Analytical Biochemistry 424 (2012) 12–20

Contents lists available at SciVerse ScienceDirect

Analytical Biochemistry
journal homepage: www.elsevier.com/locate/yabio

Harnessing aptamers for electrochemical detection of endotoxin

Sung-Eun Kim 1, Wenqiong Su 1, MiSuk Cho, Youngkwan Lee, Woo-Seok Choe ⇑
School of Chemical Engineering, Sungkyunkwan University, Suwon 440-746, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: Lipopolysaccharide (LPS), also known as endotoxin, triggers a fatal septic shock; therefore, fast and accu-
Received 1 January 2012 rate detection of LPS from a complex milieu is of primary importance. Several LPS affinity binders have
Received in revised form 7 February 2012 been reported so far but few of them have proved their efficacy in developing electrochemical sensors
Accepted 8 February 2012
capable of selectively detecting LPS from crude biological liquors. In this study, we identified 10 different
Available online 25 February 2012
single-stranded DNA aptamers showing specific affinity to LPS with dissociation constants (Kd) in the
nanomolar range using a NECEEM-based non-SELEX method. Based on the sequence and secondary struc-
Keywords:
ture analysis of the LPS binding aptamers, an aptamer exhibiting the highest affinity to LPS (i.e., B2) was
Lipopolysaccharide
Aptamer
selected to construct an impedance biosensor on a gold surface. The developed electrochemical aptasen-
NECEEM-based non-SELEX sor showed excellent sensitivity and specificity in the linear detection range from 0.01 to 1 ng/mL of LPS
Electrochemical aptasensor with significantly reduced detection time compared with the traditional Limulus amoebocyte lysate (LAL)
assay.
Ó 2012 Elsevier Inc. All rights reserved.

Lipopolysaccharide (LPS2 or endotoxin) is a major component of Aptamers are single-stranded oligonucleotides able to bind to a
the outer membrane of gram-negative bacteria [1,2] and induces a variety of targets including proteins, small molecules, and living
strong immune response on its internalization into mammalian cells. cells with high affinity and specificity [11,12]. Ease of synthesis,
When the human circulatory system is infected by LPS, various facile chemical modification, and integration into different analyt-
inflammatory cytokines are overexpressed by activation of the innate ical schemes have rendered their application in many areas of bio-
immune system and these mediators may render the patient to expe- analytical and biomedical sciences.
rience LPS-induced symptoms of septic shock and mortality [3]. Target recognizable aptamers have often been selected from
Due to its toxicity, the detection and removal of even small synthetic combinational nucleic acid libraries in general through
amounts of LPS in therapeutic products and blood purification an in vitro selection procedure known as systematic evolution of li-
are a primary concern in the treatment of septic shock. Several gands by exponential enrichment (SELEX) [13,14]. SELEX involves
LPS adsorbents such as activated carbon [4] and polymyxin B repetitive rounds comprising the following two steps: (1) fraction-
[5,6] have been developed for the removal of LPS. Nevertheless, ation of target-bound aptamers from free aptamers and (2) PCR
the use of these nonspecific LPS adsorbents usually gives rise to amplification of the selected aptamers. Typically 8 to 15 iterative
the significant loss of target proteins and/or other nonproteina- cycles of partitioning and amplification are required prior to select-
ceous products (e.g., plasmid DNA) during the removal process ing aptamers with high and specific affinity to a target from an
[7–10]. Thus ligands capable of specifically recognizing LPS are initial library [15,16]. As a result, SELEX-based conventional parti-
highly required. tioning methods take several months to complete. Moreover, these
processes often lead to aptamers that bind to the surface of the
⇑ Corresponding author. Fax: +82 31 290 7272. filters or chromatographic supports used in partitioning, rather
E-mail address: [email protected] (W.-S. Choe).
than to the target [17–19].
1
These authors equally contributed to this work. An alternative screening method termed nonequilibrium capil-
2
Abbreviations used: LPS, lipopolysaccharides or endotoxin; Kd, equilibrium disso- lary electrophoresis of equilibrium mixtures (NECEEM) was proven
ciation constant; NECEEM, nonequilibrium capillary electrophoresis of equilibrium to have the partitioning efficiency of 2 orders of magnitude higher
mixtures; SELEX, systematic evolution of ligands by exponential enrichment; LAL, than that of the conventional methods [20–23]. The outstanding
Limulus amoebocyte lysate; EIS, electrochemical impedance spectroscopy; PCR,
polymerase chain reaction; BSA, bovine serum albumin; FAM, fluorescent dye; DTT,
partitioning capabilities of NECEEM have been successfully used
1,4-dithiothreitol; MCH, 6-mercapto-1-hexanol; QCM, quartz crystal microbalance; in the aptamer selection procedure that does not include intermedi-
EU, enzyme unit; EDTA, ethylenediaminetetraacetic acid; RT-PCR, real-time poly- ate amplification steps between selection rounds (i.e., NECEEM-
merase chain reaction; SPR, surface plasmon resonance; PBS, phosphate-buffered based non-SELEX) [21]. The obviation of repetitive steps of PCR
saline; RU, resonance unit; ka, association constant; kd, dissociation constant; CV,
significantly accelerates the screening process and excludes the
cyclic voltammetry; Rs, solution resistance; C, double layer capacitance; W, Warburg
diffusion resistance; Ret, electron transfer resistance; SAM, self-assembled monolayer. quantitative errors associated with the exponential nature of PCR

0003-2697/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.ab.2012.02.016
 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20 13

amplification. This makes non-SELEX a useful tool for selection and Korea). Limulus amoebocyte lysate assay kit (C1500) was pur-
characterization of aptamer binders to different targets [22]. chased from CapeCod Inc. (East Falmouth, USA). 1,4-Dithiothreitol
Several aptamers have been isolated against LPS for different (DTT) and 6-mercapto-1-hexanol (MCH) were purchased from Sig-
purposes (e.g., neutralization of the toxicity of LPS [24], inhibition ma-Aldrich. All solutions were prepared using ultrapure water pro-
of the growth of Escherichia coli [25], and protection of mice from duced by the Satorious Arium 61316 RO system (Sartorius Stedim
LPS-induced sepsis [26]). However, few aptamer has been tested Biotech, Aubagne Cedex, France).
for its specificity toward LPS in the context of developing an elec- All the electrochemical experiments were conducted on VSP
trochemical sensor capable of selectively detecting LPS from crude Potentiostat (Bio-logic Science Instruments, USA) and the data
biological liquors. were recorded with VSP EC-Lab software. A conventional three-
Electrochemical biosensors with the use of modified electrodes electrode system with a gold disk electrode of 2 mm diameter
via self-assembled monolayers (SAM) have recently attracted (CH Instruments Inc., USA) as the working electrode, an Ag/AgCl
much attention [27–29]. Aptamers can be easily modified at their as the reference electrode, and a platinum plate as the counterelec-
50 - or 30 -end with various functional groups (e.g., amine, carboxyl, trode was employed for all the electrochemical analyses. In EIS
and thiol groups) to facilitate scaffolding of aptamer-modified SAM measurement, a sine wave of 10 mV amplitude was applied to
on a gold electrode surface [30]. Among various electrochemical the working electrode from 0.1 Hz to 100 kHz. The obtained EIS
methods, electrochemical impedance spectroscopy (EIS) is a rapid, data were calculated by ZSimpWin EIS DATA analysis software
sensitive, and nondestructive technique for characterizing the elec- (Perkin-Elmer, Version 2.00). Aptamer coverage on the gold surface
trochemical property changes at the electrode–electrolyte inter- was measured using Quartz Crystal Microbalance (QCM, Princeton
faces and has widely been used as an effective tool for detecting Applied Research, USA) equipped with 10 MHz AT-cut crystals
nanogram quantities of biomolecules including bacteria, proteins, coated with polished gold disk (5 mm in diameter) on both sides
DNA, antigens, and antibodies on electrode surfaces [31–33]. under the same conditions used for the gold electrode modification
In the present study, we have used NECEEM-based non-SELEX with the aptamer. The frequency change of 1 Hz corresponds to the
to obtain high affinity single-stranded DNA (ssDNA) aptamers to binding of 1.3 ng of aptamer to the gold disk.
LPS with dissociation constants (Kd) in the nanomolar range. The
screened aptamers were thiolated at their 50 -end and used to con- Selection of LPS–affinity aptamer by NECEEM-based non-SELEX
struct a biosensor where aptamer molecules function as a detec-
tion probe to mediate electrochemical impedance changes on All the screening experiments were conducted using the P/ACE
their interaction with LPS. The developed electrochemical aptasen- MDQ capillary electrophoresis system (144003, Beckman Coulter)
sor showed a linear detection range from 0.01 to 1 ng/mL of LPS with a LIF detector (727621) following injection of aptamer–LPS
comparable to the traditional Limulus amoebocyte lysate (LAL) as- mixture at 2 psi for 13 s to a bare fused-silica capillary set at
say but with significantly reduced detection time. Furthermore, the 30 kV and 15 °C. Prior to each run, the capillary was rinsed suffi-
developed aptasensor showed negligible cross-binding reactivity ciently with each of the following solutions: 100 mM HCl,
to various biomolecules such as pDNA, RNA, proteins, saccharides, 100 mM NaOH, ultrapure water, and running buffer (50 mM Tris-
and/or lipids which often coexist with LPS in many bioprocessing acetate, pH 8.2).
liquors. The FAM-labeled aptamer library (50 -FAM-CTTCTGCCCGCCT-
CCTTCC-(45N)-GGAGACGAGATAGGCGGACACT-30 ) was diluted in
Materials and methods running buffer, heated in a thermal cycler (170-9703, Bio-Rad Lab-
oratories Inc., Hercules, USA) to 95 °C for 1 min and cooled down to
Materials 15 °C prior to its injection to the capillary to restore the secondary
structure of each ssDNA aptamer in the library. For the screening of
LPS from E. coli 055:B5 (L4524) was purchased from Sigma-Al- LPS affinity aptamers, aptamer–LPS mixture was prepared by mix-
drich (USA). Bare fused-silica capillaries (TSP075375) with a length ing 10 lL of 50 lM library with 10 lL of 4 mg/mL (i.e., 2 lM or
of 80 cm and inner and outer diameters of 75 and 363 lm, respec- 2  107 EU/mL) of LPS. Following 1 h incubation, 150 nL of the apt-
tively, were purchased from Polymicro (Polymicro Technologies, amer–LPS mixture was injected to the capillary. During each round
USA). D-(+)-Glucose (G7021), sucrose (84097), cholesterol of screening, a fraction of the aptamer–LPS complex (designated as
(C3045), Baker’s yeast RNA (R6750), and bovine serum albumin Fraction C) was directed to an outlet vial containing 20 lL of 2 mg/
(BSA, A7906) were from Sigma (St. Louis, MO, USA). Baker’s yeast mL of fresh LPS solution. Following 1 h incubation, 150 nL of the
RNA was further purified using LPS removal solution (Sigma, Fraction C-LPS mixture was injected to the capillary to start a next
E4274) prior to its use in the experiment. Purified plasmid pDNA round of screening. After three rounds of screening, Fraction C from
pcDNA3.1D was prepared according to the protocol described else- each round of screening (i.e., Fraction C1, C2, or C3) was PCR-
where [34]. amplified and used for assessing the bulk affinity of screened apta-
FAM-labeled random single-strand library (50 -FAM-CTT CTG mers to LPS.
CCC GCC TCC TTC C-(45N)-GGA GAC GAG ATA GGC GGA CAC T-
30 ), nonlabeled or FAM-labeled forward primer (50 -CTT CTG CCC PCR amplification and purification of ssDNA aptamer
GCC TCC TTC C-30 ), and nonlabeled or biotin-labeled reverse primer
(50 -AGT GTC CGC CTA TCT CGT CTC C-30 ) were all synthesized by The screened aptamers in Fractions C1 to C3 were PCR-ampli-
Integrated DNA Technologies Co., Ltd. (Coraville, USA). Magnetic fied by Taq polymerase with the use of nonlabeled forward and
streptavidin beads (112.05D) were purchased from Invitrogen biotin-labeled reverse primers in a thermocycler for 30 thermal cy-
(Invitrogen Dynal AS, Oslo, Norway). SYBR Green supermix cles with each cycle comprising denaturation at 94 °C for 30 s,
(1708882) was obtained from Bio-Rad (Hercules, USA). Plasmid annealing at 65 °C for 10 s, and extension at 72 °C for 10 s. The
Mini purification kit (004055) and all the PCR reagents including amplified biotinylated dsDNA was used as a template for the next
Taq polymerase (CMT 1001) were purchased from LaboPass (Cos- asymmetric PCR conducted under the same conditions as above
mo Genetech, Korea). QIA quick PCR purification kit (28106) was using FAM-labeled forward primer only. The resulting PCR product
purchased from Qiagen (Hilden, Germany). TA cloning kit was incubated with streptavidin-coated magnetic beads for 15 min
(756377) was obtained from Invitrogen (CA, USA). DH5a-compe- in binding buffer (100 mM Tris-HCl, 1 mM EDTA, pH 7.5) in order
tent cells (15045) for cloning were obtained from Intronbio (Seoul, to remove dsDNA. The supernatant containing FAM-labeled
14 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20

aptamers was collected and further purified using a PCR purifica- water and ethanol for 1 min each, and dried under nitrogen. Apt-
tion kit. The concentration of purified ssDNA was analyzed by amer molecules were immobilized on the ligand channel by inject-
RT-PCR (DNA Engine Peltier Thermal Cycler, Bio-Rad) using SYBR ing 200 nM of thiolated B2 in PBS to the channel followed by 4 h
Green supermix. incubation. Subsequently the chip was immersed in 1 mM MCH/
PBS for 2 h to block the aptamer unloaded chip surfaces on the li-
Cloning and sequencing gand and reference channels. After washing the chip surface with
PBS, each LPS/PBS solution prepared at varying concentrations
The aptamers in Fraction C3 were PCR-amplified by Taq poly- (0.005, 0.01, 0.02, and 0.04 mg/mL) was injected sequentially into
merase with the use of nonlabeled forward and reverse primers. both channels for 240 s at a flow rate of 20 lL/min. Between each
The amplified products were ligated into PC 2.1 vector provided LPS injection, the channels were washed with PBS for 180 s. LPS–
in TA cloning kit and the ligates were transformed into DH5a-com- aptamer interaction was monitored based on the difference of re-
petent cells. Plasmids extracted from 32 transformants using a sponse unit (DRU) between the ligand channel and reference chan-
plasmid purification kit were sent for sequencing. nel. The DRU data were fitted to a kinetic titration 1:1 interaction
model using CLAMP (Center for Biological Interaction Analysis, UT,
Determination of dissociation constant (Kd) USA). The association and dissociation constants (i.e., ka and kd,
respectively) were determined from the fit to obtain the equilib-
Cloned aptamers or the aptamers in Fractions C1 to C3 were rium dissociation constant (Kd).
amplified, purified, and quantified as described above. To assess
the binding affinity of screened aptamers to LPS, 10 lL of 60 nM Detection of LPS using aptamer-modified gold electrodes
of each aptamer was mixed with 10 lL of 4 mg/mL LPS in running
buffer for 1 h and 150 nL of the incubated mixture was injected to An aptamer B2 exhibiting the highest affinity to LPS and a con-
the capillary. The CE running condition was the same as described trol aptamer C2 lacking affinity to LPS were modified by Integrated
above. Assessment of Kd was conducted based on the determina- DNA Technologies to introduce a disulfide modifier at the 50 end of
tion of peak areas accounting for LPS–DNA complex (A1), DNA dis- each. In order to enable self-assembly of thus obtained aptamers
sociated from complex (A2), and free DNA (A3) as shown in Eq. (1) on a gold surface, the disulfide group was reduced by DTT for 1 h
[23] and Fig. 1 (adopted and modified from Ref. [23]). at room temperature [35] to give 50 sulfurhydryl-terminated apt-
  amer (50 -HS–(CH2)6–). Following the reaction, the thiolated apt-
½LPStot 1 þ A1AþA
3
2
 ½DNAtot amer was purified by centrifugal filter (UFC900324, Millipore) at
Kd ¼   ; ð1Þ
A1 þA2
1 þ A3 10,000g for 30 min (3) and the purified aptamer was diluted by
PBS (10 mM, pH 7.4) to give an aptamer concentration of
where [LPS]tot and [DNA]tot are the total concentrations of the LPS 11.5 nM based on UV absorption at 260 nm.
and aptamer, respectively. Gold disk electrodes were prepared according to the protocol
described elsewhere [36] and then electrochemically activated by
Surface plasmon resonance (SPR) analysis to study the interaction cyclic voltammetry (CV) in 0.5 M H2SO4 aqueous solution from 0
between thiol-modified, immobilized aptamer B2 and LPS to 1.5 V at a rate of 100 mV/s for 20 cycles. Afterward, the electro-
polished electrodes were dried and immediately immersed in the
To verify if the high affinity binder B2 keeps its LPS recognition PBS containing thiolated aptamer for 3 h at room temperature.
ability following thiol modification at the 50 end of the aptamer The aptamer-modified gold electrode was incubated in 1 mM
molecules and their subsequent immobilization onto the gold elec- MCH aqueous solution for 1 h at room temperature [37]. EIS mea-
trode, SPR analysis was performed to determine the equilibrium surement in 2 mM FeðCNÞ6 3=4 ð1=1Þ=PBS was conducted to elec-
dissociation constant (Kd) for thiolated/immobilized B2–LPS inter- trochemically assess interaction of the aptamer-modified gold
action. All SPR experiments were conducted on a Reichert electrode with varying amounts of LPS. As a control, EIS measure-
SR7500DC instrument and two-channel (i.e., a ligand and a refer- ment was repeated for other biomolecules (e.g., pDNA, RNA, pro-
ence channel, respectively) gold chips (13206060, Reichert, NY, teins, saccharides, and lipids) often found in biological liquors
USA) using PBS (10 mM, pH 7.4) as running buffer. The bare chips where LPS is likely to exist in order to evaluate the specificity of
were cleaned in piranha solution (H2SO4/H2O2 (3/1)) for 15 min, the sensor. Unless otherwise noted, all the EIS measurements were
thoroughly washed with ultrapure water, sonicated in ultrapure conducted after immersing the electrodes in the corresponding
solutions for 15 min.

Results and discussion

Selection of LPS affinity aptamers

In order to screen aptamers showing high affinity to LPS using

NECEEM, it is important to define a collection window where the
LPS–aptamer complex is likely to be populated. For this, migration
times of free LPS and free ssDNA library were determined under
the CE running condition. Since LPS gave no apparent UV or fluo-
rescence signal, its migration time was estimated by collecting
LPS in six fractions (from 0 to 48 min, 8 min/fraction) and quanti-
fying LPS in each fraction by LAL assay. It was found that approxi-
mately 80% of the injected LPS molecules were distributed in the
Fig.1. A schematic of a putative electropherogram from NECEEM following the
injection of ssDNA aptamer library and LPS mixture. Al, A2, and A3 represent the
fractions collected from 8 to 24 min following their injection. The
peak areas for LPS–DNA complex, DNA dissociated from the complex, and free DNA, migration time of FAM-labeled free ssDNA library, monitored by
respectively. LIF, was found to be around 34 min. When the equilibrium mixture
 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20 15

comprising 25 lM FAM-labeled ssDNA library and 2 mg/mL of LPS B1 dominated Fraction C3 with 23 copies out of 32 ssDNA aptamer
was injected to the capillary, several low intensity peaks, in addi- sequences.
tion to the free ssDNA peak, were observed on the electrophero- The homology of the LPS binding aptamers listed in Table 1 was
gram between 21 and 27 min (Fig. 2). The migration window for analyzed by DNAman and found to be around 49.4% (Fig. 5A). Sur-
these peaks was located between those for free LPS and ssDNA li- prisingly, the LPS binders B1, B2, B3, and B4 showed more than 98%
brary, accounting for the presence of putative LPS-aptamer com- identity in their nucleotide sequences, indicating that these four
plexes in the equilibrium mixture. To ensure the efficient binders essentially evolve from a highly conserved sequence (i.e.,
recovery of ssDNA complexed with LPS, all the aptamers present TAGCCGGATCGCGNTGGCCAGATGATATAAAGGGTCANCCCCCCN,
in an aptamer collection window from 20 to 31 min were directed where N denotes any of the four nitrogenous bases) conducive to
to a reservoir containing 2 mg/mL of fresh LPS in 20 lL running exhibiting specific affinity to LPS (Fig. 5B).
buffer in order to expedite the next round of selection. Following The relative affinities of the screened aptamers to LPS were
1 h incubation, this equilibrium mixture was then injected to the judged based on the Kd values estimated for individual LPS binders
capillary to start the second round of screening. In this study, three using Eq. (1). As shown in Table 1, Kd values for the majority of 10
rounds of selection were conducted and the collected aptamer LPS binders were found to be in nanomolar range, suggesting that
pools from each round (i.e., Fractions C1, C2, and C3) were used most of the selected sequences exhibit significantly high affinity to
to assess the bulk affinity of screened aptamers to LPS. Each frac- LPS as compared with the controls. It is also noted that B1, B2, B3,
tion containing enriched LPS affinity aptamers was amplified by and B4 exhibit similar high affinity to LPS as anticipated from their
PCR, purified, and quantified by RT-PCR. Ten microliters of 60 nM high sequence homology reported in Fig. 5B. This further confirms
of these aptamers was mixed with 10 lL of 4 mg/mL of fresh LPS that the consensus sequence plays an important role in LPS recog-
solution and 150 nL of this mixture was injected to the capillary. nition and binding of these four binders. Interestingly, B9, which
A peak corresponding to LPS–aptamer complex was observed in also shows a low Kd of 15 nM, exhibits only 32–34% sequence
a time window 25–30 min in front of a small peak accounting for homology with B1 to B4 analogues, suggesting that B9 may be an
free DNA around 32–34 min (Fig. 3). Using Eq. (1), the equilibrium independent LPS binder evolved from a different ssDNA origin.
dissociation constant (Kd) for the interaction between the enriched The negative controls (i.e., the ssDNA aptamer library and/or an
aptamer pools and the LPS was determined from the analysis of unbound ssDNA sequence eliminated from the screening at an
peak areas for LPS-aptamer complexes and free DNA (Fig. 4). It early stage) showed no binding to LPS.
was found that the bulk Kd measured for the fractionated aptamer Putative secondary structures of each LPS binding aptamer were
pool following each round of selection decreased with selection predicted by submitting the corresponding aptamer sequences to
rounds, indicative of the successful enrichment of aptamers with the Mfold web server [38]. As shown in Table 2, a common second-
enhanced affinity to the target. In addition, the bulk affinity of ary structure resembling ‘‘M-shape’’ was found for B1 to B3 bind-
aptamers evolved after the third round of selection showed no fur- ers, further confirming that the high sequence homology among
ther difference when compared to that from the second round, these binders allows them to have a conservative structural fit con-
confirming that three rounds of NECEEM-based non-SELEX should ducive to recognizing LPS. Despite the high sequence homology
be sufficient to obtain high affinity aptamers to LPS. (e.g., B4) or the comparable low dissociation constant (e.g., B9),
B4 and B9 showed several putative secondary structures departing
Assessment of Kd for the screened LPS affinity aptamers from those predicted for B1 to B3, suggesting that LPS recognition
by B4 or B9 might be mediated via differential mechanisms in the
Following the three rounds of selection, 32 clones harboring the context of either structural and/or sequence fit.
PCR-amplified LPS affinity aptamers ligated into PC 2.1 plasmid Taken together, NECEEM was successfully introduced into the
were sampled and their DNA sequences analyzed. Table 1 shows non-SELEX process to greatly facilitate the screening of high affin-
the aptamer sequences recovered from 32 clones. For simplicity, ity aptamers to LPS in a few days in contrast to the conventional
each LPS affinity aptamer is designated as B1 to B10. It is noted that SELEX usually requiring several months to complete 8 to 15 rounds

Fig.2. Electropherograms for (A) 25 lM ssDNA library; (B) equilibrium mixture of 25 lM ssDNA library and 2 mg/mL LPS after 1 h incubation.
16 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20

Fig.3. Electropherograms of equilibrium mixture between LPS and (A) Fraction C1, (B) Fraction C2, (C) Fraction C3, (D) ssDNA library.

of the target molecule and aptamer library members could be

effectively retained. This would decrease the chances of nonspe-
cific bindings of aptamers to the surface of the solid supports
(e.g., filters or chromatographic matrices) often used in the con-
ventional partitioning methods to immobilize target molecules,
thereby resulting in significant affinity enhancement of the initial
DNA library just after three rounds of selection.
In order to explore the possibility of harnessing LPS binding
aptamers for construction of an electrochemical aptasensor capa-
ble of detecting LPS from complex biological liquors, at least two
factors (i.e., sensitivity and specificity) must be considered. On
the basis of the aforementioned sequence and structure analysis
for the 10 LPS binding aptamers, the highest LPS affinity binder
B2 was therefore selected to see if aptamer molecules function as
a sensor probe to mediate electrochemical impedance changes on
their interaction with LPS.

Construction of the aptasensor and assessment of its performance

Fig.4. Kd assessed for the enriched aptamer pools during the three rounds of
screening against LPS. The B2 binder was modified with a thiol group (HS–(CH2)6–) at
its 50 terminal to facilitate its anchoring to the gold electrode sur-
of repeated screening. More importantly, NECEEM-based non-SE- face. Prior to constructing an aptasensor where the aptamer B2 on
LEX was conducted in solution where the natural configurations the electrode works as a detection probe, it is important to confirm
 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20 17

Table 1
Aptamer sequences screened against LPS and Kd of each aptamer estimated using Eq. (1).

ID Sequence Frequency Kd (nM)

a a
B1 H -TAGCCGGATCGCGCTGGCCAGATGATATAAAGGGTCAACCCCCCA-T 23 20
B2 H-TAGCCGGATCGCGCTGGCCAGATGATATAAAGGGTCAGCCCCCCA-T 1 12
B3 H-TAGCCGGATCGCGTTGGCCAGATGATATAAAGGGTCAACCCCCCA-T 1 32
B4 H-TAGCCGGATCGCGCTGGCCAGATGATATAAAGGGTCAACCCCCCG-T 1 21
B5 H-CTAAGCACAGGGAAACCAGCTAATGAGTTAGGCCTGTCCCCCACG-T 1 484
B6 H-CAATGGACCTATTCGAGTACTGAATAGAACAGTCGGCGCTCTGGG-T 1 298
B7 H-TTCAAGACGATGCCTGGCGCGAGTTACACACTTGCATGGAGCTGG-T 1 61
B8 H-ATCCAATACCTCAGAACTCAGTTCGAGTCGTAAAGGGGAATCGCA-T 1 71
B9 H-ACCGATCCATCGAGTTTCTGAGAAAGGCCCGGAGAAACCGCGAGA-T 1 15
B10 H-TCAATCTAACCATGCATGCAGTTTAGGCAGGATTCGTTATCGCAA-T 1 2336
Control1b H-(45N)-T – d

Control2c H-GTCATGCTCGAGTCAGGGGTTAGTAAGTCGACGGTACTACGGGAA-T –d
a
Constant Head (H) and Tail (T) sequence regions flanking the randomized 45-mer ssDNA. H and T represent CTTCTGCCCGCCTCCTTCC and GGAGACGAGATAGGCGGACACT,
respectively.
b
Control 1 (C1) is the randomized aptamer library used for first round of screening.
c
Control 2 (C2) is recovered from the unbound fraction of ssDNA library after the first round of NECEEM-based screening, hence expected to have no affinity to LPS.
d
Kd values for the control sequences are unable to be determined due to their negligible interaction with LPS.

Fig.5. Homology analysis with DNAman for (A) the 10 LPS binding aptamers and (B) the selected 4 binders (i.e., B1 to B4).

if the B2 would preserve its LPS recognition ability following its its thiol modification at 50 terminal interacts with LPS more or less
modification and/or immobilization onto the gold electrode. In or- to the same extent in the nanomolar range. It is therefore con-
der to address this issue, the affinity of B2 binder to LPS was reas- cluded that the screened aptamer B2 retains its high affinity to
sessed using an independent technique using surface plasmon LPS molecules following its modification and immobilization on a
resonance (SPR) where the interaction between the thiol-modified gold surface, thereby likely to act as an LPS-sensing element.
B2 aptamer immobilized on the gold chip and LPS in running buffer An aptasensor harnessing the B2 aptamer as an LPS probe was
could be monitored quantitatively. The SPR-based Kd between the fabricated as described under Materials and methods. During mod-
immobilized B2 and the LPS was found to be approximately 38 nM. ification of the gold electrode, cyclic voltammetry (CV) measure-
This is very close to the aforementioned CE-based Kd value of ments in 2 mM K3Fe(CN)6 aqueous solution containing 0.1 M KCl
12 nM, indicating that either the free or the immobilized B2 after were conducted in the range of 0.3 to 0.6 V at a scan rate of
18 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20

Table 2 the solution resistance (Rs), the double layer capacitance (C), the
Putative secondary structures of each LPS bindng aptamer predicted by Mfold DNA Warburg diffusion resistance (W), and the electron-transfer resis-
and assessment of sequence homology of each LPS binder against the strongest LPS
binder B2.
tance (Ret) [33] to fit the electron transfer kinetics of the redox cou-
ples (i.e., FeðCNÞ6 3=4 ) at the interface. The model-based
ID Predicted secondary Homology regression and experimental results showed good agreements over
structuresa to B2 (%)
the 0.1 Hz and 100 kHz frequency range. The Nyquist plots for a
B1 97.78 gold electrode at each stage of the surface treatment (i.e., aptamer
conjugation followed by MCH blocking) are shown in Fig. 6A (i.e.,
curves 1–3).
The bare gold electrode exhibits an almost straight line with a
B2 –
negligible semicircle observed at a very high frequency region in
the Nyquist plot of impedance spectroscopy (Fig. 6A, curve 1), indi-
cating that a diffusion-controlled process is dominant for the entire
frequency range explored. Following deposition of aptamer mole-
B3 95.56
cules on the gold electrode, an obvious semicircle indicative of
the onset of charge-transfer-controlled process (Fig. 6A, curve 2)
appeared. However, it is noted that the contribution of diffusion-
controlled process still persists at the low frequency region. This
B4 95.56 indicates a successful formation of an aptamer layer on the elec-
trode which apparently acted as a barrier for the electrolytes
(i.e., FeðCNÞ6 3=4 ions) to access the electrode surface due to the
presence of a large number of negatively charged phosphate
B5 42.55 groups in the adsorbed aptamer layer [27]. When the aptamer-
modified gold electrode was further treated with MCH solution,
the larger semicircle in the Nyquist plot (Fig. 6A, curve 3) evidenc-
ing the formation of a mixed self-assembled monolayer of the apt-
B6 38.30 amer and the spacer thiol was observed. This monolayer on the
electrode works against migration and/or mass transfer of elec-
trons from the bulk solution to the electrode surface by further
insulating the conductive support and hindering the access of re-
B7 32.61 dox probe toward the electrode surface, thereby resulting in a sig-
nificant increase in Ret [39].
Herne and Tarlov [37] have extensively characterized the two-
step method to immobilize ssDNA probes on gold surfaces. They
reported that the posttreatment with MCH largely replaced disori-
B8 30.00
ented nonspecifically positioned aptamers (i.e., those attached on
the electrode surface through nitrogenous base or sulfur atom-
mediated physisorption), facilitating precise orientation of ssDNA
molecules. Besides, the addition of a large number of MCH mole-
B9 31.91
cules rendered topologically favored surface spread ssDNA mole-
cules in a horizontal configuration oriented in axial direction,
giving rise to a tightly packed aptamer/MCH mixed layer likely to
pose a high steric and electrostatic hindrance to redox ions in
B10 40.43 accessing the electrode surface [40].
Following the formation of a stable aptamer/MCH mixed layer
on the gold electrode, the aptamer-coupled electrode was used to
detect LPS at varying concentrations (0.01, 0.1, 0.25, 0.5, 0.75, and
a
1 ng/mL) in PBS by EIS measurement. As shown in Fig. 6A (curves
The structures for each LPS binder are arranged based on their predicted Gibbs
4–9) and Fig. 6B, the increase in Ret is proportional to the LPS con-
free energy in ascending order.
centration with R2 coefficient >0.99 in the range of 0.01–1 ng/mL
(i.e., 0.05–5 EU/mL), which is comparable to the LPS detection
100 mV s1 to characterize the changes on the electrode surface. range (i.e., 0.05–5 EU/mL) reported by the conventional LAL assay.
After immobilization of aptamer molecules followed by MCH Moreover, the time required for LPS detection by the demonstrated
blocking, the redox peak showed a corresponding significant de- aptasensor is no longer than 15 min as compared to 1.5–2.0 h for
crease for each modification step. This demonstrates that the gold the LAL assay. The results in Fig. 6 clearly show that the screened
surface was mostly covered by the mixed self-assembled mono- aptamer molecule B2 is functional in recognizing LPS following
layer of aptamer and MCH. Aptamer density on the gold electrode its conjugation to the gold electrode, and able to mediate elec-
was measured by QCM and found to be 2.0 ng/mm2 (data not tron-transfer resistance change (DRet) sufficient to detect femtom-
shown). olar levels of LPS in the solution.
The surface properties of the modified electrode have been Since LPS is one of the more abundant molecules present in the
characterized by EIS. A typical impedance spectrum (presented in outer membrane of gram-negative bacteria, many other host cell-
Nyquist plot) comprises a semicircled high frequency and linear derived components such as sugars, lipids, nucleic acids, and pro-
low frequency regions accounting for electron-transfer and diffu- teins are likely to coexist with LPS in a majority of biological li-
sion-limited processes, respectively. The electrode/electrolyte quors from which purification of a target bioproduct (e.g.,
interface impedance has been approximated by an equivalent cir- therapeutic proteins, plasmid DNA vaccines, cytokines) is often at-
cuit with mixed kinetics and charge transfer control comprising tempted. For a practical implementation of the developed LPS
 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20 19

Fig.6. (A) Nyquist plots for the gold electrode during modification (1–3: bare, aptamer deposition, MCH deposition) and detection (4–9: 0.01, 0.1, 0.25, 0.5, 0.75, and 1 ng/mL
of LPS) processes and the equivalent circuit (inner) used to model EIS data. (B) The linear relationship between DRet and LPS concentration.

Fig.7. The electron-transfer resistance (DRet) calculated from the EIS data recorded after incubation of the aptasensors with LPS (1 ng/mL), pDNA (100 ng/mL), RNA (25 lg/
mL), BSA (50 lg/mL), glucose (50 lg/mL), sucrose (50 lg/mL), and cholesterol (50 lg/mL). An aptasensor employing the control aptamer (C2) showed virtually no interaction
with 1 ng/mL of LPS as shown in LPS column. DRet represents the change in Ret of the aptasensors after 15 min incubation with the various analytes.

aptasensor to the downstream bioprocess, it is therefore important anchored to a gold electrode following the same procedure applied
to determine how specifically the aptasensor interacts with LPS to B2. From Fig. 7, it is evident that LPS detection by the aptasensor
without exhibiting cross-binding affinity to other biomolecules. is solely mediated by the presence of a high affinity LPS binder B2
In order to assess the sensor performance in view of selective LPS since the EIS response on incubation of the control aptasensor with
detection, pDNA, RNA, BSA, glucose, sucrose, and cholesterol were LPS showed a negligible DRet.
selected as the representative model analytes most likely to be
found in a LPS-rich milieu. As summarized in Fig. 7, the developed Conclusions
aptasensor showed negligible cross-reactivity (i.e., almost no DRet)
to pDNA, RNA, BSA, glucose, or sucrose despite their high concen- A high affinity LPS binder B2 was identified by a NECEEM-based
trations (i.e., 100–50000 times higher than LPS concentrations non-SELEX method. To explore the possibility of harnessing B2 apt-
used for the test). Even though the DRet obtained from cholesterol amer as a specific sensor probe to detect LPS in solution, an electro-
is a little higher than the other analytes tested, considering that a chemical apatasensor based on aptamer/MCH-mixed SAM on the
50,000 times higher (as compared to LPS) concentration of choles- gold electrode was fabricated. The developed aptasensor was dem-
terol was used for the measurement, this result is still able to dem- onstrated to detect LPS at a range of 0.01–1 ng/mL (i.e., at a fem-
onstrate that the aptasensor is highly selective in interacting with tomolar level comparable to the conventional LAL assay) in
LPS and hence compatible with complex bioprocessing liquors. 15 min. Furthermore, the aptasensor showed little cross-interac-
To further demonstrate the specificity of the LPS aptamer, the tion reactivity to plasmid DNA, RNA, proteins, saccharides, and/or
control aptamer (i.e., C2 in Table 1) lacking affinity to LPS was lipids which are most likely to coexist with LPS in a majority of
20 Electrochemical detection of endotoxin using aptamers / S.-E. Kim et al. / Anal. Biochem. 424 (2012) 12–20

biological liquors, providing a good potential in realizing practical [20] A.P. Drabovich, M. Berezovski, V. Okhonin, Selection of smart aptamers by
methods of kinetic capillary electrophoresis, Anal. Chem. 78 (2006) 3171–
applications of the developed aptasensor to crude biological li-
3178.
quors where a rapid and specific detection of LPS is of primary [21] M. Berezovski, M. Musheev, A. Drabovich, S.N. Krylov, Non-SELEX selection of
concern. aptamers, J. Am. Chem. Soc. 128 (2006) 1410–1411.
[22] M.V. Berezovski, M.U. Musheev, A.P. Drabovich, J.V. Jitkova, S.N. Krylov, Non-
SELEX: selection of aptamers without intermediate amplification of candidate
Acknowledgments oligonucleotides, Nat. Protoc. 1 (2006) 1359–1369.
[23] M. Berezovski, A. Drabovich, S.M. Krylova, M. Musheev, V. Okhonin, A. Petrov,
This work was supported by research Grants (NRF-C1AAA001– S.N. Krylov, Nonequilibrium capillary electrophoresis of equilibrium mixtures:
a universal tool for development of aptamers, J. Am. Chem. Soc. 127 (2005)
2011-0018903) through the National Research Foundation of 3165–3171.
Korea (NRF) funded by the Ministry of Education, Science, and [24] J.L. Ding, S.T. Gan, B. Ho, Single-stranded DNA oligoaptamers: molecular
Technology (MEST), Korea. recognition and LPS antagonism are length- and secondary structure-
dependent, J. Innate Immun. 1 (2009) 46–58.
[25] J.G. Bruno, M.P. Carrillo, T. Phillips, In vitro antibacterial effects of
References antilipopolysaccharide DNA aptamer–C1qrs complexes, Folia Microbiol. 53
(2008) 295–302.
[1] A. Preston, R.E. Mandrell, B.W. Gibson, M.A. Apicella, The lipooligosaccharides [26] A. Wen, Q. Yang, J. Li, F. Lv, Q. Zhong, C. Chen, A novel lipopolysaccharide-
of pathogenic gram-negative bacteria, Crit. Rev. Microbiol. 22 (1996) 139–180. antagonizing aptamer protects mice against endotoxemia, Biochem. Biophys.
[2] R.J. Ulevitch, P.S. Tobias, Receptor-dependent mechanisms of cell stimulation Res. Commun. 382 (2009) 140–144.
by bacterial endotoxin, Annu. Rev. Immunol. 13 (1995) 437–457. [27] H. Cai, T.M.-H. Lee, I.M. Hsing, Label-free protein recognition using an
[3] Y.G. Kim, C.S. Lee, W.J. Chung, E.M. Kim, D.S. Shin, J.H. Rhim, Y.S. Lee, B.G. Kim, J. aptamer-based impedance measurement assay, Sens. Actuators B 114 (2006)
Chung, Screening of LPS-specific peptides from a phage display library using 433–437.
epoxy beads, Biochem. Biophys. Res. Commun. 329 (2005) 312–317. [28] Y. Xu, L. Yang, X. Ye, P. He, Y. Fang, An aptamer-based protein biosensor by
[4] J.P. Nolan, J.J. McDevitt, G.S. Goldmann, C. Bishop, Endotoxin binding by detecting the amplified impedance signal, Electroanalysis 18 (2006) 1449–
charged and uncharged resins, Proc. Soc. Exp. Biol. Med. 149 (1975) 766–770. 1456.
[5] A.C. Issekutz, Removal of gram-negative endotoxin from solutions by affinity [29] A. Sassolas, L.J. Blum, B.D. Leca-Bouvier, Electrochemical aptasensors,
chromatography, J. Immunol. Methods 61 (1983) 275–281. Electroanalysis 21 (2009) 1237–1250.
[6] S. Minobe, T. Sato, T. Tosa, I. Chibata, Characteristics of immobilized histamine [30] A. Merkoci, M. Pumera, X. Llopis, B. Perez, M. Delvalle, S. Alegret, New
for pyrogen adsorption, J. Chromatogr. A 262 (1983) 193–198. materials for electrochemical sensing VI: carbon nanotubes, TrAC Trends Anal.
[7] F.B. Anspach, O. Hilbeck, Removal of endotoxins by affinity sorbents, J. Chem. 24 (2005) 826–838.
Chromatogr. A 711 (1995) 81–92. [31] M.K. Patel, P.R. Solanki, A. Kumar, S. Khare, S. Gupta, B.D. Malhotra,
[8] T.E. Karplus, R.J. Ulevitch, C.B. Wilson, A new method for reduction of Electrochemical DNA sensor for Neisseria meningitidis detection, Biosens.
endotoxin contaminations from protein solutions, J. Immunol. Methods 105 Bioelectron. 25 (2010) 2586–2591.
(1987) 211–220. [32] J.Y. Heras, D. Pallarola, F. Battaglini, Electronic tongue for simultaneous
[9] I.P. Wicks, M.L. Howell, T. Hancock, H. Kohsaka, T. Olee, D.A. Carson, Bacterial detection of endotoxins and other contaminants of microbiological origin,
lipopolysaccharide copurifies with plasmid DNA: implications for animal Biosens. Bioelectron. 25 (2010) 2470–2476.
models and human gene therapy, Hum. Gene. Ther. 6 (1995) 317–323. [33] J.-D. Qiu, R.-P. Liang, R. Wang, L.-X. Fan, Y.-W. Chen, X.-H. Xia, A label-free
[10] P.M. Montbriand, R.W. Malone, Improved method for the removal of endotoxin amperometric immunosensor based on biocompatible conductive redox
from DNA, J. Biotechnol. 44 (1996) 43–46. chitosan-ferrocene/gold nanoparticles matrix, Biosens. Bioelectron. 25 (2009)
[11] D.S. Wilson, J.W. Szostak, In vitro selection of functional nucleic acids, Annu. 852–857.
Rev. Biochem. 68 (1999) 611–647. [34] M.J. Shin, L. Tan, M.H. Jeong, J.H. Kim, W.S. Choe, Monolith-based immobilized
[12] M. Famulok, G. Mayer, M. Blind, Nucleic acid aptamers-from selection in vitro metal affinity chromatography increases production efficiency for plasmid
to applications in vivo, Acc. Chem. Res. 39 (2000) 591–599. DNA purification, J. Chromatogr. A 1218 (2011) 5273–5278.
[13] A. Ellington, J. Szostak, In vitro selection of RNA molecules that bind specific [35] W. Lee, B.K. Oh, W.H. Lee, J.W. Choi, Immobilization of antibody fragment for
ligands, Nature 346 (1990) 818–822. immunosensor application based on surface plasmon resonance, Colloids Surf.
[14] C. Tuerk, L. Gold, Systematic evolution of ligands by exponential enrichment: B 40 (2005) 143–148.
RNA ligands to bacteriophage T4 DNA polymerase, Science 249 (1990) 505– [36] W.Q. Su, H.T. Nguyen, M.S. Cho, Y.K. Son, Y.K. Lee, Synthesis, characterization
510. and self-assembled film of poly(3-((2,3-dihydrothieno[3,4-b][1,4]dioxin-2-
[15] B. Vant-Hull, A. Payano-Baez, R.H. Davis, L. Gold, The mathematics of SELEX yl)methoxy)propane-1-thiol) (PEDTMSHA), Synth. Met. 160 (2010) 2471–
against complex targets, J. Mol. Biol. 278 (1998) 579–597. 2475.
[16] Y. Tian, N. Adya, S. Wagner, C.Z. Giam, M.R. Green, A.D. Ellington, Dissecting [37] T.M. Herne, M.J. Tarlov, Characterization of DNA probes immobilized on gold
protein:protein interactions between transcription factors with an RNA surfaces, J. Am. Chem. Soc. 119 (1997) 8916–8920.
aptamer, RNA 1 (1995) 317–326. [38] M. Zuker, Mfold web server for nucleic acid folding and hybridization
[17] D. Schneider, L. Gold, T. Platt, Selective enrichment of RNA species for tight prediction, Nucleic Acids Res. 31 (2003) 3406–3415.
binding to Escherichia coli rho factor, FASEB J. 7 (1993) 201–207. [39] U.K. Sur, R. Subramanian, V. Lakshminarayanan, Cyclic voltammetric and
[18] H. Chen, L. Gold, Selection of high-affinity RNA ligands to reverse electrochemical impedance studies on the structure, adsorption kinetics, and
transcriptase: inhibition of cDNA synthesis and RNase H activity, barrier properties of some organic dithiol self-assembled monolayers on gold,
Biochemistry 33 (1994) 8746–8756. J. Colloid Interface Sci. 266 (2003) 175–182.
[19] A. Petrov, V. Okhonin, M. Berezovski, S.N. Krylov, Kinetic capillary [40] F. Lucarelli, Carbon and gold electrodes as electrochemical transducers for
electrophoresis (KCE): a conceptual platform for kinetic homogeneous DNA hybridisation sensors, Biosens. Bioelectron. 19 (2004) 515–530.
affinity methods, J. Am. Chem. Soc. 127 (2005) 17104–17110.