (Scheme 1A). After degradation of the hydrogel,
caused by a convenient stimulus (change in temperature or pH,
enzyme action), the drug is released.5 A second strategy is to pre-
pare drug-gelators, namely, molecules that present a therapeutic
A
drug
enzyme
self-assembling
fragment
B
linker
C
HYDROGEL
SOLUTION
* Corresponding authors. Tel.: þ34 964 729155; e-mail addresses: escuder@
Scheme 1. Hydrogel-based approaches for drug delivery.
qio.uji.es (B. Escuder), miravet@qio.uji.es (J.F. Miravet).
0040-4020/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2010.02.033

´
J.A. Saez et al. / Tetrahedron 66 (2010) 2614–2618
2615
be used in drug delivery, as they allow the release of drugs in very
specific locations. For instance, drug release in tumors as a result of
the enzymatic action of tumor-associated proteases10 or in selected
areas of the GI tract under the influence of digestive enzymes11 has
been reported.
Proteases and lipases have been typically used as enzymes being
drugs, in many cases, directly attached to the hydrogelating core by
the bond that has to be broken by the enzyme. In some cases, due
for instance to unexpected stereoelectronic effects, the enzyme
activity may be limited. Furthermore, the change of drug structure
may also change the aggregation properties of the gelator. In order
to avoid these random problems, the use of spacers may be an
excellent option for the general design of a hydrogel carrier
(Fig. 1C).9 Moreover, it has to be considered that all the molecular
fragments produced after gel fragmentation must be innocuous
and easily excreted from the body.
a PABC linker with the intention to use it as an enzyme triggered
drug delivery hydrogel system (Scheme 3). Fmoc N-protected
amino acids and small peptides have been shown to be excellent
hydrogelators.7,16 Besides, peptidic bond between Lys and PABC
linker can be easily cleaved by Trypsin, an enzyme, which selec-
tively cleaves peptides at the carboxyl side of lysine and arginine
amino acids, causing the subsequent ‘self-immolation’ of the spacer
and release of model drug. This approach aims to combine features
from hydrogels, as, for example, slow and controlled disassembly of
prodrug, with in situ enzymatic action and release of active drug.
2. Results and discussion
Hydrogelators 1 and 2 were prepared by a synthetic procedure
that involved three steps (see Scheme 4). First, Fmoc-L
-Lys (tBoc)-
OH was reacted with 4-aminobenzyl alcohol to yield intermediate
Figure 1. Cryo-SEM images of hydrogels 1 (left) and 2 (right).
A widely studied and extraordinarily useful family of linkers is
the so-called ‘self-immolative’ spacers that after the action of an
external stimulus (either chemical or enzymatic) start a domino-
like fragmentation to yield small molecular fragments.12 This con-
cept has been applied for many years by Shabat and co-workers
among others in drug delivery and signal amplification. For in-
stance, they have reported cascade-release dendrimers for
3, which was treated without further purification with 4-nitro-
phenyl chloroformate and the corresponding amine model drug to
yield the tBoc protected prodrugs 4 and 5. After deprotection with
TFA hydrogelators 1 and 2 were obtained.
Compounds 1 and 2 did not form gels in pure water, being
necessary the presence of a small amount of an alcoholic co-solvent
and fine tuning of pH. In order to optimize the gelation conditions,
the influence of variables, such as the use of different alcohols as co-
solvents or the use of different buffers in various concentrations to
control the pH of the media were studied. Optimum gels, which
presented an appropriate gelator concentration threshold and
a thermal stability (Tgel) convenient for experiments under enzy-
matic conditions were achieved by using ethanol as co-solvent (20
v/v%) and NaHCO3 0.1 M as buffer (pH¼8.2, 25 ꢀC). Gels were ac-
tually prepared by fast addition of the buffer to the ethanolic so-
lution of the gelators. Opaque gels were formed immediately and
were stable for several weeks at room temperature. In the condi-
tions described above, both gelators showed a minimum gel con-
centration of 0.1 w/v% at room temperature, and Tgel values up to
85 ꢀC (0.5 w/v%, conditions used in enzymatic experiments). The
use of other alcohols (methanol, isopropanol, propanol) or different
buffering conditions afforded a less efficient gel formation.
the liberation of multiple drugs by single trigger events13
,
self-immolative lineal polymers14 and low-molecular weight self-
eliminating drug conjugates.15 One of those linkers is p-amino-
benzyloxycarbonyl (PABC) (Scheme 2) that with the amino group
being protected is stable under physiological conditions, whereas
when the amino group is free undergoes a rapid 1,6-elimination to
carbamic or carbonic acids that are unstable and decompose to the
respective amines or alcohols.
O
R
O
O
X
1,6-elimination
R
+
O
X
H2N
H2N
X = NH, O
H2O
The microscopic morphology of hydrogels was examined by cryo-
SEM (Fig. 1). Both hydrogels showed a sponge-like aspect in which
a network made of objects with fibrillar aspect appears forming
cavities of micrometer size that in the original gel phasewere filled by
the solvent. Typically, fibrils have dimensions ranging from hundreds
of nanometers of width and several micrometers of length.
OH
CO2
+
+
HX-R
H2N
Scheme 2. 1,6-Elimination of PABC derivatives.
In order to study the enzyme-mediated drug release form
hydrogels 1 and 2, we used trypsin dissolved in the appropriate
volume of NaHCO3 0.1 M buffer and injected over the alcoholic
solution of the gelator. Then, the hydrogel was incubated at
37 ꢀC and after 30 h it could be observed that it was
Here we report on the design of a ‘proof-of-concept’ supramo-
lecular hydrogelator formed by a Fmoc- -Lysine based assembling
fragment that can be easily conjugated to model drugs through
L

´
J.A. Saez et al. / Tetrahedron 66 (2010) 2614–2618
2616
MODEL DRUG
O
O
NH
O
HN
O
HN
O
n
n
trypsin
O
O
self-immolation
HN
H2N
NH
H2N
n
SELF-IMMOLATIVE
LINKER
FREE
MODEL DRUG
OH
+
H3N
H2N
1, n = 1
2, n = 2
O
O
SELF-ASSEMBLING
FRAGMENT
O
O
CO2
H3N
Scheme 3. Enzymatic cleavage and self-immolative degradation mechanism of gelators 1 and 2.
disassembled into a solution with a fine white solid suspension
(Scheme 5). After that period of time, fully degradation of
gelator was simply proved by TLC, being observed spots similar
to the self-immolated fragments, namely, Fmoc-Lys (tBoc)-OH, 4-
aminobenzyl alcohol and the corresponding amine (see Scheme
3). Furthermore, 1H NMR of the final mixture was identical to
a blank spectrum of a mixture of the three components.
In order to determine the rate at which the gelator was released
and degradated by Trypsin 1H NMR was used. In particular, in-
tegration of aromatic signals of the immolated products in solution
was found to be useful since molecules in the gel phase are NMR
silent (see Supplementary data). Therefore, we prepared hydrogels
in NMR tubes by dissolving the appropriate amount of gelator
(5 mg/mL) in 0.1 mL of methanol-d4 followed by injection of 0.4 mL
of 0.1 M NaHCO3 buffer prepared in D2O with a 0.8% mol of Trypsin
(respect to gelator). Sample was maintained at 37 ꢀC and spectra
were taken every 10–15 min for a total of 30 h, time enough at this
enzyme concentration to achieve the total degradation of the
gelator (Fig. 2A). Additionally, blank NMR samples of both 1 and 2
were prepared and incubated at 37 ꢀC without addition of the
O
O
NH
O
EEDQ
DME
OH
NH
O
O
OH
H2N
O
HN
HN
0ºC, N2
4
OH
HN
O
4
O
O
3
O
1) THF, 0 ºC, N 2, pyr
4-nitrophenyl chloroformate
2)
H2N
n
THF, r.t., N 2, Et 3N
O
O
NH
O
HN
n
O
HN
O
HN
4
O
O
TFA, CH2Cl2, 0 ºC
4, n = 1
5, n = 2
O
NH
O
HN
O
n
O
O
HN
H3N
4
CF3COO-
1, n = 1
2, n = 2
Scheme 4. Synthesis of hydrogelators 1 and 2.

´
J.A. Saez et al. / Tetrahedron 66 (2010) 2614–2618
2617
enzyme-breakable bond is moved away from the drug moiety
and different chemical functions can be used to link the spacer
to the desired drug. In this way the possible influence of the
drug structure in both enzyme action and hydrogel assembly
have been minimized, leading to a highly versatile system. The
use of reversible supramolecular gels allows for a sustained re-
lease of the actives associated to the progressive gel dissociation.
It is envisaged that changes in the gelling scaffold and media
conditions (temperature, ionic strength, chemical species.)
would permit to control the release rate to fit with the desired
drug dosage requirements. In particular, these systems appear as
a promising approach for the sustained delivery of amine-based
drugs, such as dopamine and related mono-amine neurotrans-
mitters of high biomedical interest and could be also extended
to hydroxylated drugs, such as steroids or taxol among others. In
this sense, current work is being carried out in order to conju-
gate such important drugs.
gel
cleavage
drug
release
self-assembly
gelator
prodrug
hydrogel
solution
trypsin
free gelator
(NMR silent)
gel fiber
enzyme degradation +
self-immolation
free drug
Scheme 5. Schematic representation of the release of a model drug from the hydrogels.
Figure 2. A) Evolution of 1H NMR aromatic region of hydrogel 1 with time. (B) Kinetic profile of amine release.
enzyme. In those cases, the gels remained intact for at least one
week and integral values did not change during this time. As can be
seen in Figure 2B, release of amines shows a typical enzymatic
product versus time profile with identical slopes, confirming the
reliability of the sequence of events (disassembly–enzymatic
cleavage–self-immolation).
4. Experimental
4.1. General
Fmoc-Lys (tBoc)-OH, 4-aminobenzyl alcohol, 2-ethoxy-1-
ethoxycarbonyl-1,2-dihydroquinoline
(EEDQ),
4-nitrophenyl
Furthermore, an important issue in matrix-controlled drug de-
livery systems is the burst release of the drug. Burst release may
have negative effects in some cases, as for instance, increased
toxicity, shorter half-life of the drug in vivo or a shortened release
profile requiring more frequent dosing.17 In the present case, the
release profiles shown in Figure 2 clearly reveal the absence of this
effect. Moreover, it is important to note that the concentration of
trypsin used in these in vitro experiments (2.6 mg/mL) is much
higher than physiological concentration (trypsin in human serum
chloroformate, and Trypsin Type IX-S from porcine pancreas
(liophilized powder, 14,712 units/mg solid, 23.3 kDa) were pur-
chased from Sigma–Aldrich and were used without further
purification.
4.1.1. Synthesis of hydrogelators
1 and 2 (Scheme 4). 3.0 g,
(6.4 mmol) of Fmoc-Lys (tBoc)-OH was dissolved in 30 mL of dry
DME under nitrogen and then 1.89 g (7.68 mmol) of EEDQ were
added at 0 ꢀC (see Scheme 4). After stirring for 1 h, 0.94 g
(7.68 mmol) of 4-aminobenzyl alcohol was added dropwise and
then stirring continued for 24 h. Then, the solvent was evaporated
under vacuum to give a solid, which was washed thoroughly with
ether and filtered off. After drying, 3.66 g of crude compound 3
were obtained as a white solid, which was used in the next step
without further purification.
covers a range between 112 and 637 m
g/l in healthy patients),18
therefore, under an hypothetical in vivo experience, the present
system would show prolonged, sustained release.
3. Conclusion
Prodrug models with excellent hydrogelating properties (with
minimum gelator concentrations of 0.1 w/v%) have been syn-
thesized and have been shown to degrade through the action of
trypsin provoking a self-immolative degradation and release of
the drug model fragment. It has to be noted that the preparation
of gelators containing a self-immolative linker constitutes an
interesting alternative to previous related work because the
Then 0.5 g (0.87 mmol) of crude 3 were dissolved in 15 mL dry
THF at 0 ꢀC under nitrogen. To that solution, 0.10 g (1.3 mmol) of
freshly distilled pyridine and 0.26 g (1.3 mmol) of 4-nitrophenyl
chloroformate were added dropwise and stirring continued for 16 h
at room temperature (see Scheme 3). Then, 0.11 g (1.1 mmol) of
freshly distilled triethylamine were added in first place followed by
dropwise addition of 1.2 mmol of the corresponding amine

´
J.A. Saez et al. / Tetrahedron 66 (2010) 2614–2618
2618
(benzylamine or phenethylamine) dissolved in THF. The mixture
was left stirring at room temperature for 16 h. Then the solvent was
removed at reduced pressure and the yellow solid obtained was
washed with methanol and filtered off. Compounds 4 and 5 were
obtained as white solids after column chromatography purification
(Silica gel, CH2Cl2–MeOH 99:1 to 97:3) with a 67% and 63% yield,
respectively. Finally, the tBoc protecting group was removed by
dropwise addition of 8.5 equiv of trifluoroacetic acid to a solution of
4 or 5 in 30 mL dry CH2Cl2 at 0 ꢀC followed by stirring for 8 h at
room temperature. After that, the solvent was removed at reduced
pressure to yield quantitatively compounds 1 or 2 as off-white
yellowish solids.
Supplementary data
Supplementary data associated with this article can be found in
the online version, at doi:10.1016/j.tet.2010.02.033.
References and notes
1. (a) Lee, K. Y.; Mooney, D. J. Chem. Rev. 2001, 101, 1869–1880; (b) Brandon, V. S.;
Shahana, Z. F.; Omar, Z. F.; Ali, K.; Nicholas, A. P. Adv. Mater. 2009, 21, 3307–
3329; (c) Yu, L.; Ding, J. Chem. Soc. Rev. 2008, 37, 1473–1481; (d) Gupta, P.;
Vermani, K.; Garg, S. Drug Discovery Today 2002, 7, 569–579; (e) Peppas, N. A.;
Bures, P.; Leobandung, W.; Ichikawa, H. Eur. J. Pharm. Biopharm. 2000, 7, 27–46.
2. Weis, R. G.; Terech, P. Molecular Gels: materials with Self-Assembled Fibrillar
Networks; Springer: Dordrecht, 2006.
For a detailed characterization of compounds, see Supplementary
data.
3. (a) Estroff, L. A.; Hamilton, A. D. Chem. Rev. 2004, 104, 1201–1217; (b) Sangeetha,
N. M.; Maitra, U. Chem. Soc. Rev. 2005, 34, 821–836; (c) Hirst, A. R.; Escuder, B.;
Miravet, J. F.; Smith, D. K. Angew. Chem., Int. Ed. 2008, 47, 8002–8018; (d)
Dankers, P. Y. W.; Meijer, E. W. Bull. Chem. Soc. Jpn. 2007, 80, 2047–2073; (e)
Banerjee, S.; Das, R. K.; Maitra, U. J. Mater. Chem. 2009, 19, 6649–6687; (f)
Menger, F. M.; Caran, K. L. J. Am. Chem. Soc. 2000, 122, 11679–11691; (g) Bom-
mel, K. J. C. v.; Pol, C. V. D.; Muizebelt, I.; Friggeri, A.; Heeres, A.; Meetsma, A.;
Feringa, B. L.; Esch, J. v. Angew. Chem., Int. Ed. 2004, 43, 1663–1667; (h) Sree-
nivasachary, N.; Lehn, J. M. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 5938–5943; (i)
Yoshimura, I.; Miyahara, Y.; Kasagi, N.; Yamane, H.; Ojida, A.; Hamachi, I. J. Am.
Chem. Soc. 2004, 126, 12204–12205; (j) Estroff, L. A.; Hamilton, A. D. Angew.
Chem., Int. Ed. 2000, 39, 3447–3450.
4. Zhao, F.; Ma, M. L.; Xu, B. Chem. Soc. Rev. 2009, 38, 883–891.
5. (a) Vemula, P. K.; Cruikshank, G. A.; Karp, J. M.; John, G. Biomaterials 2009, 30,
383–393; (b) Murdan, S.; Andrysek, T.; Son, D. Int. J. Pharm. 2005, 300, 113–124;
(c) Friggeri, A.; Feringa, B. L.; Esch, J. v. J. Controlled Release 2004, 97, 241–248.
6. Xing, B.; Yu, C. W.; Chow, K. H.; Ho, P. L.; Fu, D.; Xu, B. J. Am. Chem. Soc. 2002,
124, 14846–14847.
4.1.2. Preparation of gels. Typically, a given amount of the spec-
ified compound (0.1–0.5 wt/v%) was weighted in a 2 mL glass
vial and was dissolved in 0.2 mL with EtOH (methanol-d4 in the
1H NMR experiments) and 0.8 mL of NaHCO3 buffer (0.1 M, pH
8.3; D2O was used in the 1H NMR experiments), which con-
tained the required amount of trypsin. The aqueous solution was
quickly poured over the ethanolic solution. Immediately, an off-
white to yellowish opaque gel was obtained. A gel was consid-
ered being formed when no gravitational flux was observed
upon vial inversion.
7. Yang, Z.; Gu, H.; Zhang, Y.; Wang, L.; Xu, B. Chem. Commun. 2004, 208–209.
8. (a) Gao, Y.; Kuang, Y.; Guo, Z. F.; Guo, Z.; Krauss, I. J.; Xu, B. J. Am. Chem. Soc.
2009, 131, 13576–13577; (b) Bhuniya, S.; Seo, Y. J.; Kim, B. H. Tetrahedron Lett.
2006, 47, 7153–7156.
9. Bommel, K. J. C. v.; Stuart, M. C. A.; Feringa, B. L.; Esch, J. v. Org. Biomol. Chem.
2005, 3, 2917–2920.
10. (a) Rosseboom, M.; Commandeur, J. N. M.; Vermeulen, N. P. E. Pharmacol. Rev.
2004, 56, 53–102; (b) Trouet, A.; Passioukov, A.; Derpoorten, A. M. v.; Abar-
ca-Quin˜ones, J.; Baurain, R.; Lobl, T. J.; Oliya, C.; Shochat, D.; Dubois, V. Cancer
Res. 2001, 61, 2843–2846; (c) Curran, S.; Murray, G. I. Eur. J. Cancer 2000, 36,
1621–1630.
4.1.3. Gel melting temperatures. Gel to solution transition temper-
ature (Tgel) was determined by typical ‘inversion-tube-method’. In
a 2 mL glass vial the required amount of gelator was weighted and
the gel was prepared as described in the section above. The vial was
immersed in a water bath up-side-down and slowly heated. The
temperature where the gel melted and flowed to bottom of the vial
was considered as Tgel
.
4.1.4. 1H NMR kinetic measurements. In the 1H NMR measures of
the enzymatic cleavage of 1 and 2 gelators, a pre-acquisition
delay of 20 s before every NMR pulse was introduced in the
experience to avoid errors derived from the interaction between
the gelator degradation products and the gel fibers (which are
NMR silent).
11. (a) Sinha, V. R.; Kumria, R. Eur. J. Pharm. Sci. 2003, 18, 3–18; (b) Chourasia, M. K.;
Jain, S. K. J. Pharm. Pharmaceut. Sci. 2003, 6, 22–66.
12. Carl, P. L.; Chakravarty, P. K.; Katzenellenbogen, J. A. J. Med. Chem.1981, 24, 479–480.
13. (a) Keren, H.; Mikhail, P.; Marina, S.; Richard, A. L.; Carlos, F. B.; Doron, S. Angew.
Chem., Int. Ed. 2005, 44, 716–720; (b) Marina, S.; Doron, S. Chem.dEur. J. 2007,
13, 4523–4528; (c) Groot, F. M. H. d; Albrecht, C.; Koekkoek, R.; Beusker, P. H.;
Scheeren, H. W. Angew. Chem., Int. Ed. 2003, 42, 4490–4494.
14. Sagi, A.; Weinstain, R.; Karton, N.; Shabat, D. J. Am. Chem. Soc. 2008, 130,
5434–5435.
15. (a) Toki, B. E.; Cerveny, C. G.; Wahl, A. F.; Senter, P. D. J. Org. Chem. 2002, 67,
1866–1872; (b) Erez, R.; Shabat, D. Org. Biomol. Chem. 2008, 6, 2669–2672.
16. (a) Smith, A. M.; Williams, R. J.; Tang, C.; Coppo, P.; Collins, R. F.; Turner, M. L.;
Saiani, A.; Ulijn, R. V. Adv. Mater. 2008, 20, 37–41; (b) Thornton, P. D.; Mart, R. J.;
Webb, S. J.; Ulijn, R. V. Soft. Mater 2008, 4, 821–827.
17. Huang, X.; Brazel, C. S. J. Controlled Release 2001, 73, 121–136.
18. Ruddell, W. S.; Mitchell, C. J.; Hamilton, I.; Leek, J. P.; Kelleher, J. Br. Med. J.
(Clin. Res. Ed.) 1981, 283, 1429–1432.
Acknowledgements
We thank the Spanish Ministry of Science and Innovation (Grant
CTQ2009-13961) and Universitat Jaume I (Grants P11B2007-11 and
P1$1B2009-42) for funding. We also thank Dr. M. J. Vicent for
helpful discussion.