A low molecular weight hydrogel which exhibits electroosmotic flow and its use as a bioreactor and for electrochromatography of neutral species

Article abstract of DOI:10.1039/b802155d

A low molecular weight hydrogel which exhibits electroosmotic flow is described, and its use for separation and biocatalytic applications that require passage of a solvent stream through the gel is demonstrated. The Royal Society of Chemistry.

Full text of DOI:10.1039/b802155d

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
A low molecular weight hydrogel which exhibits electroosmotic flow and
its use as a bioreactor and for electrochromatography of neutral speciesw
Shaul Mizrahi, Dan Rizkov, Netanel Hayat and Ovadia Lev*
Received (in Cambridge, UK) 6th February 2008, Accepted 18th March 2008
First published as an Advance Article on the web 14th April 2008
DOI: 10.1039/b802155d
A low molecular weight hydrogel which exhibits electroosmotic
flow is described, and its use for separation and biocatalytic
applications that require passage of a solvent stream through the
gel is demonstrated.
allow for formation of gel slabs before cooling. In order to use the
galactonamide gel, the alkyl chain was shortened to increase
solubility and lower the gelation temperature. The structure of
the hexyl-galactonamide gelator which we used is shown in
Scheme 1.
Small molecules which self-assemble through non-covalent
bonds to form gels in organic solvents and water continue to
attract attention, and their potential applications continue to
increase.1 They are now potentially useful in many fields, some of
which require mass transport and interaction with the environ-
ment such as sensors,2 drug delivery,3 and chemical4 and bio-
chemical5 catalysis. For these applications a means to introduce a
stream of solvent carrying reactants and products is needed.
Needless to say, a pressure driving force would result in defor-
mation or even total destruction of these soft gels. Our group has
recently introduced the use of such a low molecular weight
organogel (LMOG) as a matrix for electrophoresis.6 Amino
acids and peptides were separated in both slab and capillary
form. The most important feature of these gels is that they can be
made to revert reversibly to solutions by elevating the tempera-
ture or by exertion of shear force which allows easy separation of
analytes, reactants and products from the matrix and even direct
mass spectrometry analysis. One drawback is that because the gel
inhibits electroosmosis as noted in our earlier work, only charged
species can be separated, as the solvent itself is not introduced
into the capillary. In this work we show that it is possible to
promote electroosmotically driven solvent flow through the
capillary without destroying the gel. Moreover, electroosmosis
can be promoted even in flat slab form where the external walls
cannot induce electroosmosis.
Scheme 1 N-Hexyl-galactonamide gelator.
While performing separations in slab hydrogels, it was
observed that even uncharged molecules were mobile within
the gel and were fractionated. Thus there appeared to be flow
of the bulk fluid in the gel, a phenomenon that could only be
caused by electroosmosis. As the Teflon walls of the planar cell
were certainly not charged, the only possible remaining source
of charge was the gel itself. Electroosmosis is usually observed
in silica capillaries, the walls of which have a negative charge.
Electroosmosis can be altered by modifying the capillary walls
with charged or uncharged groups or by coating the walls with
polymers to block the charged silanol groups. Our previous
separation work6 relied on electrophoresis in order to intro-
duce charged analytes into the bulk bis(dodecylamido)-cyclo-
hexane organogel as the gel completely inhibited
electroosmosis.
As no clear ionogenic group is present in the gelator, the
gelation process must provide the environment necessary for the
gel to acquire charge. Of all the hydroxyl groups present, the one
adjacent to the amide is the most acidic due to inductive effects,
yet not sufficiently so to explain the electroosmosis. It has been
observed in gels, polymers, and proteins that hydrogen bonding
lowers the pKa of acidic groups by stabilizing the basic form.8 In
the case of the galactonamide gel, the a-hydroxyl is surrounded
not only by the other hydroxyls in the same molecule but also all
the other adjacent molecules in the bundle of gel fibers. This great
potential for hydrogen bonding stabilization of the resulting base
allows for the deprotonation of the gelator molecules. The neutral
molecule mesityl oxide was used as a probe for electroosmosis. Its
electroosmotic mobility was observed at different pH values. The
results are depicted in Fig. 1. Electroosmosis was observable
down to pH 4. A parallel can be drawn between these gel fibers
and the walls of a silica column whose silicic acid precursor has a
pKa of about 9.69 and exhibits electroosmosis down to about pH
2. In both cases, the very same functional group exhibits a
different proton dissociation constant depending on its close
environment, nearby fibrillar charge, and the availability of
As a continuation of our prior work, we searched for a low
molecular weight hydrogel in order to be able to work with
biomolecules. Very few small molecule hydrogelators are applic-
able to electrophoresis as most are charged and would thus move
in the electric field and destroy the gel. For example, an aqueous
N-hexadecyl lactosylamine gel was destroyed when we applied an
electric field. As a potential matrix we selected a class of hydro-
gelators consisting of N-octyl-hexonamides originally introduced
by Furhop et al.7 In order to form organogels for electrophoresis
there must be sufficient time to cast the hot solutions into the slabs
or capillaries before gelation takes place. Octyl-galactonamide
gels at very high temperature (close to 100 1C) which does not
The Institute of Chemistry and the Casali Institute, The Hebrew
University of Jerusalem, Jerusalem 91904, Israel.
E-mail: Ovadia@vms.huji.ac.il
w Electronic supplementary information (ESI) available: Conditions
and results of the buffer experiments, results of the enzyme kinetics.
See DOI: 10.1039/b802155d
ꢀc
This journal is The Royal Society of Chemistry 2008
2914 | Chem. Commun., 2008, 2914–2916

View Article Online
Fig. 3 Effect of the concentration of glucosamine in the buffer on the
mobility of mesityl oxide at 200 V, pH 6, 7 mg mlꢁ1 gelator.
Fig. 1 Effect of pH on mobility of mesityl oxide at 200 V, 7 mg mlꢁ1
gelator. Bars represent standard deviation of three gels.
stabilizing hydrogen bonding. This results in a spread of the
increase of the fibrillar charge over several pH units.
When glucosamine, a compound which should be easily
incorporated into the gel, was added to the buffer, the mobility
did decrease as the positive amine partially negated the negative
charge of the gel. The change was linear with the concentration
of glucosamine as seen in Fig. 3. Ethanolamine, which does not
have a strong affinity to the gel like glucosamine, did not cause a
decrease in electroosmosis. This demonstrates a method of
controlling the EOF. Another way of doing this is to add a
polymer such as polyethylene glycol to the buffer.
The effect of changing the concentration of the gelator on
the mobility of mesityl oxide was also measured. The results at
pH 7 are shown in Fig. 2. A maximum is observed at 9 mg
mlꢁ1. We believe that there is a trade-off between two compet-
ing factors. On the one hand, the more gelator there is, the
more charge accumulates on the fibrillar structure per unit
volume and the electroosmotic mobility should increase. On
the other hand, as the gel becomes more dense the viscosity
increases and the pore volume decreases. Therefore the mobi-
lity decreases.6 The result is the observed local maximum.
In order to rule out the possibility of the electroosmosis being
caused by ions from the buffer being adsorbed onto the gel
surface, the mobility of mesityl oxide was measured in a variety of
buffers. The initial experiments were performed using a wide pH
range citrate–phosphate buffer. Changing the concentration of
the buffer (at the same pH) did not affect the mobility. Changing
to a trisamine based buffer also did not change the mobility. Were
the electroosmosis to arise from adsorbed ions, an increase in
buffer concentration should have increased the amount of ad-
sorbed ions and hence the electroosmotic flow (EOF). If negative
phosphate ions were what caused the electroosmotic flow, repla-
cing them with positive tris ions should have either eliminated the
flow or perhaps reversed its direction if the tris ions were
adsorbed onto the gel. Details of these experiments are available
in the electronic supplementary information (ESI).w
One potential use of the EOF is for the separation of neutral
compounds. Dichlorophenols (DCPs) were chosen as model
compounds to demonstrate this application. 2,3-, 2,6- and
3,5-DCP were separated at pH 4. The retention times are shown
in Table 1. 2,3-, 2,4- and 2,5-DCP were not resolved. 3,4- and 3,5-
DCP were not resolved either. The chlorine location appears to
play a key role in the separation, as only compounds with
different numbers of chlorine atoms adjacent to the hydroxyl
were separated. The hydroxyl group should show affinity to the
gel due to hydrogen bonding. 2,6-DCP, which has two chlorines
adjacent to the hydroxyl, elutes first, suggesting that the chlorine
interferes with the attraction to the gel. 2,3-DCP, which has one
chlorine adjacent to the hydroxyl, eluted second and 3,5-DCP
with no chlorines adjacent to the hydroxyl eluted last. This can
only result from a chromatographic separation based on the
differing interactions with the gel. Were charge to play a role, the
order of the DCPs would have been reversed, as they would be
anions migrating electrophoretically against the EOF and there-
fore the most acidic would elute last. Since the more acidic
compounds eluted first, it must be concluded that charge was not
a factor, which is not surprising at a pH much below the pKa of
the compounds.
Another potential use for the EOF promoting gel is in
enzyme immobilization, be it for sensing or packed bed
bioreactors. Despite the importance of enzymatic reactions
in contemporary materials and the continual search for in-
novative enzyme immobilization systems, the first and only
report of enzyme encapsulation in an LMOG was very
Table 1 Retention times and properties of dichlorophenols
Retention time/min
pKa
2,6-Dichlorophenol
2,3-Dichlorophenol
3,5-Dichlorophenol
55.7
57.9
60.7
7.02
7.44
8.04
Fig. 2 Effect of gelator concentration on mobility of mesityl oxide at
200 V, pH 7. Bars represent standard deviation of three gels (in each
gel at least five distance points were taken as a function of time).
ꢀc
This journal is The Royal Society of Chemistry 2008
Chem. Commun., 2008, 2914–2916 | 2915

View Article Online
increase the substrate concentration to observe a Michaelis–
Menten dependence.
In conclusion, LMOGs based on octyl hexonamides were
found to exhibit electroosmosis, even in slab gels. This is the
first demonstration of electroosmosis in an LMOG. This
phenomenon is both novel and unexpected. Intuitively, elec-
troosmosis could not be driven by uncharged moieties. How-
ever, post priori, the occurrence of EOF can be readily
explained by the supramolecular structure of LMOGs which
imposes environmental constraints which spread the proton
dissociation susceptibility of the very same gelator function-
ality (i.e., the a-hydroxyl) over a wide pH range when it is
exposed to different solid environments. This structure can
similarly explain the somewhat surprising existence of enzy-
matic activity within the gel. The gelator, which like most
LMOG gelators is a surface active agent having distinct
hydrophobic and hydrophilic parts, should have denatured
the enzyme. However, this does not happen due to the
supramolecular structure which binds and hinders the mobility
of the surfactant gelator and thus interferes with its denaturing
activity.
Fig. 4 Electropherograms of pyrogallol (23 min) with N-benzyl-1-
(1-naphthyl)-ethylamine as internal standard (8 min) in (a) a plain gel
capillary; and (b) a capillary with added HRP and hydrogen peroxide.
Capillary effective length 42 cm; driving force 20 kV.
recent.10 Since the presence of electroosmosis in the gel allows
the flow of uncharged compounds through the capillary, it is
possible to use the gel-filled capillary as a packed bed reactor.
To date non-destructive accessibility to the gel interior could
be attained only by diffusion or electrophoresis. The first is a
slow process and the latter is limited to charged species. This
communication expands the possibilities of enzyme immobili-
zation and paves the way for LMOG applications requiring
fast transport of analytes, reactants or products between the
interior of LMOG gels and their surroundings.
The gelator, hexyl-galactonamide was tailored to meet the
needs of two model applications. First it was used as a
stationary phase for electrochromatography. Dichlorophenols
were separated based on the different interactions with the gel
resulting from the location of the chloride groups relative to
the hydroxyl. The gelator was further used as the solid phase
of a packed bed bioreactor hosting HRP inside a fused silica
capillary. Pyrogallol was oxidized by hydrogen peroxide cat-
alyzed by the immobilized HRP. The unique properties of
LMOGs and especially their thixotropy, reversibility and
porosity combined with the accessibility due to electroosmosis
thus provide an environment suitable for many analytical
applications, two of which were demonstrated here.
In order to demonstrate this idea, horseradish peroxidase
(HRP) and hydrogen peroxide were added to the gel, and
pyrogallol was injected into the capillary along with a refer-
ence species N-benzyl-1-(1-naphthyl)-ethylamine. The conver-
sion was calculated from the ratio of the pyrogallol peak to the
reference peak in relation to the ratio in a blank run with no
enzyme. All runs were carried out at pH 5 since at higher pH
the stronger EOF causes some instability in the gel. A run at
20 kV is shown in Fig. 4. The top window shows a run without
HRP. The first peak is the N-benzyl-1-(1-naphthyl)-ethyl-
amine internal standard, and the second is pyrogallol which
amounts to 46% of the internal standard peak. In the lower
window is the run with HRP in which the pyrogallol peak has
been reduced to 8% of the internal standard peak. This
amounts to an 83% conversion. The oxidation of pyrogallol
to purpurogallin is a multi-step reaction with several inter-
mediates.11 Because the reaction is taking place under electro-
phoretic conditions, the products are formed at different
locations along the capillary and therefore do not emerge as
chromatographic peaks, rather they are smeared through the
capillary as seen by the rise in baseline starting a few minutes
before the pyrogallol peak. No reaction (or decrease of the
pyrogallol peak) was observed for thermally denatured HRP.
The oxidation of pyrogallol is first order with respect to
each of pyrogallol, peroxidase and peroxide at low substrate
concentrations.12 The effective residence time of pyrogallol in
the reactor was changed by varying the voltage and/or the
capillary length, keeping all other parameters constant. The
data fit the expected logarithmic relationship for first order
reactions (R2 ¼ 0.991). The observed rate constant for the
We gratefully acknowledge the financial support of the
Israel Ministry of Science.
Notes and references
1. M. George and R. G. Weiss, Acc. Chem. Res., 2006, 39, 489;
N. M. Sangeetha and U. Maitra, Chem. Soc. Rev., 2005, 34, 821;
M. de Loos, B. L. Feringa and J. van Esch, Eur. J. Org. Chem.,
2005, 3615.
2. P. Mukhopadhyay, Y. Iwashita, M. Shirakawa, S. Kawano,
N. Fujita and S. Shinkai, Angew. Chem., Int. Ed., 2006, 45, 1592.
3. A. Friggeri, B. L. Feringa and J. Van Esch, J. Controlled Release,
2004, 97, 241.
4. Q. Wang, Z. Yang, X. Zhang, X. Xiao, C. K. Chang and B. Xu,
Angew. Chem., Int. Ed., 2007, 46, 4285.
5. S. Kiyonaka, K. Sada, I. Yoshimura, S. Shinkai, N. Kato and
I. Hamachi, Nat. Mater., 2004, 3, 58.
6. S. Mizrahi, J. Gun, Z. G. Kipervaser and O. Lev, Anal. Chem.,
2004, 76, 5399.
7. J. H. Furhop, P. Schnieder, E. Boekema and W. Helfrich, J. Am.
Chem. Soc., 1988, 110, 2861.
8. M. A. Porter, J. R. Hall, J. C. Locke, J. H. Jensen and
P. A. Molina, Proteins, 2006, 63, 621.
9. CRC Handbook of Chemistry and Physics, ed. R. C. West and
M. J. Astle, CRC Press, Boca Raton, FL, 63rd edn, 1982.
10. Q. Wang, Z. Yang, L. Wang, M. Ma and B. Xu, Chem. Commun.,
2007, 1032.
enzyme and peroxide concentrations used is 0.0677 minꢁ1
.
11. K. Miyazaki, S. Arai, T. Iwamoto, T. Masaru and A. Tomoda,
Tohoku J. Exp. Med., 2004, 203, 319.
12. M. K. Mahanti, Afinidad, 1991, 48, 59.
Fig. 2 of the ESIw shows the characteristic logarithmic con-
version–retention time dependence. We did not attempt to
ꢀc
This journal is The Royal Society of Chemistry 2008
2916 | Chem. Commun., 2008, 2914–2916