Letter
pubs.acs.org/ac
Well-Defined Microapertures for Ion Channel Biosensors
Erik Halža,†,‡ Tobias Hedegaard Bro,§,‡ Brian Bilenberg,*,§ and Armağan Koçer*,†
†
Groningen Biomolecular Sciences and Biotechnology Institute & BioMaDe Technology Foundation, Nijenborgh 4, 9747 AG,
Groningen, The Netherlands
§
NIL Technology ApS, Diplomvej 381, DK-2800, Kongens Lyngby, Denmark
S Supporting Information
*
ABSTRACT: Gated ion channels are excitable nanopores in
biological membranes. They sense and respond to different
triggers in nature. The sensory characteristics of these channels
can be modified by protein engineering tools and the channels
can be functionally reconstituted into synthetic lipid bilayer
membranes. The combination of the advances in protein
engineering with electrical and/or optical signal detection
possibilities makes ion channels perfect detection modules for
sensory devices. However, their integration into analytical
devices is problematic due to the instability of lipid bilayers.
Here, we report on developing a stable sensory chip containing a mechanosensitive channel in a Si/SiO2 chip with a 3 μm pore.
Our new fabrication strategy was straightforward. It required only lithography and dry etching for the pore definition and
membrane release and reduced the risk of membrane rupture in the fabrication process. A gated ion channel could be inserted,
with the retention of its function, into the pores of Si/SiO2 chips and be detectable at the single channel level upon activation.
Excitable ion channels in stable small pores can serve as very sensitive detectors of specific molecules.
I
channel into the lipid bilayer. This method is suitable for
generating wafers with closely positioned individual apertures
that can accommodate ion channels and for volume production
of devices, which could be important in future higher-level
integrated sensory devices.
BLM7 has been used previously for ion channel activity
measurements or receptor−analyte binding assays;9−12 however, they are not suitable for sensory devices because the
conventional BLM cups have a large aperture (50−200 μm)
and the lipid bilayer of about 5−20 nm thickness over this
opening is very fragile. They easily break with mechanical and
electrical disturbances.13 One of the solutions to obtain
reproducible and stable BLMs has been forming the BLM
over well-defined sub-10 μm diameter apertures in electrically
inert materials.14−17 The approaches to generate such openings
have been based on different combinations of silicon nitride
and silicon oxide membranes where the pores have been
created by the use of lithography followed by dry etching for
pore definition and wet etching for membrane release or direct
definition of the pores by ion beam milling. The major
drawbacks of these approaches are the materials and fabrication
processes. For instance, pinhole free silicon nitride films are
challenging to make. Wet etching to release the membranes
without membrane rupture is difficult because the membranes
need to be very thin in order to make small apertures. Processes
such as ion beam milling of the apertures are not suitable for
n recent years, there have been significant efforts on the use
of synthetic or biological nanopores in single-molecule
sensing platforms.1−3 The most attractive features of such
systems are the ease of detection as the binding to or passage of
analytes through the pores generates detectable changes in
ionic pore current, no requirement of labeling or surface
attachment of the analytes, and their costs.4 Among the
nanopores, gated ion channels that are natural excitable
nanopores embedded in biological lipid bilayer membranes
stand out for their intrinsic high sensitivity for different stimuli
and for functioning as “on/off” switches. Upon stimulation with
various chemicals or electrical or mechanical triggers,5 these
channels form temporary openings in the membrane and
communicate with the other side of the membrane by allowing
the rapid flux of ions (106−107 ions/s), which can be measured
electrically with a millisecond time resolution.6−8 Furthermore,
some of these channels stay functional in synthetic lipid
bilayers. The combination of the advances in protein
engineering for the modification of the sensitivity of gated
ion channels with electrical and/or optical signal detection
possibilities makes them perfect detection modules for sensory
devices. However, the main challenge in integrating their
excellent detection capacity into the sensory devices is in
creating an environment suitable for the proper functioning of
these channels, i.e., a lipid bilayer membrane. Lipid bilayers are
not robust platforms for use in diagnostic devices. We
developed a straightforward method to generate chips from
silicon on insulator (SOI) wafers with microfabricated 3 μm
apertures. We formed a stable, suspended artificial bilayer lipid
membrane (BLM) across it and could reconstitute a gated-ion
© 2012 American Chemical Society
Received: October 29, 2012
Accepted: December 21, 2012
Published: December 21, 2012
811
dx.doi.org/10.1021/ac303005g | Anal. Chem. 2013, 85, 811−815
Analytical Chemistry
Letter
standard procedures for silicon processing, and they are suitable
for high throughput production as well as for fabricating even
smaller apertures by the use of a thinner device layer and, e.g.,
electron beam lithography, deep UV lithography, or nanoimprint lithography.
The formation of BLM on the aperture and the channel
activities were followed by planar bilayer measurements. The
chip was integrated into a custom-made transparent polymethyl
metacrylate (PMMA) measurement chamber (75 mm × 30
mm × 10 mm) (Figure 2A) as explained in the Material and
production of large numbers of devices. The aim and the
novelty of our work is the introduction of a new microfabrication method for generating well-defined microapertures
and showing its compatibility not only with self-inserting pore
forming peptides but also with bigger, not self-inserting, gatedion channels as sensory elements.
■
RESULTS AND DISCUSSION
To fabricate Si/SiO2 chips with microapertures, silicon on
insulator (SOI) wafers were used. The fabrication of the devices
was a three step dry etch process using UV-lithography on a
SOI wafer with a 300 μm Si handle layer, a 500 nm buried
oxide (BOX) layer, and a 5 μm Si device layer (Figure 1A).
Figure 1. Schematic diagram of the fabrication procedure. The
fabrication process involves only UV lithography and 3 dry etches.
Steps A−C define the through hole in the silicon device layer of the
SOI wafer. Steps D and E define the cavity on the backside of the
wafer. Step F removes the burried oxide layer on the backside of the
silicon membrane thus opening up the through hole. In step G, a 150
nm thermal oxide is grown as electrical insulation. Representative SEM
image of a 3 μm diameter aperture is given in G (Magnification is
15.00 KX). NOTE: In the drawings, the pore is represented by the slit.
Figure 2. Micro-BLMs could be formed over the 3 μm apertures, and
they could accommodate functional alamethicin pores. (A) Schematic
drawing of the PMMA holder for Si/SiO2 chip. (B) A representative
activity of alamethicin pores at 90 mV applied voltage. The triangle
marks the position that is magnified in panel C. (C) Magnified image
of alamethicin activity. Inset shows the conductance histogram of
alamethicin pores.
First, a 3−30 μm hole was defined on the device layer by UVlithography based on AZ5214E resist and transferred into the
device layer by SF6/C4F8 Bosch dry etching using the BOX
layer of the SOI wafer as etch stop (Figure 1B,C). Second, a
750 μm × 750 μm opening was defined on the handle layer side
of the wafer by UV-lithography and by SF6/C4F8 Bosch dry
etching (Figure 1D). The handle layer etch stopped at the BOX
layer of the SOI wafer as well (Figure 1E). Third, the BOX was
removed from the handle layer side of the wafer using a CF4/
CHF3 based reactive ion etch (Figure 1F). Finally, a 150 nm
high quality thermal SiO2 layer was grown (Figure 1G). The
scanning electron microscopy (SEM) image of a 3 μm aperture
of the Si/SiO2 chip is given in Figure 1G. The fabrication
process has been thoroughly tested and has repeatedly been
used to fabricate a full 4 in. wafer containing 36 working devices
with different aperture diameters. All fabrication steps are
Methods. The trans side of the aperture was formed from 0.8%
agarose (in 1 M NaCl). After a drop of buffer (10 mM Tris, 250
mM NaCl, pH = 7.4) was placed on the cis side of the chip, a
lipid bilayer across the 3 μm aperture was formed from
Azolectine (20 mg/mL) lipids using the painting method.18
The bilayer formation was followed by measuring the
membrane capacitance.19,20 We tested the generated microbilayer for the functional insertion of membrane proteins by
reconstituting first a relatively simple, self-inserting peptide
“Alamethicin”. Alamethicin is an antimicrobial peptide that kills
gram-positive bacteria by forming pores in their membranes.21
In this work, 2 μL of alamethicin from a 0.1 mg/mL stock (in
methanol) was added into the buffer on the cis side of the
bilayer. A few minutes after applying voltage across the bilayer
812
dx.doi.org/10.1021/ac303005g | Anal. Chem. 2013, 85, 811−815
Analytical Chemistry
Letter
Figure 3. Functional MscL channels could be inserted into the Si/SiO2 chip with 3 μm aperture. (A) Reconstitution strategy. (B) A representative
activity of MscL channels at −50 mV applied voltage. (The channel openings were shown as upward deflections.) The triangle marks the position
magnified in C. (C) Detailed channel activity. (D) Single channel conductance histogram of MscL. (E) Open time distribution histogram.
(90 mV, sampling at 50 kHz and filtering at 5 kHz), distinct
alamethicin channels could be observed (Figure 2B). A detailed
conductance trace of alamethicin channels and the corresponding conductance histogram is shown in Figure 2C. The bilayers
could be stored at ambient temperature (18 °C) for at least 2
days, and at the end of the period, they were still intact and
could accommodate alamethicin pores.
Next, we tested the suitability of our setup for the
reconstitution of a bigger, gated ion channel, MscL, which
does not insert into the membrane by itself. In nature, MscL
senses tension in a bacterial membrane.22 In its closed state, it
does not allow even the passage of ions. However, upon sensing
the membrane tension, it opens a temporary, large (∼3 nm in
diameter),23 nonselective pore and allows the passage of not
only ions but also small molecules and even small peptides.24
Previously, we have proven its potential as a remote-controlled
nanovalve in liposomal drug delivery vehicles25 and in lipid−
polymer hybrid nanocontainers.26 We also could modify its
specificity toward different triggers. For instance, in one case,
we made it recognize the ambient pH.27 In another case, we
engineered it to respond to the wavelength of the light.28 The
combination of its potential for sensing desired signals with its
functioning in conventional BLMs29 and tethered lipid bilayer
membranes30 makes MscL a suitable candidate for biosensors.
Figure 3A shows a schematic diagram depicting the
procedure for reconstitution of MscL channels into preformed
micro-BLMs on the Si/SiO2 chip. Since MscL does not insert
into lipid bilayers spontaneously, it was first reconstituted into
large unilamellar vesicles (LUVs) as explained in the Material
and Methods. Proteo-LUVs (5 μL) were added into the cis
chamber that had a low salt buffer (10 mM Tris, 250 mM NaCl,
pH = 7.4). The fusion was facilitated by swelling of
proteoliposomes with the help of an osmotic gradient between
the inside and the outside of the proteoliposomes. Since MscL
in LUVs were in their closed conformation, we could not detect
the fusion events in real time. In order to test the functioning of
813
dx.doi.org/10.1021/ac303005g | Anal. Chem. 2013, 85, 811−815
Analytical Chemistry
Letter
vesicles (25 g wet weight) were suspended in 25 mM Tris−
Cl, pH 8.0, and stored at −80 °C. For protein isolation,
typically, 3 g wet weight of membrane vesicles were solubilized
in 30 mL of extraction buffer (10 mM Na2HPO4/NaH2PO4,
pH 8.0, 300 mM NaCl, 35 mM imidazole, 2% (v/v) TritonX100). The solubilized fraction was cleared by centrifugation
(40 000 rpm, 45 min, 4 °C) and applied onto a nickelnitriloacetic acid (Ni-NTA) metal-affinity column. After
washing steps, the pure MscL was eluted with 10 mL of
elution buffer (histidine buffer containing 235 mM histidine).
Typical isolations yielded 3.4 mg of MscL protein (Bradford
assay) which was >98% pure as analyzed by sodium dodecyl
sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE).
In order to reconstitute MscL into lipid membranes, LUVs
were titrated with 0.1 volume of detergent Triton-X100.
Detergent-solubilized pure MscL was added to LUVs at a
protein−lipid ration of 1:75 (w/w), and this mixture was
incubated at 50 °C for 30 min. After this period, one volume of
1.5 M glucose was added to the MscL−LUV mixture. The
detergent was removed by incubating the mixture with 300 mg
wet weight of Biobeads (Bio-Beads SM2, Bio-Rad) at 4 °C for
overnight. This procedure generated uniform proteoliposomes
with 100 nm average diameter containing 750 mM glucose.
Forming a Suspended Lipid Bilayer over a 3 μm Pore.
A horizontal, custom-made transparent polymethyl metacrylate
(PMMA) measurement chamber (75 mm × 30 mm × 10 mm)
was used for the chips with a 3 μm hole. The chamber has a 10
mm × 10 mm central indentation with a pore (7 mm in
diameter) in its center. For BLM formation, the chip was glued
to the center of the chamber at 60 °C, 1 h using
polydimethylsiloxane (PDMS). Next, the chip was inverted,
and the trans compartment was filled with 0.8% melted agarose
(in 1 M NaCl). After the agarose was solidified, the holder was
turned back. A lipid bilayer across a 3 μm hole was formed
using the painting method.8 In these smaller pores, azolectine
(20 mg/mL) was used as a lipid and 10 mM Tris, 250 mM
NaCl, pH = 7.4, was used as the buffer on the cis side.
Electrophysiology Measurements. Ionic currents passing
through the alamethicin or MscL channels were measured by a
Warner Instruments planar lipid bilayer workstation. Data were
amplified (BC-535D Bilayer Amplifier; Warner Instruments,
Hamden, CT), digitized (DigiData 1440A; Axon Instruments,
Foster City, CA), and stored on a computer using the Axoscope
program (version 9.0; Axon Instruments, Foster City, CA).
Data analysis was performed using pClamp suite software
(version 10.2; Axon Instruments, Foster City, CA).
MscL in the bilayer, 5 min after the addition of LUVs, a
positively charged, cysteine specific compound MTSET was
added into the cis compartment (60 mM final concentration)
as an analyte. Binding of MTSET into the Cys residues in the
pore region of MscL opened the channel and allowed ionic
conduction. The channel activity was recorded at −50 mV
constant voltage. Signals were sampled at 50 kHz and filtered at
5 kHz. A representative current trace showing single MscL
channel fluctuations are shown in Figure 3B,C. Bilayers alone
did not give any signal upon treatment with MTSET
(Supporting Information Figure S1). The all points histogram
shown in Figure 3D indicated that, upon binding to MTSET,
the channel started switching between on and off states as
observed before on conventional patch clamp experiments.31,32
The activated channel mainly stayed open often for less than 1
ms and rarely 4 ms (Figure 3E).
In summary, we have developed a new, straightforward
method in order to improve the stability of reconstituted microBLMs by generating microfabricated small apertures on Si/SiO2
chips only from SOI wafers with no SiN deposition and with no
risk of destroying the membranes in the release process. We
could successfully demonstrate that these chips are suitable for
accommodating not only self-inserting, pore-forming peptides
but also bigger, selective, two-state (i.e., open/on and closed/
off) gated ion channels. Sensory chips with gated ion channels
that can be engineered to sense different analytes could offer
opportunities ranging from very early detection of disease
markers to environmental control.
■
MATERIAL AND METHODS
Si/SiO2 Pore Chips. The devices containing the apertures
in which the lipid bilayers are inserted were fabricated from
silicon on insulator (SOI) wafers. The dimensions of the
devices were 10 mm × 10 mm with a 750 μm × 750 μm Si/
SiO2 membrane area. The silicon thickness in the center was 5
μm. A single through hole 3−30 μm in diameter was situated in
the center of this cavity. The device was covered on all edges by
a 150 nm layer of high quality thermally grown SiO2.
Large Unilamellar Vesicle (LUV) Preparation. The
LUVs containing nystatin were prepared as explained before.33
The mixture of azolectine (20 mg/mL) containing 20 mol %
ergosterol and 25 μL/mL of nystatin (from a 2.5 mg/mL stock
solution in dry methanol) in chloroform was vacuum-dried
under reduced pressure. The dried thin lipid film was
rehydrated in 2 mL of sodium phosphate buffer (10 mM
sodium phosphate, 150 mM NaCl, pH 8). The suspension was
vortexed for 5 min and sonicated in a water bath for 2 min. The
lipid vesicle suspension was subjected to six freeze−thaw cycles
in liquid nitrogen and a 50 °C water bath, respectively. Finally,
they were sized using a miniextruder (Avanti Polar Lipids) by
eleven times passage through a polycarbonate filter with a pore
size of 400 nm.
Reconstitution of Mechanosensitive Channel of Large
Conductance into LUVs. MscL channels were isolated as
explained before.34 Briefly, G22C MscL with C-terminal 6Histag31 was expressed in the mscL-knockout Escherichia coli strain
PB104,35 using the pB10a expression vector.22 A 10 L culture
was grown in Luria−Bertani (LB) medium containing 100 μg/
mL ampicillin in a batch fermentor at 37 °C, and the protein
expression was induced by 1 mM isopropyl-β-D-thiogalactopyrapyranoside (IPTG) (Sigma). The cells were passed twice
through a French Press at 15 000 psi, and membrane fractions
were isolated by differential centrifugation.35 Membrane
■
ASSOCIATED CONTENT
S Supporting Information
*
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*Tel: +31 50 3632797 (A.K.). Mobile:+31618183280 (A.K.).
E-mail: [email protected] (A.K.); [email protected] (B.B).
Author Contributions
‡
These authors contributed equally.
Notes
The authors declare no competing financial interest.
814
dx.doi.org/10.1021/ac303005g | Anal. Chem. 2013, 85, 811−815
Analytical Chemistry
■
Letter
(32) Birkner, J. P.; Poolman, B.; Kocer, A. Proc. Natl. Acad. Sci. U.S.A.
2012, 109, 12944−12949.
(33) Woodbury, D. J.; Miller, C. Biophys. J. 1990, 58, 833−839.
(34) Kocer, A.; Walko, M.; Feringa, B. L. Nat. Protoc. 2007, 2, 1426−
1437.
(35) Blount, P.; Sukharev, S. I.; Schroeder, M. J.; Nagle, S. K.; Kung,
C. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 11652−11657.
ACKNOWLEDGMENTS
This work was financially supported by the Seventh Framework
Programme (BISNES 214538 coordinated by Dr. D. Nicolau,
to E.H., T.H.B., and B.B.), the European Research CouncilIdeas Program (ERC-Starting Grant 208814 to A.K.), and The
Netherlands Organisation for Scientific Research (NWO-VIDI
Grant 700.57.427 to A.K.).We acknowledge Dr. G.T. Robillard
for critical reading of the manuscript.
■
REFERENCES
(1) Janshoff, A.; Steinem, C. Anal. Bioanal. Chem. 2006, 385, 433−
451.
(2) Bally, M.; Bailey, K.; Sugihara, K.; Grieshaber, D.; Vörös, J.;
Städler, B. Small 2010, 6, 2481−2497.
(3) Stava, E.; Choi, S.; Kim, H.-S.; Blick, R. H. IEEE Trans.
Nanobiosci. 2012, 11, 169−175.
(4) Howorka, S.; Siwy, Z. Chem. Soc. Rev. 2009, 38, 2360−2384.
(5) Hille, B. Ionic channels of excitable membranes; Sinauer Associates
Inc: Sunderland, MA, 1992.
(6) Hamill, O. P.; Marty, A.; Neher, E.; Sakmann, B.; Sigworth, F. J.
Pflugers Arch. 1981, 391, 85−100.
(7) Hladky, S. B.; Haydon, D. A. Nature 1970, 225, 451−453.
(8) Mueller, P.; Rudin, D. O.; Tien, H. T.; Wescott, W. C. Nature
1962, 194, 979−980.
(9) Bayley, H.; Cremer, P. S. Nature 2001, 413, 226−230.
(10) Guan, X.; Gu, L.-Q.; Cheley, S.; Braha, O.; Bayley, H.
ChemBioChem 2005, 6, 1875−1881.
(11) Howorka, S.; Cheley, S.; Bayley, H. Nat. Biotechnol. 2001, 19,
636−639.
(12) Sackmann, E. Science 1996, 271, 43−48.
(13) Nazaruk, E.; Bilewicz, R.; Lindblom, G.; Lindholm-Sethson, B.
Anal. Bioanal. Chem. 2008, 391, 1569−1578.
(14) Schmidt, C.; Mayer, M.; Vogel, H. Angew. Chem., Int. Ed. Engl.
2000, 39, 3137−3140.
(15) Fertig, N.; Tilke, A.; Blick, R. H.; Kotthaus, J. P.; Behrends, J. C.;
ten Bruggencate, G. Appl. Phys. Lett. 2000, 77 (8), 1218−1220.
(16) McGeoch, J. E. M.; McGeoch, M. W.; Carter, D. J. D.; Shuman,
R. F.; Guidotti, G. Med. Biol. Eng. Comput. 2000, 38 (1), 113−119.
(17) Peterman, M. C.; Ziebarth, J. M.; Braha, O.; Bayley, H.;
Fishman, H. A.; Bloom, D. M. Biomed. Microdevices 2002, 4, 231−236.
(18) Mueller, P.; Rudin, D. O.; Tien, H. T.; Wescott, W. C. J. Phys.
Chem. 1963, 67 (2), 534−535.
(19) Wonderlin, W. F.; Finkel, A.; French, R. J. Biophys. J. 1990, 58,
289−297.
(20) Latorre, R.; Labarca, P.; Naranjo, D. Meth. Enzymol. 1992, 207,
471−501.
(21) Meyer, C. E.; Reusser, F. Experientia 1967, 23, 85−86.
(22) Sukharev, S. I.; Blount, P.; Martinac, B.; Blattner, F. R.; Kung, C.
Nature 1994, 368, 265−268.
(23) Cruickshank, C. C.; Minchin, R. F.; Le Dain, A. C.; Martinac, B.
Biophys. J. 1997, 73, 1925−1931.
(24) van den Bogaart, G.; Krasnikov, V.; Poolman, B. Biophys. J.
2007, 92, 1233−1240.
(25) Kocer, A. J. Liposome Res. 2007, 17, 219−225.
(26) Dudia, A.; Kocer, A.; Subramaniam, V.; Kanger, J. S. Nano Lett.
2008, 8, 1105−1110.
(27) Kocer, A.; Walko, M.; Bulten, E.; Halza, E.; Feringa, B. L.;
Meijberg, W. Angew. Chem., Int. Ed. Engl. 2006, 45, 3126−3130.
(28) Kocer, A.; Walko, M.; Meijberg, W.; Feringa, B. L. Science 2005,
309, 755−758.
(29) Price, C. E.; Kocer, A.; Kol, S.; van der Berg, J. P.; Driessen, A. J.
M. FEBS Lett. 2011, 585, 249−254.
(30) Andersson, M.; Okeyo, G.; Wilson, D.; Keizer, H.; Moe, P.;
Blount, P.; Fine, D.; Dodabalapur, A.; Duran, R. S. Biosens. Bioelectron.
2008, 23, 919−923.
(31) Yoshimura, K.; Batiza, A.; Schroeder, M.; Blount, P.; Kung, C.
Biophys. J. 1999, 77, 1960−1972.
815
dx.doi.org/10.1021/ac303005g | Anal. Chem. 2013, 85, 811−815