A fibre-optic biosensor for detection of microbial contamination

Article abstract of DOI:10.1139/v03-070

A fibre-optic biosensor is described for detection of genomic target sequences from Escherichia coli. A small portion of the LacZ DNA sequence is the basis for selection of DNA probe molecules that are produced by automated nucleic acid synthesis on the s

Full text of DOI:10.1139/v03-070

339
AWARD LECTURE / CONFÉRÉNCE D’HONNEUR
A fibre-optic biosensor for detection of microbial
contamination1
Amer Almadidy, James Watterson, Paul A.E. Piunno, Inge V. Foulds,
Paul A. Horgen, and Ulrich Krull
Abstract: A fibre-optic biosensor is described for detection of genomic target sequences from Escherichia coli. A
small portion of the LacZ DNA sequence is the basis for selection of DNA probe molecules that are produced by auto-
mated nucleic acid synthesis on the surface of optical fibres. Fluorescent intercalating agents are used to report the
presence of hybridization events with target strands. This work reviews the fundamental design criteria for development
of nucleic acid biosensors and reports a preliminary exploration of the use of the biosensor for detection of sequences
that mark the presence of E. coli. The research work includes consideration of the length of the strands and non-selective
binding interactions that can potentially block the selective chemistry or create background signals. The biosensors
were able to detect genomic targets from E. coli at a picomole level in a time of a few minutes, and dozens of cycles
of use have been demonstrated. In a step towards the preparation of a completely self-contained sensor technology, a
new intercalating dye known as SYBR 101 (Molecular Probes, Inc.) has been end-labelled to the LacZ nucleic acid
probe, to examine whether dye tethered onto an oligonucleotide terminus could fluorimetrically transduce the formation
of hybrids. The results obtained from experiments in solution indicate that the use of tethered dye provides fluores-
cence signals that are due to hybridization, and that this process is functional even in the presence of a high concentra-
tion of non-selective background DNA obtained from sonicated salmon sperm.
Key words: biosensor, DNA, fibre optic, hybridization, fluorescence, pathogen, E. coli.
Résumé : On décrit une fibre optique servant de biosenseur pour la détection du génome cible dans la séquence du
Escherichia coli. Une petite portion de la séquence du LacZ-ADN sert de base pour la sélection des molécules sondes
d’ADN obtenues par la synthèse automatique d’acides nucléiques à la surface des fibres optiques. On a utilisé des
agents intercalants fluorescents pour mettre en évidence la présence de phénomènes d’hybridation avec la fibre cible.
Ce travail passe en revue le critère de conception fondamental de développement de biosenseurs d’acides nucléiques, et
rapporte également une exploration préliminaire de l’utilisation des biosenseurs pour détecter les séquences qui témoi-
gnent de la présence du E.coli. Les travaux de recherche tiennent compte de la longueur de la fibre et des interactions
liantes non sélectives qui peuvent potentiellement bloquer la sélectivité chimique ou créer un bruit de fond. Les biosen-
seurs ont été capables de détecter le génome cible à partir du E. coli au niveau de la picomole et en quelques minutes,
on a ainsi pu effectuer des douzaines de cycles d’études. Dans une étape en vue de la préparation de la technologie du
senseur auto-contenu, on a marqué un nouveau colorant intercalaire connue sous le nom de SYBR 101 (Molecular
Probes, Inc.) à l’extrémité de la sonde d’acide nucléique LacZ, pour voir si le colorant attaché dans l’oligonucléotide
terminal peut fluorométriquement induire la formation d’hybrides. Les résultats obtenus à partir des expériences en so-
lution indiquent que l’utilisation de colorant attaché fournit des signaux fluorescents qui sont dus à l’hybridation, et
que ce processus est fonctionnel même en présence d’une forte concentration d’ADN d’arrière plan non sélectif ob-
tenue à partir du sperme de saumon traité à l’ultrason.
Mots clés : biosenseur, ADN, fibre optique, hybridation, fluorescence, pathogène, E. coli.
Almadidy et al. 349
Received 18 February 2003. Published on the NRC Research Press Web site at http://canjchem.nrc.ca on 29 May 2003.
A. Almadidy, J. Watterson, and U. Krull.2 Department of Chemistry, University of Toronto at Mississauga, Mississauga, ON
L5L 1C6, Canada.
P.A.E. Piunno. FONA Technologies, Inc., 785 Bridge Street, Waterloo, ON N2V 2K1, Canada.
I.V. Foulds and P.A. Horgen. Department of Biology, University of Toronto at Mississauga, Mississauga, ON L5L 1C6, Canada.
1MAXXAM Award Lecture. Presented by U.J. Krull at the Canadian Society for Chemistry Conference, 2002.
2Corresponding author (e-mail: ukrull@utm.utoronto.ca).
Can. J. Chem. 81: 339–349 (2003)
doi: 10.1139/V03-070
© 2003 NRC Canada

340
Can. J. Chem. Vol. 81, 2003
This may be owing to local density changes in the nucleic
acid films as a result of the denaturation. The results sug-
gested that there may be significant differences in the nature
of the base pairing in an interfacial environment compared
with that which occurred in bulk solution. There did not ap-
pear to be a relationship between the packing density of im-
mobilized oligonucleotides and the reduction in the
endothermicity of the denaturation. The observed Tm values
were still of comparable magnitude to those that were ob-
served in experiments done in bulk solution, so it is likely
that there is a significant difference in entropy changes ac-
companying hybridization and denaturation in an interfacial
environment relative to those observed in experiments done
in bulk solution.
The data further suggest that the selectivity of hybridiza-
tion in an interfacial environment may be substantially dif-
ferent and advantageous in comparison with that observed in
a bulk solution environment. Furthermore, the selectivity of
hybridization does not necessarily follow the trend of Tm,
which is seen as a function of ionic strength and oligo-
nucleotide immobilization density. These results corroborate
the notion that there is an ensemble of interactions that will
occur, along with the hybridization–denaturation transition,
in an interfacial environment. These interactions contribute
to the overall stability of the binding of target DNA and
therefore play an important role in defining the Tm values of
a particular probe–target complex, as well as the shape of
the thermal denaturation profile, and therefore ultimately af-
fect the selectivity of hybridization.
The results of such work (15, 17) indicate that the choice
of density of immobilized single-stranded DNA (ssDNA)
does provide for control of selectivity. The optimization of
the analytical function of a fibre-optic biosensor for any par-
ticular hybridization assay must consider both the issues of
selectivity and the amount of fully complementary double-
stranded DNA (dsDNA) that is formed, as this is the source
of the analytical signal. A review of the data suggests that it
is possible to select a combination of ssDNA density, solu-
tion ionic strength, and temperature that provides an im-
provement in the selectivity coefficient of about two orders
of magnitude when comparing the formation of the fully
complementary duplex to that which contains one central
base-pair mismatch. Bulk solution experiments do not attain
the same magnitude of selectivity coefficient. Importantly,
only about 30% of the maximum amount of fully matched
dsDNA is available under the conditions in bulk solution
where the best case of selectivity is achieved. This does not
compare favourably with the immobilized ssDNA system,
where about 55% of the maximum amount of fully matched
dsDNA is available under conditions where selectivity is
maximized.
An important consideration in the evaluation of the sensi-
tivity and selectivity of hybridization for a given sensor sys-
tem is the nature of the sample that is being introduced.
Samples may contain various levels of large, non-
complementary genomic DNA and RNA molecules, which
may interfere with analysis. Also, most nucleic acid sensor
systems will be exposed to the target DNA of interest in
double-stranded form. This imposes the requirement of de-
naturing these double-stranded targets so that selective hy-
bridization may subsequently take place at the sensor
Introduction
Design of a nucleic acid biosensor
Methods that are suitable for the routine determination of
the presence of bacterial species are of obvious importance
in the quality control of foodstuff and water resources every-
where. Classical methods of nucleic acid hybridization assay
are often time consuming and so may be inappropriate for
laboratories that require rapid turnover of results and high
sample throughput (1–6). As a result, significant research
has been devoted to the development of biosensors in an at-
tempt to generate assays that are reversible, reusable, sensi-
tive, selective, and relatively simple to use. Methods of
signal transduction to detect nucleic acid hybridization on
biosensor surfaces include electrochemical, piezoelectric,
and optical approaches (7–11) and compete favourably with
the recent developments in wet-chemical methods such as
those based on the polymerase chain reaction (12, 13).
The development of biosensors must take into consider-
ation the effects of the local environment on the binding ca-
pacity of the immobilized selective molecular recognition
elements. This involves careful consideration of the nature
of the solid substrate used, the method of immobilization,
solution conditions under which experiments will generally
be done, and probe density and length. Each of these param-
eters has direct consequences on the reproducibility and sen-
sitivity of signal generation (14–17).
To examine the effects of immobilization on nucleic acid
hybridization, thermal denaturation experiments were done
at the surface of fibre-optic nucleic acid biosensors to deter-
mine whether trends in hybridization observed in bulk solu-
tion could be extrapolated to describe nucleic acid
hybridization in an interfacial environment (17).
The immobilization of oligonucleotide probes onto the
surface of fused-silica optical fibre substrates was achieved
by means of a modification using a silane reagent (15, 17).
The modified optical fibre substrates were then subjected to
standard β-cyanoethyl-phosphoramidite oligonucleotide syn-
thesis protocols in order to undergo the stepwise synthesis of
oligonucleotides at controlled densities onto the surface of
the substrates.
To examine the energetics of interfacial hybridization, the
van’t Hoff enthalpy changes and temperature-corrected stan-
dard enthalpy changes were computed for a series of dena-
turation experiments, conducted based on the method
developed by Piunno and co-workers (17). This model ap-
plies to denaturation occurring within a film of immobilized
nucleic acids, with the complementary DNA freely able to
float in and out of the membrane. The model assumes no
interaction between neighbouring strands and that the dena-
turation is a two-state transition. The enthalpic change ac-
companying denaturation in an interfacial environment was
significantly lower than that observed in experiments con-
ducted in bulk solution. The sensitivities of ∆HVH (Tm) to
changes in the characteristic melt temperature (Tm) were a
factor of 2–4 smaller for the transitions occurring at the in-
terface of the optical biosensors relative to those observed
for the experiments done in bulk solution and were usually
opposite in sign. This suggested that the changes in heat ca-
pacity that accompanied the denaturation were not the same
in an interfacial environment as they were in bulk solution.
© 2003 NRC Canada

Almadidy et al.
341
surface. In practice, this may result in a competition for
hybridization of target strands in bulk solution between im-
mobilized probe oligonucleotides and the complementary
DNA in bulk solution. This competition for hybridization
may impart some significant limitations on the sensitivity
and selectivity of the assay. A balance can be struck be-
tween the desired sensitivity and selectivity of a given hy-
bridization assay, and this balance is somewhat tunable by
means of controlling the density of ssDNA immobilization.
If non-selective adsorption occurs predominantly in regions
between the immobilized oligonucleotide probes, then it
may be that optimal assay sensitivity and selectivity would
be achieved using a sensor with a higher density of probe
molecules, where the number of exposed surface sites for
non-selective adsorption is decreased.
To more accurately model the effects of interferences as
experienced in a real sample, experiments were done to in-
vestigate the effects of the presence of large genomic DNA
strands (20 kbp average size) on the response of the sensors
to labeled oligonucleotides. The experiments were designed
such that the relative concentrations of oligonucleotides and
genomic DNA were adjusted. Experiments were done using
concentration regimes where both fully complementary
(cDNA) and non-complementary (ncDNA) oligonucleotide
20mers were introduced at a concentrations of about 1015
molecules L–1, while genomic DNA from E. coli was intro-
duced at a concentration of 1012 to 1014 molecules L–1 (15).
The results suggest that the presence of genomic DNA as a
background species does not substantially block hybridiza-
tion of very short target oligonucleotides or the extent of
non-selective adsorption of short non-complementary
oligonucleotides. This trend was observed for all concentra-
tion regimes used. Additionally, the presence of genomic
DNA did not affect the response times of the sensors in the
first few minutes of an analysis. Interestingly, the pretreat-
ment of the sensor surface with genomic DNA for 10 min
reduced the response time of the sensors to cDNA. It may be
that the larger genomic DNA acted to reduce the effective
solution volume near the sensor surface and to increase the
effective analyte concentration. This effect was more signifi-
cant at the lower analyte concentration, where response
times were more sensitive to changes in analyte concentra-
tion. The response time of sensors to the addition of
1015 molecules L–1 cDNA was 224 ± 5 s, while the response
time after pretreatment with genomic DNA was 192 ± 5 s.
The response time for addition of 1016 molecules L–1 cDNA
was 28 ± 1 s, while the response time after pretreatment of
the surface with genomic DNA was 21 ± 1 s.
tential bacterial pathogen contamination. Using such indica-
tors, researchers in the U.S.A. estimate that 40% of private
water supplies and 70% of spring-fed supplies contain
coliform bacteria (4). Coliform bacteria concentrations are
determined using methods specified by the Environmental
Protection Agency (EPA) and those found in ref. 5. These
methods can be slow, and new biosensor technologies may
offer substantial advantages in providing analyses within
seconds to minutes.
Currently, several methods are used for the detection or
enumeration of E. coli cells in water, including microbiolog-
ical, serological, and immunological procedures. Polymerase
chain reaction (PCR) methods have been developed, where
LacZ, lamB, and uid genes have been used as targets for the
design of primers for coliform detection (6). False positive
and negative results can arise when using these techniques,
and only a very limited number of strains have been used to
test for specificity of the primers. Standard PCR generally
only provides information about detection, and even when
using quantitative real-time PCR, the analyses still often re-
quire hours.
Herein we report the development of a fibre-optic biosen-
sor for the detection of short sequences of oligonucleotides
that indicate the presence of E. coli. Single-stranded DNA
(ssDNA) was immobilized by covalent binding to a fused
silica optical fibre. Hybridization on the solid surface was
detected by use of the fluorescent intercalating dye, ethidium
bromide (EB). Testing to detect coliform contamination of
water was demonstrated using selective hybridization of nu-
cleic acid sequences. A 25mer sequence on the LacZ gene of
the E. coli was targeted using a 25mer ssDNA probe. The in-
vestigation has shown that the biosensor was capable of de-
tecting minute amounts of synthetic cDNA and also genomic
DNA that was extracted from E. coli. The biosensor could
provide analytical information in less than 1 min and was
regenerable for many cycles of application.
Preliminary work that targets the development of a self-
contained biosensor has involved attachment of the interca-
lating fluorescent reporter dye to the probe by means of a
short molecular tether. The intercalating fluorescing dye
(SYBR 101) was covalently attached through a short tether
to the 25mer ssDNA (labelled DNA, L-DNA), and the fluo-
rescence changes caused by hybridization have been investi-
gated in bulk solution using free L-DNA. In the design of a
self-contained biosensor, this approach may help reduce
background fluorescence from free dye in solution, will al-
low internal standardization, and will substantially reduce
the risk of exposure of the operator to toxic chemicals by
confining the intercalating dye to the surface of the device.
These results have ramifications for analyses of real-world
samples. The preliminary results suggest that the process of
non-selective adsorption by interfering short and long nu-
cleic acid sequences in solution may not occur in such a
manner as to substantially inhibit the extent of hybridization
of a target sequence when early in an analytical experiment.
Experimental
Chemicals
SYBR 101, succinimidyl ester, was donated by Molecular
Probes, Eugene, Oregon. Biosynthesis-grade solvents were
purchased (EM Science, Toronto, ON) and further purified
or dried by standard laboratory protocols. Reagents for DNA
synthesis were purchased from Dalton Chemical Labora-
tories Inc. (Toronto, ON) and were used as received or were
prepared as below. Anhydrous acetonitrile (EM Science) was
pre-dried by distillation from P2O5 and redistilled from
A fibre-optic biosensor for E. coli
Coliforms are aerobic and facultatively anaerobic, gram-
negative, non-spore forming bacilli, encompassing members
of Escherichia, Citrobacter, Klebsiella, and Enterobacter
(1–3). Although several of the coliform bacteria are not usu-
ally pathogenic themselves, they serve as an indicator of po-
© 2003 NRC Canada

342
Can. J. Chem. Vol. 81, 2003
calcium hydride under dry argon. Tetrahydrofuran (EM Sci-
ence) was pre-dried over CaH2, filtered, and distilled imme-
diately prior to use from sodium metal (Aldrich) /
benzophenone (Aldrich). Water was double distilled in
glass, treated with diethyl pyrocarbonate (Aldrich), and
autoclaved. Molecular-biology-grade polyacrylamide gel
electrophoresis reagents and apparatus were obtained
through Bio-Rad (Hercules, California). Silica gel (Toronto
Research Chemicals, Toronto, ON) had a particle size of 30–
70 microns.
dium of DNA Grade, in distilled water containing 0.15%
Kathone CG/ICP Biocide as a preservative.
Preparation of optical fibres
The jacket material surrounding the fused silica optical
fibres (400 µm core diameter, 3M Power Core™ Series Op-
tical Fibre, FT-400-URT or FP-400-UHT, distributed by
Thor Labs Inc., Newton, NJ, U.S.A.) was mechanically re-
moved by use of a fibre-stripping tool (Thor Labs Inc.) to re-
veal the fused silica core material and cladding layer.
Optical fibre pieces 48 mm in length were then made by use
of a custom-built, diamond-edged fibre-scoring device. The
termini of the fibre pieces were visually inspected at 40 ×
magnifications to ensure the fibre termini were flat, orthogo-
nal to the length of the fibre, and free of chips and nicks.
The fused silica fibre segments were cleaned prior to sur-
face modification according to the published methods (14).
CPG was used to grow ssDNA in tandem with fused silica
fibres and was subsequently used for recovery of ssDNA to
determine the quality and quantity of synthesis. CPG was
treated identically to fused silica fibres. The fibre substrates
were first immersed and gently agitated in a 1:1:5 (v/v) solu-
tion of 30% ammonium hydroxide – 30% hydrogen peroxide
– water at 80°C for 5 min. The substrates were then recov-
ered, washed with copious amounts of water, and then
treated with 1:1:5 (v/v) concd. HCl – 30% hydrogen perox-
ide – water for 5 min at 80°C with gentle agitation. The sub-
strates were recovered and washed with 100 mL portions of
water, methanol, chloroform, and diethyl ether, respectively,
dried under reduced pressure, and stored in vacuo and over
P2O5 until required.
Instrumentation
Fluorescence studies of LacZ-target hybridization onto the
probe at the surface of the E. coli biosensors were done us-
ing an optic-fibre spectrofluorimeter operated in an intrinsic-
mode configuration (17). The spectrofluorimeter was
equipped with a fluid-handling system for stop-flow fluores-
cence investigations of nucleic acid hybridization. Prelimi-
nary fluorescence studies of LacZ target to the dye labelled
ssDNA in solution were done using a spectrofluorimeter in-
strument.
Attachment of SYBR 101 to ssDNA probe
The SYBR 101 – ssDNA probe consisted of three parts:
the ssDNA, a C6 aminomodifier as a tether, and the fluores-
cent dye SYBR 101 (absorption 483 nm, emission 515 nm).
The tether was attached to the ssDNA using an ABI 392
DNA/RNA synthesizer. C6 aminomodifier is a phosphor-
amidite synthon containing a six-carbon atom chain termi-
nated by a protected amine moiety. The reagent was used in
analogy to a phosphoroamidite nucleoside. The modifier was
activated with tetrazole to form an active intermediate that
coupled to the 5′-hydroxyl terminus of the oligonucleotide,
which was bound to controlled pore glass (CPG) in the final
coupling cycle. Oxidation and ammonium hydroxide cleav-
age – deprotection yielded the 5′-amine-modified oligonu-
cleotide.
The ssDNA-linker at this stage bore a nucleophilic, unpro-
tected primary amine group, which reacted with the electro-
philic N-hydroxy succinamide group of the SYBR 101.
SYBR 101 succinimidyl ester (reactive dye) was used to la-
bel the amine-modified oligonucleotide because it forms a
very stable amide bond between the dye and the amine-
modified oligonucleotide probe. The reactive dye is a hydro-
phobic molecule, so it was dissolved in high-purity
dimethylsulfoxide before reaction with the amine-modified
oligonucleotide probe. The reaction was done in a tetra-
borate buffer at pH 8.5, so that the reactive dye reacted with
the non-protonated amine group on the modified oligo-
nucleotide probe.
Functionalization of fused silica substrates with 3-
glycidoxypropyltrimethoxysilane (GOPS)
The cleaned fused silica substrates were suspended in an
anhydrous solution of xylene – 3-glycidoxypropyltrimeth-
oxysilane – diisopropylethylamine (100:30:1 v/v/v). The re-
action was stirred under argon at 80°C for 24 h. The fibres
were then collected and twice washed with 50 mL portions
of methanol, chloroform, and diethyl ether, respectively, and
then dried and stored in vacuo and over P2O5 at room tem-
perature until required.
Synthesis of dimethoxytrityl hexaethylene glycol (DMT-
HEG)
A solution of dimethoxytrityl chloride (7.1 g, 21 mmol) in
dry pyridine (10 mL) was added dropwise to a stirred solu-
tion of hexaethylene glycol (5.6 mL, 21 mmol in 5 mL
pyridine) under an argon atmosphere. Stirring was continued
overnight, after which time the reaction mixture was com-
bined with dichloromethane (50 mL). The mixture was
shaken against 5% aqueous bicarbonate (2 × 900 mL) and
then with water (2 × 900 mL) to remove unreacted HEG,
pyridine, and salts. The organic layer was dried under re-
duced pressure to yield the crude product. The product was
purified by liquid chromatography using a silica gel column
and an eluent of 1:1 dichloromethane – diethyl ether con-
taining 0.1% triethylamine (2.9 g, 24% yield). 1H NMR
(200 MHz, CDCl3) δ: 7.47–7.19 (m, 9H), 6.81 (d, 4H, J =
8.8 Hz), 3.78 (s, 6H), 3.74–3.51 (m, 22H), 3.22 (t, 2H, J =
5.8 Hz). Purity DMT-HEG = 96%.
In this protocol, 250 µg of the reactive dye was dissolved
in 14 µL dimethylsulfoxide. To this vial, 7 µL of deionized,
distilled H2O was added followed by 75 µL of the sodium
tetraborate buffer and 4 µL of a 25 µg µL–1 of gel-purified
5′-amine-modified oligonucleotide. The vial was placed on
an oscillating platform stirrer at low speed to insure that the
reaction remained well mixed. The reaction was allowed to
continue overnight. The labelled oligonucleotide (L-DNA)
was purified from the reaction mixture by use of a
Pharmacia NAP-10 column containing Sephadex G-25 me-
© 2003 NRC Canada

Almadidy et al.
343
Fig. 1. HPLC anion-exchange chromatogram of the 25mer mixed-base LacZ probe that was grown on CPG and then quantitatively re-
moved by base cleavage. The chromatogram indicates the synthetic purity of the sample used in this investigation.
cur were capped prior to oligonucleotide assembly using
chlorotrimethylsilane (TMS-Cl), as per the method of
Watterson et al. (17). The substrates that had been dried by
storage in vacuo and over P2O5 for a minimum duration of
16 h were suspended in a solution of 1:10 (v/v) chloro-
trimethylsilane–pyridine for 16 h under an argon atmosphere
at room temperature. The fused silica substrates were thrice
washed with 20 mL portions of pyridine, methanol, and di-
ethyl ether, respectively, and stored in vacuo and over P2O5
at 25°C until required.
Linkage of DMT-HEG onto GOPS-functionalized
substrates
DMT-HEG (10 equiv relative to the quantity of surface
hydroxyl moieties, 700 mg DMT-HEG – 100 mg CPG) that
had been dried by extended storage in vacuo and over P2O5
(>72 h) was dissolved in 20 mL of anhydrous pyridine and
introduced to an excess of NaH (10 equiv) that had been
thrice washed with dry hexane to remove the oil in which it
had been suspended. The reaction was permitted to proceed
with stirring for 1 h at room temperature under an argon at-
mosphere. The reaction mixture was filtered through a
sintered glass frit under a positive pressure of argon and the
filtrate immediately introduced to the reaction vessel con-
taining the GOPS-functionalized substrates. One batch of
GOPS-functionalized substrates containing both optical
fibres and CPG was created, for which the DMT-HEG cou-
pling reaction was permitted to proceed under a positive
pressure of argon gas at room temperature with gentle agita-
tion on an oscillating platform stirrer for durations of 4 h.
Following the coupling reaction, the substrates were quickly
recovered by filtration over a fritted glass funnel and washed
with 150 mL portions of methanol, water, methanol, and di-
ethyl ether, respectively, to quench the coupling reaction and
remove non-specifically adsorbed reactants. The DMT-
protected polyether-functionalized substrates were dried by
placement in vacuo and over P2O5 and were maintained un-
der these conditions until further required.
Solid-phase phosphoramidite synthesis of
oligonucleotides
All oligonucleotide synthesis was done using a PE-ABI
391-EP DNA synthesizer (PerkinElmer Applied Biosystems,
Foster City, CA, U.S.A.). The manufacturer-supplied synthe-
sis cycles were employed for oligonucleotide assembly with
modifications to the delivery times of the reagents as re-
quired to completely fill the synthesis columns that were
used. Oligonucleotide synthesis onto optical fibres (400 µm
i.d. × 48 mm) was done in a custom-manufactured Teflon®
synthesis column (6 mm i.d. × 50 mm) capable of holding 8
fibres in an evenly distributed and non-contacting fashion
via cylindrical bores (400 µm i.d. × 2 mm deep) machined
into one of the end caps (9).
A nucleic acid oligonucleotide having the sequence
(5′)CAGGTAATGTGGCGGATGAGCGGCA(3′) was syn-
thesized onto the sensor surface as the ssDNA probe. The
target nucleic acid (cDNA) used to challenge the probe was
an oligonucleotide having the sequence (5′)TGCCGC-
TCATCCGCCACATATCCTA(3′), which was derived from a
portion of the LacZ gene sequence. The 25mer
Capping of unreacted silanol and hydroxyl
functionalities with chlorotrimethylsilane
Sites on the surfaces of the fused silica fibres and CPG
onto which undesired nucleotide synthon coupling could oc-
© 2003 NRC Canada

344
Can. J. Chem. Vol. 81, 2003
Fig. 2. E. coli biosensor response to serial dilutions of cDNA as a function of cDNA concentration at various times. Maximum re-
sponse was observed at about 2 min for all concentrations of cDNA.
oligonucleotides were prepared by use of a phosphoramidite
synthon (Dalton) and standard protocols for oligonucleotide
assembly, purification, and quantitation, as have previously
been reported (15).
Determination of the extent of surface coverage of CPG
substrates with covalently immobilized oligonucleotide–
polyether conjugates was done by anion-exchange HPLC us-
ing methods that have been reported elsewhere (16).
by precipitation of DNA by centrifugation at 2000g for
5 min at 2–8°C. Phenol–ethanol supernatant was removed
and the DNA pellet was washed twice with ethanol and
dried under vacuum. DNA was finally reconstituted by add-
ing 1 mL of 1 × PBS and measurement of absorption at
260 nm showed that the DNA concentration of the resultant
solution was about 350 µg mL–1. The extracted E. coli DNA
was sheared by syringe and then by sonication for 5 min
with a Vibra cell sonicator (Sonics & Materials, Inc.)
equipped with a 5-mm tip and set to 125 W maximum
power at 20 kHz. Samples were kept on ice at all times until
they were used for examination of hybridization.
Lyophilized salmon sperm genomic DNA was reconsti-
tuted in 1 mL of 1 × PBS to a concentration of 350 µg mL–1,
as indicated by measurement of absorption at 260 nm. This
DNA was sheared by syringe and then by sonication for
5 min, as was done for the DNA from E. coli.
E. coli and salmon sperm DNA preparation
A 60 mL culture of E. coli was grown overnight at 37°C
in LB media. Bacteria were harvested by centrifugation at
3000g for 10 min. Cells were lysed using TRIZOL reagent
(Life Technologies, Canada) by repetitive pipetting, using
1 mL of the reagent per 1 × 107 cells of the E. coli. The ho-
mogenized sample was incubated for 5 min at room temper-
ature, and chloroform was then added (0.2 mL chloroform
per 1 mL of TRIZOL). Sample tubes were capped and
shaken vigorously for 15 s and incubated at room temp for
2–3 min. The sample tubes were then centrifuged at 12 000g
for 15 min at 2–8°C, to separate the mixture into a lower,
red phenol–chloroform phase, an interphase, and a colour-
less upper aqueous phase. DNA was precipitated from the
interphase and the organic phase by the addition of 0.3 mL
of 100% ethanol per 1 mL of TRIZOL reagent originally
used. Samples were mixed by inversion and permitted to
equilibrate at room temperature for a few minutes, followed
Hybridization assays of the E. coli biosensor
All sensors were cleaned by sonication in ethanol in a
40 W bath sonicator for 30 min to remove adsorbed contam-
inants from the sensor surface. In all cases, sensors were ac-
tivated for hybridization by undergoing three consecutive
thermal denaturation – re-annealing cycles, in which the sen-
sors were exposed to a 1 × 10–7 M solution of the comple-
mentary 25mer oligonucleotide sequence (cDNA) in
phosphate-buffered saline (PBS) hybridization buffer
© 2003 NRC Canada

Almadidy et al.
345
Fig. 3. Chronofluorimetric response profile of a biosensor for E. coli using immobilized 25mer mixed-base probe on a fused silica op-
tical fibre. Biosensor was exposed to 10 pmols of fully complementary LacZ 20mer (cDNA), 35 ng ssDNA from E. coli (prepared by
treatment of whole genomic DNA by sonication and shearing), 35 ng of dsDNA from E. coli, and 35 ng ssDNA from salmon sperm
(prepared by treatment of whole genomic DNA with sonication and shearing), in 1 × PBS containing 10–7 M ethidium bromide at
40°C with full washing and chemical regeneration with water at 90°C and formamide solution (90% in TE buffer) between samples.
(1.0 mol L–1 NaCl, 50 mmol L–1 total phosphate ion,
pH 7.0) and subsequently subjected to a temperature ramp of
0.3°C per min, over a range from 20 to 80°C. Hybridization
assays were done for optical sensors that were exposed to
solution-phase cDNA mixed with the staining intercalator
ethidium bromide at a final concentration of 0.1 µmol L–1.
Each sample solution had a total volume of 26 µL, once in
the reaction chamber. The flow was stopped and the signal
was recorded over a period of up to 10 min. After reaction,
the sensor was washed at a flow rate of 3 mL min–1. Re-
moval of the bound DNA that had associated with the sensor
surface from the previous analysis was done prior to each
experiment by flushing 15 mL of 90°C water through the
flow cell (3 mL min–1, 5 min), followed by 1 mL of 95%
ethanol, and final wash with 90% formamide in TE buffer
(10 mm L–1 Tris HCl, 5 mmol L–1 EDTA, pH = 8.3).
Salmon sperm DNA (0.1 µmole L–1) served as a control for
genomic non-complementary DNA. Assays were performed
using a solution temperature of 40°C. All hybridization as-
says were done in triplicate.
The L-DNA probe hybridization assays were done in bulk
solution, by titrating the tethered SYBR 101-probe with a
fully complementary synthetic LacZ sequence. A concentra-
tion range, from 0.133 to 6.00 µg mL–1 of each sequence,
was used in a total volume of 700 µL.
Results and discussion
The rapid detection of microbes in samples of water is be
-
coming more critical as the population of the world in-
creases. Our research group has focused on developing a
biosensor that is rapid and sensitive for the detection of
© 2003 NRC Canada

346
Can. J. Chem. Vol. 81, 2003
Fig. 4. HPLC anion-exchange chromatogram of the mixed base 5′ amine-modified LacZ probe that was grown on CPG and then quan-
titatively removed by base cleavage. The chromatogram indicates the high purity of the synthetic samples used in this investigation.
Fig. 5. MALDI-TOF mass spectrometric analysis of the mixed
base SYBR 101-LacZ probe showing the molecular weight of
the probe m/z (8402 amu). The results confirm the tethering of
the dye to the probe.
the strand terminus to a linker molecule, which in turn is
covalently linked to the substrate. This strategy has proven
advantageous in terms of the enhanced nucleotide coupling
efficiency realized during solid-phase assembly of the oligo-
nucleotide onto the linker and the rapid kinetics of hybrid
formation of the immobilized strand with target sequences.
The single-stranded probe was immobilized via hexaethyl-
ene glycol linker (HEG) to functionalized fused silica sub-
strates. This effectively provided an oligonucleotide surface
where each molecule covered an approximate area of 400–
1300 Å2 when using linear co-polymer strands of ca. 100 Å
lengths (18). This chemistry is highly stable toward water
hydrolysis and physical cleavage, which can occur, owing to
treatment of the surface by heating, cooling, and detergent
washing. The use of HEG also provides advantages based on
high solubility and hydrophilicity (19).
All sensors that were used were checked for quality of
oligonucleotide immobilization by an indirect method based
on concurrent immobilization of DNA on fibres and CPG.
The material on the CPG was cleaved quantitatively from
the surface with base, and anion exchange HPLC was then
used for separation and analysis of recovered material (16).
Figure 1 shows the HPLC analysis of the ssDNA LacZ
probe that was synthesized on CPG. The chromatogram
demonstrates the sequence integrity of the oligonucleotides
that were used throughout this investigation.
Prior to hybridization, sensors were thermally activated by
multiple cycles of heating (90°C) and cooling (30°C) in the
presence of 1 × 10–7 M cDNA. The response of the biosen-
sor to a series of samples introduced sequentially is shown
in Fig. 2. The fluorescence signal was a function of the con-
centration of cDNA. In these experiments, the maximum sig-
nal was obtained at about 120 s after the cDNA was
coliforms as an indicator of microbial contamination in sur-
face and ground water. Immobilized oligonucleotides on
solid supports for use as molecular recognition elements in
bioassays and biosensors are short sequences conjugated at
© 2003 NRC Canada

Almadidy et al.
347
Fig. 6. Spectrofluorimetric scan of SYBR 101-LacZ probe. The probe was treated with an equivalent amount of synthetic cDNA
(LacZ) in the presence of an equivalent number of molecules of salmon sperm ssDNA, in 1 × PBS at 515 nm (room temperature).
Salmon sperm ssDNA (SP-DNA) was prepared by treatment of whole genomic DNA with sonication.
introduced to the sensor. Figure 3 provides an indication of
the shape of the curves and the speed of response. Quantities
as low as about 100 fmole provided signals at the three-
standard-deviation level. Fibres were washed and chemically
regenerated using water at 90°C and 90% formamide solu-
tion in TE buffer between samples. The reproducibility was
excellent, and dozens of cycles of use have been demon-
strated.
Figure 3 shows that the sensor does respond selectively to
the synthetic cDNA LacZ sequence in comparison to the
non-complementary genomic sonicated DNA from salmon
sperm. The fact that the signal for cDNA was present and re-
producible after challenging the biosensor with sonicated
genomic salmon sperm DNA provides an indication that, for
real environmental samples, the possible co-existence of
other non-complementary DNA would not block the biosen-
sor from functioning. Upon challenging the biosensor with
sonicated genomic E. coli DNA, the time dependence of the
signal obtained demonstrates hybridization between the
probe and the genomic LacZ.
longer product have demonstrated that signal magnitude is
somewhat dependent on the location of the target sequence
(20), with the signals of greatest magnitude appearing when
the target sequence is farthest removed from the biosensor
surface.
Tethered dye
In a preliminary experiment that was designed as a first
step towards preparation of tethered dye, a second probe was
constructed in which the fluorescent intercalator SYBR 101
was chemically conjugated to the probe via a tether. The
dye-conjugated probe was purified and examined by anion
exchange HPLC for purity (Fig. 4). The mass spectral analy-
sis confirmed the formation of SYBR 101 labelled LacZ
probe (Fig. 5).
The ability of the L-DNA to hybridize with fully comple-
mentary DNA was shown by the change in fluorescence
when complementary, in comparison to non-complementary,
DNA was added. Importantly, the results confirmed that the
tethered dye could still bind into double-stranded DNA and
achieve a substantial change in quantum yield. A second im-
portant observation was that the fully complementary target
Further results investigating the use of PCR products with
the target sequences located in various positions within a
© 2003 NRC Canada

348
Can. J. Chem. Vol. 81, 2003
Fig. 7. Spectrofluorimetric titration curve of SYBR 101-LacZ probe (5.7 µg mL–1) against concentrations of LacZ cDNA from
0.1 µg mL–1 to 7.7 µg mL–1 in 1 × PBS at 515 nm (room temperature).
could still be detected in the presence of a large background
of salmon sperm DNA. Figure 6 provides spectral informa-
tion about the tethered dye and a summary of the fluorescent
signal due to binding events of the labelled probe at a con-
centration of 1 × 1014 molecules in 700 µL PBS solution.
When the probe was mixed with an equivalent number of
molecules of salmon sperm DNA, the signal remained rela-
tively high.
To observe the response to hybridization at different con-
centrations of target cDNA of the SYBR 101 labelled probe,
a titration curve was generated using 25mer fully comple-
mentary target LacZ. In this experiment, a PBS solution of
cDNA containing 5.7 µg mL–1 (1 × 1014 molecules) was
titrated against the SYBR101-LacZ probe. A maximum of
hybridization was achieved when a stoichiometrically equiv-
alent amount of the labelled probe was added to the cell.
The intensity of fluorescence did not substantially change
beyond the 1:1 stoichiometric equivalence point, indicating
that hybridization was necessary to bring the tethered dye
into close proximity to a duplex for stable intercalation to
occur (Fig. 7).
peared to be largely an electrostatic phenomenon where the
positively charged dye interacted with the relatively concen-
trated DNA at the solid interface. The non-selective adsorp-
tion was largely eliminated by moving to high salt (3 ×
PBS), with the concurrent advantage being that the dsDNA
stability was improved. The background intensity effect pro-
duced by non-complementary DNA could be reduced to less
than 10%. Another limitation that was identified was sensi-
tivity to photobleaching (21), which was easily ameliorated
by use of gated detection. Further research has now demon-
strated that a more complicated time-dependent chemical
process, based on availability of intercalant after denatur-
ation and biosensor regeneration, was the main cause of re-
duction in signal intensity as a series of experiments were
done sequentially.
Conclusions
A fibre-optic biosensor for a portion of the LacZ gene was
constructed as a diagnostic device to provide a surface for
the hybridization with markers from E. coli.
Preliminary work using tethered thiazole orange labels on
short mixed nucleotide probes that were immobilized to the
biosensor surface confirmed that such tethered dyes can re-
port selective hybridization (21). The results also indicated
that adsorption of non-complementary DNA had a signifi-
cant effect on the environment of the tethered TO. This ap-
The LacZ gene of E. coli was selected because conven-
tional coliform monitoring is based on detection of the activ-
ity of the gene product (β galactosidase) produced by
coliform bacteria. Also, the LacZ sequence was selected as a
target because it is specific to total coliforms, while the
lamB gene is within E. coli, Salmonella, and Shigella spp.,
© 2003 NRC Canada

Almadidy et al.
349
4. P.B. Kubek and P.D. Robillard. Drinking water solutions 2.0:
Computer information system. Pennsylvania State University
Department of Agricultural and Biological Engineering., Uni-
versity Park, PA. 1990.
5. American Public Health Association. Standard methods for the
examination of water and wastewater. 18th ed. Washington,
D.C. 1992.
and the uid gene is within E. coli and Shigella spp. The
25mer length of the probe was shown to be sufficiently se-
lective to hybridize genomic target from E. coli bacteria and
was further investigated by searching within the GenBank
nucleotide sequence data to insure that there were no other
homologies with potential non-target sequences (22). The
short length of the probe provides advantages in terms of re-
versibility and high speed of hybridization.
6. A.K. Bej, R.J. Steffan, L. Haff, and R.M.Atlas. Appl. Environ.
Microbiol. 56, 307 (1990).
The LacZ probe was covalently attached through a C6
amino-modifier tether to SYBR 101 fluorescent intercalating
dye. This labelled probe was used free in solution to investi-
gate whether the dye could still participate in intercalation as
hybridization with fully complementary target in the pres-
ence of a high concentration of non-complemetary DNA
(salmon sperm DNA) proceeded. The results are encourag-
ing, as there was no indication that the fluorescent signal
from hybridization was dramatically affected by the pres-
ence of the non-complementary material.
7. M. Thompson and A.K. Deisingh. Analyst, 126, 2153 (2001).
8. J. Wang. Nucleic Acids Res. 28, 3011 (2000).
9. A. Bensimon, A. Simon, A. Chiffaudel, V. Croquette, F.
Heslot, and D. Bensimon. Science, 265, 2096 (1994).
10. F. Caruso, E. Rodda, D.N. Furlong, K. Niikura, and Y.
Okahata. Anal. Chem. 69, 2043 (1997).
11. A.P. Abel, M.G. Weller, G.L. Duveneck, M. Ehrat, and H.M.
Widmer. Anal. Chem. 68, 2905 (1996).
12. S. Fanning, J. O’Mullane, D. O’Meara, A. Ward, C. Joyce, M.
Delaney, and B. Cryan. Br. J. Biomed. Sci. 52, 317 (1995).
13. S. Iqbal, M. Mayo, J. Bruno, B. Bronk, C. Batt, and J. Cham-
bers. Biosens. Bioelectron. 15, 549 (2000).
14. A.H. Uddin, P.A.E. Piunno, R.H. Hudson, M.J. Damha, and
U.J. Krull. Nucleic Acids Res. 25, 4139 (1997).
15. J. Watterson, P.A.E. Piunno, C.C. Wust, S. Raha, and U.J.
Krull. Fresenius J. Anal. Chem. 369, 601 (2001).
16. B. Sojka, P.A.E. Piunno, C.C. Wust, and U.J. Krull. Anal.
Chim. Acta, 395, 273 (1999).
Acknowledgments
This work was financially supported by the Natural Sci-
ences and Engineering Research Council of Canada
(NSERC). We would like to acknowledge Dr. Steven Yue
and Dr. Nabi Malikzada of Molecular Probes Inc. for provi-
sion of the SYBR 101 reactive dye.
17. J.H. Watterson, P.A.E. Piunno, C.C. Wust, and U.J. Krull.
Langmuir, 16, 4984 (2000).
18. Z. Guo, R. Guilfoyle, A. Thiel, R.Wang, and L. Smith. Nucleic
Acids Res. 22, 5456 (1994).
References
1. R. Bordner and J. Winter. Microbiological methods for moni-
toring the environment: Water and wastes. EPA/600/8–78/017.
Environmental Protection Agency, Washington, D.C. 1978.
2. A.F. Gaudy and E. T. Gaudy. Microbiology for environmental
scientists and engineers. McGraw-Hill Book Co., New York.
1980.
19. M.S. Shchepinov, S.C. Case-Green, and E.M. Southern. Nu-
cleic Acids Res. 25, 1155 (1997).
20. A. Almadidy, J. Watterson, P.A.E. Piunno, S. Raha, I.V.
Foulds, P.A. Horgen, A. Castle, and U.J. Krull. Anal. Chim.
Acta, 461, 37 (2002).
21. X. Wang and U.J. Krull. Anal. Chim. Acta, 470, 57 (2002).
22. A.K. Bej, J.L. DiCesare, L. Haff, and R.M. Atlas. Appl. Envi-
ron. Microbiol. 57, 1013 (1991).
3. American Water Works Association. Water quality and treat-
ment. McGraw-Hill, Inc., New York. 1990.
© 2003 NRC Canada