MINIREVIEW
Optical-fiber bundles
Hans H. Gorris, Timothy M. Blicharz and David R. Walt
Department of Chemistry, Tufts University, Medford, MA, USA
Introduction
The focus of biological research has shifted consider-
ably during the last few years. Traditionally, biomole-
cules such as DNA, RNA, peptides, proteins, lipids, or
carbohydrates have been identified and characterized
one by one by isolating and purifying the molecule of
interest. Although this procedure remains important, it
is not practical for gaining a comprehensive view of
the myriad processes in living systems. In the post-
genomic area, it has become apparent that the linear
DNA sequence of a genome holds only a small frac-
tion of the information required for understanding
whole organisms. The need to take a broader picture
represents the motivation for systems biology [1].
The need to make many measurements requires new
analytical tools. High-density arrays represent one of
the tools for making many measurements simulta-
neously and for elucidating complex patterns [2,3].
Arrays consist of a large number of spatially arranged
sensing elements that can be interrogated simulta-
neously for high-throughput and cost-effective mea-
surements. The feature density of arrays is constantly
increasing because the physical size of each sensing
element is decreasing as more sophisticated fabrication
methods become available, e.g. photolithography [4,5],
spotting with piezoelectric [6] or inkjet dispensers [7],
or assembling beads into an array format [8–10].
Arrays can be interrogated by optical, electrochemical,
thermal, and mass-transduction mechanisms. Array
support materials are chosen according to the trans-
duction mechanism, but also to minimize nonspecific
interactions with the target molecules [11]. With the
availability of more computing power for data storage
and processing to apply to these arrays, the informa-
tion that can be collected is staggering.
This minireview focuses on optical-fiber bundles
(Fig. 1), an array format that enables high-density and
multianalyte sensing [12–16]. In contrast to conven-
tional single fiber-optic biosensors that have been
reviewed previously [17–19], optical-fiberbundles con-
sist of thousands of individual glass or plastic fibers.
Individual fibers are bundled, melted, and pulled
through a fiber drawing tower in an iterative process
to fuse the individual fibers into a unitary substrate.
Each of the fiber cores is surrounded by a cladding
material of lower refractive index such that an optical
signal is transmitted by total internal reflection within
the fiber core. Each fiber acts as an independent wave-
guide that enables light to be carried over long
Keywords
array; artificial olfaction; bead; cell; DNA;
lab-on chip; optical-fiber bundle; single
molecule
Correspondence
D. Walt, Department of Chemistry, Tufts
University, 62 Talbot Avenue, Medford,
MA 02155, USA
Fax: +1 617 627 3443
Tel: +1 617 627 3470
E-mail: david.walt@tufts.edu
(Received 9 July 2007, accepted 29 August
2007)
doi:10.1111/j.1742-4658.2007.06078.x
Optical-fiber bundles have been employed as a versatile substrate for the
fabrication of high-density microwell arrays. In this minireview, we discuss
the application of optical-fiber-bundle arrays for a variety of biological
problems. For genomics studies and microbial pathogen detection, individ-
ual beads have been functionalized with DNA probes and then loaded into
the microwells. In addition, beads differentially responsive to vapors have
been employed in an artificial olfaction system. Microwell arrays have also
been loaded with living cells to monitor their individual response to biolog-
ically active compounds over long periods. Finally, the microwells have
been sealed to enclose single enzyme molecules that can be used to measure
individual molecule catalytic activity.
5462 FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS
distances with minimal attenuation. A typical fiber
array consists of a few thousand to 100 000 individual
fibers, with an overall bundle diameter of < 1 mm.
The individual fiber size can be specified and can range
between 2 and 20 lm.
As each individual fiber in the array maintains its
relative position throughout the bundle, the bundle
can be cut into pieces of any length. To make a high-
density sensor array, the bundle surface is first
polished on both ends, and a homogenous array of
microwells is formed at one end of the fiber bundle by
selectively acid etching the fiber cores [8]. The resulting
wells have the same diameter as the fiber cores and the
depth of the wells is dependent on the acid, its concen-
tration, the exposure time, and the core material.
Microwell volume can be tailored by etching to differ-
ent depths. A microwell typically has a volume of a
few tens of fL (lm
3
) and each well can be loaded with
DNA-, antibody-, or dye-coupled beads (Fig. 2A), or
even living cells (Fig. 2B). Furthermore, microwells
can be sealed to form microchambers enclosing single
molecules of analyte.
The optical-fiber bundle can be mounted on a
microscope and all microwells can be observed simul-
taneously using fluorescence microscopy. Excitation
light is introduced into the proximal end of the fiber
and excites fluorescent molecules located at the distal
end of the fiber array. Changes in analyte concentra-
tion result in a change in fluorescence intensity or
emission wavelength, and the light from each micro-
well is propagated back through the fiber to the detec-
tor. After filtering the excitation light, the fluorescence
emission is projected onto a charge-coupled device
camera.
Nucleic acid analysis
The complexity of genomic analysis has stimulated the
development of multiplexed nucleic acid arrays that
can rapidly and efficiently analyze vast amounts of
genetic information. An inherent trait of nucleic acids
is their highly specific base pair recognition, which is
employed in DNA microarray technology, where
immobilized single-stranded oligonucleotide probes are
used to detect complementary target strands. A num-
ber of optical fiber arrays for fluorescence-based oligo-
nucleotide detection have been described [10,20–22].
Optical-fiber bundles are excellent platforms for DNA
array fabrication, because they possess a very high fea-
ture density with the capability to collect thousands of
signals simultaneously. DNA arrays can be constructed
from fiber-optic bundles simply by chemically etching
the fiber cores and randomly depositing individual
beads into the resulting microwells. Each bead is func-
tionalized with several hundred thousand copies of a
particular single-stranded oligonucleotide probe mole-
cule. Different beads can be modified with different
oligonucleotide probes to increase multiplexing ability.
Each bead type can be encoded with one or more fluo-
rescent dyes for identifying or registering its location
on the array [23,24]. Bead registration has also been
demonstrated by linking unique sequence markers to
each bead type and decoding the bead array using a
series of hybridization reactions with fluorescently
A
B
Fig. 1. Scheme of an optical-fiber bundle. (A) A typical optical-fiber
bundle consists of a few thousand to 100 000 individually address-
able fibers that share a common cladding material (black).
(B) Because of the difference between the refractive indices of the
core and cladding material, light propagates along the entire fiber
length by total internal reflection and cannot escape from individual
fibers. Light is transmitted in both directions such that each fiber can
act as a waveguide for the excitation as well as the emission signals.
AB
Fig. 2. Atomic force micrographs of uni-
formly etched fiber bundles. (A) Microwells
of 3.1 lm diameter and 2 lm depth filled
with 3.1 lm diameter beads. (B) Microwells
of 22 lm diameter and 15 lm depth loaded
with single mammalian cells (Chinese ham-
ster ovary cells).
H. H. Gorris et al. Optical-fiber bundles
FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS 5463
labeled complementary target decoding DNA strands
[15]. Bead-based fiber-optic arrays offer several advan-
tages over traditional spotted DNA arrays [15,23,25].
Optical fiber DNA arrays are randomly assembled
from a bead-probe pool using a facile wet loading
method, in contrast to mechanical spotting methods
used for traditional patterned DNA arrays. Because of
the large number of beads in a preparation
(> 10
11
Æg
)1
), the same bead pool can be used to fabri-
cate hundreds to thousands of optical fiber arrays with
high uniformity. A single bead-based optical fiber
DNA array can be reused over many hybridization–
dehybridization cycles without significant signal degra-
dation [23]. In addition, the content of bead arrays is
flexible, because any desired combination of oligonu-
cleotide probes can be mixed from a library of beads
with different sequence specificities and included in a
bead pool. In contrast, patterned arrays require sub-
stantial modifications in their fabrication protocols to
change the probe content. Finally, the density of
detection elements in a fiber-optic microwell array
( 25 000Æmm
)2
) is much higher with a smaller foot-
print compared with that of spotted DNA arrays,
enabling smaller sample volumes to be used for analy-
sis. The high density of optical-fiber-bundle arrays also
allows multiple replicates of the same probe to be ana-
lyzed in each array, improving the signal-to-noise
(S ⁄ N) ratio of measurements [26]. Along with genotyp-
ing and gene expression applications that are common
to DNA array technology, optical-fiber bundle DNA
arrays can also be used for rapid and sensitive detec-
tion of biological warfare agents as well as food and
waterborne pathogens.
In one study, a fiber-optic DNA array was con-
structed to detect six biological warfare agents [24].
The multiplexed array was fabricated with 18 different
probes that included specificity for multiple strains of
bacteria. DNA from autoclaved bacterial culture sam-
ples was isolated, amplified with fluorescently labeled
primers via PCR, and then used to spike wastewater
samples. The array platform was used to correctly
identify the biological warfare agent target DNA in
30 min with a detection limit of 10 fm. Extremely low
detection limits can be achieved with fiber-optic micro-
well DNA array platforms, as one study demonstrated
the detection of 100 am of fluorescently labeled syn-
thetic target DNA, corresponding to 600 molecules
in a 10 lL sample volume [26].
Unlabeled chromosomal DNA from Salmonella spp.
was detected with a fiber-bundle array using a sand-
wich-type assay [25]. In this assay format, the bacterial
DNA was first captured by complementary probe-
functionalized beads in the bundle. Detection was per-
formed after incubating the array with a solution of
fluorescently labeled signal probes that were comple-
mentary to another region of the bacterial DNA.
Using this format, 10
3
)10
4
cfuÆmL
)1
of Salmonella
could be detected after 1 h hybridization without
DNA amplification, even in the presence of potentially
interfering organisms such as Escherichia coli and Yer-
sinia enterocolitica. A similar sandwich-type assay was
conducted using a fiber-optic DNA array to detect the
harmful algal bloom species Alexandrium fundyense,
A. ostenfeldii, and Pseudo-nitzschia australis [27]. Ribo-
somal RNA from these three species was simulta-
neously detected with bead-based probes specific to the
respective organisms. Detection limits as low as five
cells were observed, due to the high copy number of
rRNA per cell ( 8 · 10
6
in A. fundyense).
Genotyping arrays can also be important when
profiling pathogenic microorganisms because different
strains of the same bacterial species can be either
benign or virulent. Serotype differentiation can there-
fore be crucial in pathogen detection. Using an array
of only six bead-based sequence probes, a fiber-optic
bundle DNA array was used to discriminate 12 dif-
ferent strains of E. coli. This efficiency was accom-
plished by selecting probe sequences specific for both
virulent and nonpathogenic strains, such that a binary
yes ⁄ no pattern generated from all six probes could be
used to classify each strain. In principle, an array
with only six probes should have the capability to
discern 2
6
¼ 64 different strains using this method.
With a similar strategy, it should be possible to apply
this platform to other genomic studies, such as detect-
ing single- or multiple-nucleotide polymorphisms or
insertions ⁄ deletions.
Fiber-optic bundles have been used in a variety of
DNA array applications. The flexible array fabrication
procedure allows customizable array content, high
degrees of multiplexing, and the capability to detect
amplified or unamplified samples in multiple assay for-
mats with extremely low detection limits. With these
advantages, fiber-optic bundle DNA array platforms
show great promise for rapid pathogen detection, gene
expression, and genotyping studies.
Laboratory-on-a-chip arrays for
biomarker screening and diagnostics
Microfluidic devices are now considered a common
tool for various natural and life science applications
[28]. Microfluidics, which involves the manipulation of
small fluid volumes at the lL and nL scale, has the
capability to reduce sample, reagent, and assay time
requirements, as well as improve assay detection limits.
Optical-fiber bundles H. H. Gorris et al.
5464 FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS
The use of microfluidic systems also permits assay
automation, which can reduce human error associated
with multistep protocols. The fusion of a multiplexed
detection platform with microfluidics creates a unified
analytical device that is commonly referred to as a lab-
on-a-chip.
The high feature density and small footprint of
fiber-optic bundles allow them to be integrated into
microfluidic platforms that may lead to lab-on-a-chip
devices. A study was conducted where a bead-based
fiber-optic DNA array was integrated with a microflui-
dic sample delivery platform and was used to detect
10 am of fluorescently labeled target DNA in 15 min.
The hybridization efficiency was substantially
improved with the microfluidic platform compared
with static measurements [29].
The optical-fiber bundle platform is flexible and has
been applied to nucleic acid-based detection, as dis-
cussed above, as well as protein detection using multi-
plexed immunoassays [30,31]. A logical extension of
this technique would be the multiparameter measure-
ment of both nucleic acid and protein biomarkers
within the same array. Our laboratory is currently
involved in a multi-institution collaboration with the
goal of developing a universal platform for the multi-
plexed detection of proteins and nucleic acid markers
implicated in the exacerbation of obstructive pulmo-
nary inflammatory diseases, such as asthma and
chronic obstructive pulmonary disease, using saliva as
a noninvasive diagnostic fluid [32].
Multiplexed optical-fiber bundle assays are ideal for
diagnostics applications, because there are numerous
proteins and DNA sequences that have been identified
as potential biomarkers in various diseases. A desirable
diagnostics platform for pulmonary inflammatory dis-
eases would monitor changes in inflammatory proteins
when patients are in exacerbation relative to their nor-
mal state. In addition, it would be extremely useful to
concurrently monitor the presence of bacteria (e.g.
Haemophilus influenzae) or viruses (e.g. rhinovirus,
respiratory syncytial virus) that could trigger an exac-
erbation. The lab-on-a-chip device we are developing
will incorporate both multiplexed protein and nucleic
acid bead-based detection elements with an automated
sample delivery system and an optical platform that
will be small enough for point-of-care testing. With
increasing knowledge about numerous protein and
nucleic acid markers associated with different disease
states, progress in disease research can be made more
rapidly and more effective treatments can be discov-
ered. Furthermore, the information available from a
point-of-care device capable of multiparameter mea-
surements could help physicians to better adjust the
therapy for individual patients and assess treatment
effectiveness. Microarrays, such as the fiber-optic bun-
dle arrays reviewed here, are promising tools for future
diagnostic applications.
Adaptive sensing: artificial olfaction
In the previous sections, we have considered arrays in
which each analyte is measured by a highly specific
sensor. Although this type of analyte sensing offers the
highest selectivity, it is not useful for detecting complex
mixtures such as odors and flavors, where the smell or
taste is defined not by a single component, but by the
entire composition. For most organisms, the relevant
information about food, attractants, and repellents is
encoded in such mixtures.
The vertebrate olfactory system is based on highly
versatile adaptive sensing that can recognize and evalu-
ate many complex odor patterns. The olfactory system
contains millions of neuron receptor cells in the olfac-
tory epithelium. Every cell expresses one out of 1000
different types of receptors [33,34]. The receptors
respond differentially to a wide variety of vapors, and a
single pure organic vapor triggers a response in about
50% of the cells [35]. Thus, 1000 receptor types can
decode a nearly infinite number of patterns using a
combinatorial code. Millions of neurons are wired to
the olfactory bulb, which acts as a preprocessor for the
incoming olfactory signals and sends a reduced number
of signals to the higher brain structures where the final
odor identification is carried out by neuronal networks.
A multisensor array based upon mammalian olfac-
tion principles was first introduced by Persaud and
Dodd [36], and fiber-optic fluorescence sensors for the
detection of pure vapors were used by the Wolfbeis
group [37–39]. Our group has combined both
approaches in an artificial olfaction system based on
fiber-optic bundles [40]. Other groups using fluores-
cence techniques have since adapted similar strategies
[41–43]. Although fluorescence sensor arrays are most
common, some groups use absorbance measurements,
for example, with porphyrin ring systems that show a
chromogenic shift upon vapor exposure [44–46].
Our group employs a solvatochromic dye indicator
immobilized in a variety of polymer beads as receptors.
These receptors respond differentially to various odors
by changes in their fluorescence. The emission spec-
trum of the solvatochromic dye Nile Red changes with
its local environment. Each different polymer bead
type sets a baseline polarity for the dye. The polymer
beads are deposited in the glass-fiber bundle and pat-
tern recognition is accomplished with artificial compu-
tational networks. When the different beads are
H. H. Gorris et al. Optical-fiber bundles
FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS 5465
exposed to a vapor, their spectral properties change
(Fig. 3). For example, Nile Red entrapped in a rela-
tively nonpolar bead will exhibit a red shift when
exposed to a more polar vapor. Conversely, Nile Red
entrapped in a highly polar bead will exhibit a blue
shift when exposed to the same vapor as long as its
polarity is less than that of the polymer. The optical
properties of the immobilized dye are also influenced
by other factors, such as the pore size of the bead and
its swelling tendency. Time traces of the changes in
these optical properties of multiple beads are recorded
simultaneously through the optical-fiber bundle. A
charge-coupled device camera is used to collect the
time responses and different computational networks
are used for classifying the vapors [47].
The different receptor bead types in the array are
randomly distributed, but each bead is defined by a
characteristic change in its optical properties upon
exposure to a known test vapor. Thus, the beads are
‘self-encoded’ and can be easily identified. Because the
response patterns from all receptors within the entire
array can be combined to create one profile per odor
stimulus, decoding each bead is not necessary, but the
decoding improves the sensing accuracy [48].
Self-encoded bead receptors in fiber-optic bundles
have revolutionized artificial olfaction. The small bead
size results in a large surface-to-volume ratio, which
enables good vapor interactions, rapid responses, and
a high sensitivity. The sensitivity of artificial olfaction
is further enhanced through sensor redundancy as
each receptor is represented by thousands of beads,
such that the signals from a particular type of recep-
tor can be combined [49]. A similar effect is used by
the vertebrate olfaction system where each receptor is
expressed on a large number of neuron cells and the
combined signals from identical cell types improve
sensitivity.
In an ideal artificial olfaction system, the receptors
would be completely regenerated after each exposure,
but as in natural systems, the receptors degrade with
time. A problem inherent to fluorescent dyes is photo-
bleaching. We have been able to reduce the effects of
photobleaching by illuminating subsections of the bead
array and slowly increasing the light exposure as dye is
depleted [50]. Another weakness of previous artificial
olfactory arrays lies in the network training, which
cannot be transferred from one array to another, as
there is too much variation in the array preparation.
The bead-based approach, however, allows new beads
to be reloaded from libraries, in which billions of vir-
tually identical beads from one preparation are stored,
such that the same training can be used from array to
array [51].
Such bead array optical sensors were able to recog-
nize the presence or absence of nitroaromatic explosive-
like compounds in the presence of varying concentra-
tions of background vapors at levels over 100 000 times
higher than the explosive vapor [52]. Furthermore,
three pure odors (toluene, acetone, and 1,3-dinitrotolu-
ene) and three complex odors (e.g. coffee beans) were
classified with 100% accuracy when measured at their
highest relative concentrations. At lower concentra-
tions, the classification was still better than 85% [40].
These examples of artificial olfaction are important for
both security applications and the food industry, but
there are many more potential applications for adaptive
sensing systems that remain to be explored.
Cellular analysis
Fiber-optic microwell arrays have also been applied to
cellular analyses, primarily for environmental toxicity
studies [53–55]. Cell-based arrays provide a unique
capability, because live cells can demonstrate dynamic
responses to a wide variety of biologically active com-
pounds. In analogy to bead-based arrays, individual
cells can be randomly deposited in the microwells of
an optical fiber array. The viability of the cells is main-
tained by keeping the fiber bundle surface containing
the cells submerged in medium. When cell arrays are
stored under the proper conditions, they have been
shown to have lifetimes of up to 14 days [56]. Because
of the individually addressable optics of the fiber
bundle, many single cells can be observed simulta-
neously and multiple cell types can be discriminated by
encoding them with different fluorescent dyes [57].
Fig. 3. Differential responses to a single vapor by three types of
bead sensors. Each sensor type (depicted in different colors) is rep-
resented by multiple identical beads in a fiber-bundle array. The
beads were exposed to dimethyl methylphosphonate for 1.6 s (gray
panel). The responses were monitored at a single wavelength.
Optical-fiber bundles H. H. Gorris et al.
5466 FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS
Generally, cells used in these assays are genetically
modified to express either fluorescent proteins or
enzymes that catalyze the formation of fluorescent
products [56,58]. Whereas bulk cellular analyses are
typically conducted in microtiter plates and only moni-
tor the average response from thousands of cells, fiber-
optic cell arrays can be used to observe the responses
from both individual cells and an entire population of
cells, permitting differences within a population to be
observed and analyzed. Flow cytometric assays [59]
are useful for cell sorting and the analysis of single-cell
variation within large populations, but only provide
information at a single time point. Fiber-optic bundle
microwell arrays, however, allow for monitoring gene
and protein expression kinetics as well as toxicity
screening from a large population of individual cells
over long periods.
The capability of the fiber-optic cell array platform
to discriminate numerous individual cells simulta-
neously was initially demonstrated using mouse fibro-
blast cells encoded with three different lipophilic dyes
[57]. The metabolic activities of these cells resulted in
extracellular pH changes that were detected with nano-
particles embedded with a pH-sensitive fluorescent
indicator.
A fiber-bundle assay was used to conduct lacZ
reporter gene expression on three different yeast two-
hybrid strains [60]. Each cell type (positive control,
negative control, and wild-type) was labeled with a dif-
ferent fluorescent dye-conjugated lectin to identify its
location on the array. Following incubation with a flu-
orogenic b-galactosidase substrate, the locations of the
three cell types on the array were found to match areas
where high or low levels of b-galactosidase were
expressed, as indicated by the catalytic production of a
fluorescent product. Interestingly, variability in the
level of b-galactosidase expression was observed within
the positive and negative control groups. A more
detailed analysis of this system with a larger number
of cell genotypes and stringency conditions confirmed
that a distribution of responses can exist for a seem-
ingly homogeneous cell population, illustrating the
utility of single cell measurements in gene expression
studies [61]. Similarly, gene expression kinetics and
genetic noise were monitored with a fiber-bundle array
containing two E. coli strains carrying different gene
fusion constructs [58]. The two cell types were modi-
fied to express green fluorescent protein when either
the lacZ or recA promoters were induced. Variation in
gene expression rates for hundreds to thousands of
individual cells was observed over time. The study
showed that a substantial amount of information
could be gleaned about gene expression kinetics and
cell-to-cell variation within a population using fiber-
optic bundles.
Cell arrays serve as functional assays in that they
measure both bioavailability and efficacy, in contrast
to most binding assays that simply measure affinity. A
living cell biosensor was developed to conduct toxicity
screening using a fiber-optic microwell array loaded
with genetically modified E. coli cells [62]. The lacZ
reporter gene was fused to the mercury-responsive gene
promoter zntA. After incubation in medium containing
different amounts of HgCl
2
followed by incubation
with a fluorogenic b-galactosidase substrate, the
authors observed increased enzyme expression in cells
exposed to higher concentrations of Hg
2+
. This
method was used to detect Hg
2+
levels down to
100 nm. In another study, a biosensor array of E. coli
cells modified with a recA::gfp fusion plasmid was
used to detect several genotoxins, including mitomy-
cin C at concentrations as low as 1 ngÆmL
)1
[56].
Fiber-optic bundles have also been used for cell
migration studies [63]. Migration assays are used to
identify and evaluate compounds that prevent tumor
cell migration, an important aspect of cancer metasta-
sis. In this case, the proximal end of an unetched opti-
cal-fiber bundle was modified with the cell adhesion
protein fibronectin, and fluorescently labeled mouse
fibroblast cells were allowed to adhere to the modified
fiber-bundle surface. By monitoring the fluorescent
cells through the fibers, their movement could be
tracked as they traversed the surface of the bundle.
Antimigratory agents slowed the movement of cells on
the surface. Using this method, the effect of an anti-
migratory drug could be determined in minutes, in con-
trast to the several hours required for more common
migration assay techniques.
Optical-fiber bundles containing living cells are a
potentially valuable platform for high-throughput
drug-screening applications because they enable the
monitoring of response dynamics from single cells as
well as from entire cell populations. For example, cell-
to-cell variations within a population after exposure to
a drug or a combination of drugs might provide useful
information about the efficacy of the drugs as well as
rare effects on outlier cells. Such ‘promiscuous’ drug
effects were observed using a duplexed living cell array
containing two genetically modified E. coli strains [64].
The response dynamics of both strains to two cyto-
toxic drugs were observed over time, permitting a
more comprehensive understanding of how the drugs
affected each cell type.
Living cellular arrays constructed with optical-fiber
bundles can be used for a variety of high-throughput
screening, gene expression, and biosensor applications.
H. H. Gorris et al. Optical-fiber bundles
FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS 5467
The unique nature of the array permits the simulta-
neous observation of thousands of individual cells
comprised of one or more cell types, and offers distinct
advantages over traditional cell-based assays.
Single molecules
A new and emerging field for optical-fiber-bundle
arrays is their use for single molecule studies. Single
molecule measurements provide unique information
about heterogeneous molecular behaviors that are
hidden using bulk methods in which the behaviors of
vast numbers of molecules are averaged [65,66]. The
microwells of a fiber-bundle array can be sealed
mechanically with a silicone gasket to form micro-
chambers containing fL volumes. Enclosing single
enzyme molecules in ultrasmall reaction chambers is a
straightforward method that requires no enzyme
immobilization [67]. If an appropriately dilute solution
is enclosed in these microchambers, individual mole-
cules can be isolated. We employed the enzyme
b-galactosidase at a concentration such that each
microchamber contained either a single enzyme mole-
cule or no enzyme molecule [68]. The enzyme solution
also contained the fluorogenic substrate resorufin-
b-galactopyranoside. Each enzyme in the array of
microchambers was detected individually by its cata-
lytic activity, which resulted in the production of fluo-
rescent resorufin. In microchambers containing an
enzyme molecule, the product accumulated to high
local concentrations and was detected through the
optical fibers. The ratio of fluorescent to nonfluores-
cent microchambers yielded a digital readout of the
enzyme. When a slow-binding inhibitor was added to
b-galactosidase, stochastic events of inhibitor release
and binding could be observed by changes in the cata-
lytic activity of single enzyme molecules [69]. The
stochastic behavior of the single enzyme molecules
agreed well with results derived from bulk reactions.
An alternative approach for loading single enzyme
molecules into the glass-fiber array is to capture
ligand-labeled enzymes by affinity binding. For this
purpose, we modified the surfaces of the microwells
with streptavidin to capture biotin-labeled b-galactosi-
dase [70]. Because of the high affinity of the biotin–
streptavidin binding pair, it was possible to capture
and observe single molecules when only 3 am of
enzyme were present in a sample.
Conclusion
Since we first reported the use of optical-fiber bundles
for sensing arrays [71], they have been employed for a
variety of applications. Optical-fiberbundles are read-
ily available for array fabrication and biosensing.
Because of the array’s high feature density, small sam-
ple volumes can be investigated with a large number of
different sensors. Furthermore, optical-fiber bundles
can be fabricated using a variety of sensor materials,
such as beads, cells, or single molecules. The optical-
fiber bundle enables the use of fluorescent indicators
that can be readily detected with standard fluorescence
microscopes – a common piece of equipment in most
laboratories. Signal transduction via total internal
reflection also provides the opportunity to separate
sensing and detection elements over a long distance.
This flexibility is useful when examining harmful
agents that need to be handled at a distance. A major
additional advantage of fiber-bundle arrays is the abil-
ity to add more sensing elements from a library to an
array without synthesizing an entirely new array. In
this minireview, we have demonstrated that arrays
based on optical-fiberbundles allow for multiplexed
measurements and are an ideal tool for the analysis of
complex mixtures. The flexibility of the fiber-bundle
array format distinguishes it for future applications in
chemical and biological analyses.
Acknowledgements
The authors thank Ragnhild Dragoy Whitaker, Chris-
topher LaFratta, and Matthew Aernecke for providing
figures for this manuscript.
References
1 Hood L & Galas D (2003) The digital code of DNA.
Nature 421, 444–448.
2 Werner T (2007) Regulatory networks: linking micro-
array data to systems biology. Mech Ageing Dev 128,
168–172.
3 Albeck JG, MacBeath G, White FM, Sorger PK, Lauf-
fenburger DA & Gaudet S (2006) Collecting and orga-
nizing systematic sets of protein data. Nat Rev Mol Cell
Biol 7, 803–812.
4 Fodor SPA, Read JL, Pirrung MC, Stryer L, Lu AT &
Solas D (1991) Light-directed, spatially addressable par-
allel chemical synthesis. Science 251, 767–773.
5 Warrington JA, Shah NA, Chen X, Janis M, Liu C,
Kondapalli S, Reyes V, Savage MP, Zhang Z, Watts R
et al. (2002) New developments in high-throughput
resequencing and variation detection using high density
microarrays. Hum Mutat 19, 402–409.
6 Lin H, Ho AS, Haley-Vicente D, Zhang J, Bernal-Fus-
sell J, Pace AM, Hansen D, Schweighofer K, Mize NK
& Ford JE (2001) Cloning and characterization of
Optical-fiber bundles H. H. Gorris et al.
5468 FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS
IL-1HY2, a novel interleukin-1 family member. J Biol
Chem 276, 20597–20602.
7 Shoemaker DD, Schadt EE, Armour CD, He YD, Gar-
rett-Engele P, McDonagh PD, Loerch PM, Leonardson
A, Lum PY, Cavet G et al. (2001) Experimental
annotation of the human genome using microarray tech-
nology. Nature 409, 922–927.
8 Pantano P & Walt DR (1996) Ordered nanowell arrays.
Chem Mater 8, 2832–2835.
9 Brenner S, Johnson M, Bridgham J, Golda G, Lloyd
DH, Johnson D, Luo S, McCurdy S, Foy M, Ewan M
et al. (2000) Gene expression analysis by massively par-
allel signature sequencing (MPSS) on microbead arrays.
Nat Biotechnol 18, 630–634.
10 Yeakley JM, Fan JB, Doucet D, Luo L, Wickham
EYeZ, Chee MS & Fu XD (2002) Profiling alternative
splicing on fiber-optic arrays. Nat Biotechnol 20, 353–
358.
11 Angenendt P, Glokler J, Murphy D, Lehrach H &
Cahill DJ (2002) Toward optimized antibody micro-
arrays: a comparison of current microarray support
materials. Anal Biochem 309, 253–260.
12 Walt DR (1998) Fiber-optic imaging sensors. Acc Chem
Res 31, 267–268.
13 Steemers FJ & Walt DR (1999) Multi-analyte sensing
from site-selective deposition to randomly-ordered
addressable optical sensors. Microchim Acta 131, 99–105.
14 Walt DR (2002) Imaging optical sensor arrays. Curr
Opin Chem Biol 6, 689–695.
15 Epstein JR, Ferguson JA, Lee K-H & Walt DR (2003)
Combinatorial decoding: an approach for universal DNA
array fabrication. J Am Chem Soc 125, 13753–13759.
16 Brogan KL & Walt DR (2005) Optical fiber-based sen-
sors: application to chemical biology. Curr Opin Chem
Biol 9, 494–500.
17 Marazuela D & Moreno-Bondi MC (2002) Fiber-optic
biosensors – an overview. Anal Bioanal Chem 372, 664–
682.
18 Gauglitz G (2005) Direct optical sensors: principles and
selected applications. Anal Bioanal Chem 381, 141–155.
19 Bosch ME, Sanchez AJR, Rojas FS & Ojeda CB (2007)
Recent development in optical fiber biosensors. Sensors
7, 797–859.
20 Piunno PAE, Krull UJ, Hudson RHE, Damha MJ &
Cohen H (1995) Fiberoptic DNA sensor for fluoro-
metric nuclei acid determination. Anal Chem 67,
2635–2643.
21 Abel AP, Weller MG, Duveneck GL, Ehrat M & Wid-
mer HM (1996) Fiber-optic evanescent wave biosensor
for the detection of oligonucleotides. Anal Chem 68,
2905–2912.
22 Watterson JH, Raha S, Kotoris CC, Wust CC, Ghara-
baghi F, Jantzi SC, Haynes NK, Gendron NH, Krull
UJ, Mackenzie AE et al. (2004) Rapid detection of sin-
gle nucleotide polymorphisms associated with spinal
muscular atrophy by use of a reusable fibre-optic bio-
sensor. Nucleic Acids Res 32 XX–XX.
23 Ferguson JA, Steemers FJ & Walt DR (2000) High-den-
sity fiber-optic DNA random microsphere array. Anal
Chem 72, 5618–5624.
24 Song L, Ahn S & Walt DR (2006) Fiber-optic micro-
sphere-based arrays for multiplexed biological warfare
agent detection. Anal Chem 78, 1023–1033.
25 Ahn S & Walt DR (2005) Detection of Salmonella spp.
using microsphere-based, fiber-optic DNA microarrays.
Anal Chem 77, 5041–5047.
26 Epstein JR, Lee M & Walt DR (2002) High-density
fiber-optic genosensor microsphere array capable of ze-
ptomole detection limits. Anal Chem 74, 1836–1840.
27 Ahn S, Kulis DM, Erdner DL, Anderson DM & Walt
DR (2006) Fiber-optic microarray for simultaneous
detection of multiple harmful algal bloom species. Appl
Envir Microbiol 72, 5742–5749.
28 Dittrich PS, Tachikawa K & Manz A (2006) Micro
total analysis systems. Latest advancements and trends.
Anal Chem 78, 3887–3907.
29 Bowden M, Song L & Walt DR (2005) Development of
a microfluidic platform with an optical imaging micro-
array capable of attomolar target DNA detection. Anal
Chem 77, 5583–5588.
30 Rissin DM & Walt DR (2006) Duplexed sandwich
immunoassays on a fiber-optic microarray. Anal Chim
Acta 564, 34–39.
31 Blicharz TM & Walt DR (2006) Detection of inflamma-
tory cytokines using a fiber optic microsphere immuno-
assay array. Proc SPIE Int Soc Opt Eng 6380,
638010 ⁄ 1–638010 ⁄ 638010 ⁄ 6.
32 Walt DR, Blicharz TM, Hayman RB, Rissin DM, Bow-
den M, Siqueira WL, Helmerhorst EJ, Grand-Pierre N,
Oppenheim FG, Bhatia JS et al. (2007) Microsensor
arrays for saliva diagnostics. Ann NY Acad Sci 1098,
389–400.
33 Halasz N (1980) The Vertebrate Olfactory System:
Chemical Neural Anatomy, Function and Development.
Akade
´
miai Kiado
´
, Budapest.
34 Buck L & Axel R (1991) A novel multigene family may
encode odorant receptors: a molecular basis for odor
recognition. Cell 65, 175–187.
35 Kauer JS (1991) Contributions of topography and par-
allel processing to odor coding in the vertebrate olfac-
tory pathway. Trends Neurosci 14, 79–85.
36 Persaud K & Dodd G (1982) Analysis of discrimination
mechanisms in the mammalian olfactory system using a
model nose. Nature 299, 352–355.
37 Posch HE, Wolfbeis OS & Pusterhofer J (1988) Optical
and fibre-optic sensors for vapors of polar solvents.
Talanta 35, 89–94.
38 Wolfbeis OS, Posch HE & Kroneis HW (1985) Fiber
optical fluorosensor for determination of halothane
and ⁄ or oxygen. Anal Chem 57, 2556–2561.
H. H. Gorris et al. Optical-fiber bundles
FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS 5469
39 Wolfbeis OS & Sharma A (1988) Fibre-optic fluorosen-
sor for sulfur dioxide. Anal Chim Acta 208, 53–58.
40 Albert KJ, Walt DR, Gill DS & Pearce TC (2001) Opti-
cal multibead arrays for simple and complex odor dis-
crimination. Anal Chem 73, 2501–2508.
41 Bakken GA, Kauffman GW, Jurs PC, Albert KJ & Stit-
zel SS (2001) Pattern recognition analysis of optical sen-
sor array data to detect nitroaromatic compound
vapors. Sens Actuators B Chem 79, 1–10.
42 Cho EJ & Bright FV (2001) Optical sensor array and
integrated light source. Anal Chem 73, 3289–3293.
43 Kermani BG, Fomenko I, Kotseroglou T, Forood B,
Clark L, Barker D & Lebl M (2006) Decoding beads in
a randomly assembled optical nose. Sens Actuators B
Chem 117, 282–285.
44 Di Natale C, Salimbeni D, Paolesse R, Macagnano A &
D’Amico A (2000) Porphyrins-based opto-electronic
nose for volatile compounds detection. Sens Actuators B
Chem 65, 220–226.
45 Filippini D, Alimelli A, Di Natale C, Paolesse R, D’Ami-
co A & Lundstrom I (2006) Chemical sensing with famil-
iar devices. Angew Chem Int Ed 45, 3800–3803.
46 Rakow NA & Suslick KS (2000) A colorimetric sensor
array for odour visualization. Nature 406, 710–713.
47 Jurs PC, Bakken GA & McClelland HE (2000) Compu-
tational methods for the analysis of chemical sensor
array data from volatile analytes. Chem Rev 100, 2649–
2678.
48 Albert KJ & Walt DR (2003) Information coding in
artificial olfaction multisensor arrays. Anal Chem 75,
4161–4167.
49 Dickinson TA, Michael KL, Kauer JS & Walt DR
(1999) Convergent, self-encoded bead sensor arrays in
the design of an artificial nose. Anal Chem 71, 2192–
2198.
50 Bencic-Nagale S & Walt DR (2005) Extending the lon-
gevity of fluorescence-based sensor arrays using adaptive
exposure. Anal Chem 77, 6155–6162.
51 Stitzel SE, Cowen LJ, Albert KJ & Walt DR (2001)
Array-to-array transfer of an artificial nose classifier.
Anal Chem 73, 5266–5271.
52 Albert KJ & Walt DR (2000) High-speed fluorescence
detection of explosives-like vapors. Anal Chem 72,
1947–1955.
53 Corbisier P, van der Lelie D, Borremans B, Provoost A,
de Lorenzo V, Brown NL, Lloyd JR, Hobman JL, Cso-
regi E, Johansson G et al. (1999) Whole cell- and pro-
tein-based biosensors for the detection of bioavailable
heavy metals in environmental samples. Anal Chim Acta
387, 235–244.
54 Hakkila K, Green T, Leskinen P, Ivask A, Marks R &
Virta M (2004) Detection of bioavailable heavy metals
in EILATox-Oregon samples using whole-cell lumines-
cent bacterial sensors in suspension or immobilized onto
fibre-optic tips. J Appl Toxicol 24, 333–342.
55 Ivask A, Green T, Polyak B, Mor A, Kahru A, Virta M
& Marks R (2007) Fibre-optic bacterial biosensors and
their application for the analysis of bioavailable Hg and
As in soils and sediments from Aznalcollar mining area
in Spain. Biosens Bioelectron 22, 1396–1402.
56 Kuang Y, Biran I & Walt DR (2004) Living bacterial
cell array for genotoxin monitoring. Anal Chem 76,
2902–2909.
57 Taylor LC & Walt DR (2000) Application of high-den-
sity optical microwell arrays in a live-cell biosensing sys-
tem. Anal Biochem 278, 132–142.
58 Kuang Y, Biran I & Walt DR (2004) Simultaneously
monitoring gene expression kinetics and genetic noise in
single cells by optical well arrays. Anal Chem 76, 6282–
6286.
59 Shapiro HM (2000) Microbial analysis at the single-cell
level: tasks and techniques. J Microbiol Methods 42,
3–16.
60 Biran I & Walt DR (2002) Optical imaging fiber-based
single live cell arrays: a high-density cell assay platform.
Anal Chem 74, 3046–3054.
61 Dragoy Whitaker R & Walt DR (2007) Fiber-based sin-
gle cell analysis of reporter gene expression in yeast
two-hybrid systems. Anal Biochem 360, 63–74.
62 Biran I, Rissin DM, Ron EZ & Walt DR (2003) Optical
imaging fiber-based live bacterial cell array biosensor.
Anal Biochem 315, 106–113.
63 DiCesare C, Biran I & Walt DR (2005) Individual cell
migration analysis using fiber-optic bundles. Anal Bio-
anal Chem 382, 37–43.
64 Kuang Y & Walt DR (2005) Monitoring ‘promiscuous’
drug effects on single cells of multiple cell types. Anal
Biochem 345, 320–325.
65 Lu HP, Xun L & Xie XS (1998) Single-molecule enzy-
matic dynamics. Science 282, 1877–1882.
66 Smiley RD & Hammes GG (2006) Single molecule
studies of enzyme mechanisms. Chem Rev 106, 3080–
3094.
67 Rondelez Y, Tresset G, Tabata KV, Arata H, Fujita H,
Takeuchi S & Noji H (2005) Microfabricated arrays of
femtoliter chambers allow single molecule enzymology.
Nat Biotechnol 23, 361–365.
68 Rissin DM & Walt DR (2006) Digital concentration
readout of single enzyme molecules using femtoliter
arrays and Poisson statistics. Nano Lett 6, 520–523.
69 Gorris HH, Rissin DM & Walt DR (2007) Stochastic
inhibitor release and binding from single enzyme mole-
cules. Proc Natl Acad Sci USA , in press.
70 Rissin DM & Walt DR (2006) Digital readout of target
binding with attomole detection limits via enzyme
amplification in femtoliter arrays. J Am Chem Soc 128,
6286–6287.
71 Barnard SM & Walt DR (1991) Chemical sensors based
on controlled-release polymer systems. Science 251,
927–929.
Optical-fiber bundles H. H. Gorris et al.
5470 FEBS Journal 274 (2007) 5462–5470 ª 2007 The Authors Journal compilation ª 2007 FEBS
. reported the use of optical-fiber bundles
for sensing arrays [71], they have been employed for a
variety of applications. Optical-fiber bundles are read-
ily. with optical-fiber
bundles can be used for a variety of high-throughput
screening, gene expression, and biosensor applications.
H. H. Gorris et al. Optical-fiber