Biosensors and Bioelectronics 20 (2004) 345–349

Detection of trophic factor activated signaling molecules

in cells by a compact fiber-optic sensor
Rakesh Kapoor a,∗ , Navjot Kaur a , Emmanuel T. Nishanth a ,
Stanley W. Halvorsen b , Earl J. Bergey a , Paras N. Prasad a
a Institute for Lasers, Photonics and Biophotonics, University at Buffalo, State University of New York, 458 NSC, Buffalo, NY 14260, USA
b Department of Pharmacology and Toxicology, University at Buffalo, State University of New York, Buffalo, NY 14214, USA

Received 11 December 2003; received in revised form 2 February 2004; accepted 4 February 2004

Available online 16 March 2004

Abstract

This paper describes a highly sensitive method to detect trophic factor activated signaling molecules in cells using a compact fiber optic
biosensor. The method is demonstrated by quantitative detection of phosphorylation of signal transducers and activators of transcription 3
(STAT3) in neuroblastoma cells. A single fiber-optic probe based on total internal reflection fluorescence sensing system is used. A 405 nm
diode laser is used for evanescent wave excitation of immobilized labelled analyte on the probe surface. A compact charged coupled device
(CCD) based spectrometer is used for recording the fluorescence signal. The method is two orders of magnitude more sensitive than the
Western blotting technique.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Fiber-optic; TIRF; STAT; Biosensor; Fluorescence; Evanescent wave; Diode laser

1. Introduction molecules attached to an analyte (Hirschfeld and Block,

In recent years considerable research effort has been de-
voted to the development of fiber-optic biosensors because
of their potential sensitivity, detection speed and adaptability
to a wide variety of assay conditions (Mehrvar et al., 2000;
Prasad, 2003). Fiber-optic biosensors are analytical tools in
which a fiber-optic device serves as a transduction element.
The usual aim is to produce a signal that is proportional to
the concentration of a chemical or biochemical to which the
biological element reacts. The bioactive compound could be
an enzyme, an antibody, or a nucleic acid. The signal gen-
erated in such a sensor is the result of a specific interaction
between the analyte and the immobilized bioactive com-
pound. The signal measured by the fiber-optic system, upon
analyte binding, can originate from surface changes in the
absorbance, scattering or fluorescence.
The total internal reflection fluorescence (TIRF), or flu-
orescence using evanescent-wave excitation is especially
well suited for measuring the concentration of fluorescent
∗ Corresponding author. Tel.: +1-716-645-6800x2221;
fax: +1-716-645-6945.
E-mail address: rkapoor@acsu.buffalo.edu (R. Kapoor).
1984; Andrade et al., 1985; Shirver-Lake et al., 1995; Zhou
et al., 1997). In this technique an analyte has to be selec-
tively immobilized on the surface of the fiber probe. Several
strategies have been used for immobilization on fiber-optic
biosensors. The sandwich assay technique is one such strat-
egy (Wyatt et al., 1992) in which one antibody is immobi-
lized on the surface of the waveguide, and a second antibody
labelled with a fluorescent dye is added to the bulk solution.
In the absence of the antigen, the second antibody remains
in solution and little fluorescence is observed. However
upon addition of antigen, a “molecular sandwich” is formed
on the waveguide and the labelled second antibody is within
the evanescent-wave excitation volume. Evanescent-wave
excites the labelled antibody and produces the characteristic
fluorescence of the dye molecule attached to the antibody.
The fluorescence signal from these labelled antibodies is
proportional to the number of bound antigens. In this paper
we have demonstrated that, based on sandwich assay binding
and TIRF detection, a single fiber-optic probe can be used
for detection of trophic factor activated signaling molecules
in cells. Phosphorylation of specific amino acids is criti-
cal for the regulation of protein activity involved in many
cellular processes. The method has been demonstrated by

0956-5663/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.bios.2004.02.004
346 R. Kapoor et al. / Biosensors and Bioelectronics 20 (2004) 345–349
detecting the fraction of phosphorylated signal transducers
and activators of transcription 3 (STAT3) in neuroblastoma
cells. This quantitative method is rapid and sensitive. Cil-
iary neurotrophic factor (CNTF) induces phosphorylation
of STAT3 on tyrosine residues. After stimulation, activated
STATs dimerize and translocate to the nucleus and interact
with specific DNA response elements. This regulates the
expression of targeted genes (Darnell, 1997; Stahl et al.,
1994), required for proper growth, maintenance and devel-
opment of cells and tissues (Leonard and Oshea, 1998). In
addition, STAT3 plays an important role in regulating both
innate and acquired immunity. The constitutive activation
of STAT3 is associated with cancer (Bromberg et al., 1999).
2. Materials
All solvents and chemical were either of analytical grade
or chemically pure. Aminopropyltriethoxysilane (APTS)
was obtained from Sigma (Milwaukee, WI, USA). The im-
mobilization reagent, N-hydroxysuccinimidyl-4-azidosali-
cylic acid (NHS-ASA) was obtained from Pierce (Rockford,
IL, USA). Monoclonal anti-STAT3 antibody was obtained
from Transduction Laboratories (Lexington, KY, USA)
and polyclonal anti-phospho STAT3 (Tyr705) antibody was
obtained from Cell Signaling Tech. (Beverly, MA, USA).
Alexa fluor 430 dye was obtained from Molecular Probes
(Eugene, OR, USA) and Centrispin-10 columns were from
Princeton Separations (Adelphia, NJ, USA).
2.1. Cell culture and whole cell extraction
BE (2)-C human neuroblastoma cells were grown in a
1:1 mixture of Ham’s F12 and Eagle’s minimal essential
medium supplemented with 10% (v/v) fetal calf serum,
50 U/ml penicillin and 50 ␮g/ml streptomycin (Kaur et al.,
2002). Before treatment, cells were placed in serum free
media for 2 h. Whole cell extracts were prepared either be-
fore or after CNTF stimulation (1 nM, 30 min). Cells were
rinsed with phosphate buffered saline (PBS) thrice and ex-
tracted in lysis buffer (20 mM HEPES (pH 7.9), 1.5 mM
MgCl2 , 0.2 mM EDTA, 420 mM NaCl, 0.2 mM PMSF,
1 mM DTT, 100 ␮M Na3 VO4 , and 20% glycerol) by keep-
ing on ice for 30 min (Bhattacharya and Schindler, 2003)
then freeze-thawed to insure complete cell lysis. Lysates
were centrifuged at 14 000 × g at 4 ◦ C and supernatants
were collected.
2.2. Antibody-dye conjugation
Anti-phospho-STAT3 (PY-STAT3) antibody (1.1 ␮g/ml)
and anti-STAT3 (1.2 ␮g/ml) antibodies were conjugated
with optimum concentration of Alexa Flour 430 dye. The
reaction mixture was incubated for 2 h at room temperature
with constant shaking. The labelled-antibody conjugates
were separated from free dye using gel filtration columns
(Centrispin-10).
2.3. Western blotting
Cell extract proteins were separated on 7.5% sodium
dodecyl sulphate polyacrylamide gels before transfer to
polyvinylidene difluoride membrane (ImmobilonP, Milli-
pore). Proteins on immunoblots were probed with anti-
phospho-STAT3 antibodies as described previously (Kaur
et al., 2002) using Amersham enhanced chemiluminescence-
Plus (ECL-Plus). Exposed films were scanned with an
Epson 636 Professional Series scanner.
3. Probe preparation
The probe was a 10 cm long 600 ␮m core multimode op-
tical fiber (Ocean Optics Inc.). Protective polymer (2.5 cm)
surrounding the cladding of the fiber was removed from one
end by heating. The fiber was then decontaminated by soni-
cating it in a soap solution (DeContam from ESPI). This was
followed by sonicating the fiber in a solution of de-ionized
water to get rid of any carbon soot on the surface of the
fiber. The probe was then transferred to a solution of 10%
hydrofluoric acid for the appropriate duration of time to etch
the cladding. To immobilize antibodies tips were cleaned
and neutralized in ethanol, and then the etched part of the
fiber was silanized for 3 h with a 2% solution of APTS. Af-
ter silanization, about 1 cm of the tip of the etched part of
fiber was exposed to 400 ␮l of 0.5 mM NHS-ASA in the
dark for 1 h. Then the fiber was washed and rinsed with PBS
(pH 7.2) to remove any unbound NHS-ASA from the fiber.
This procedure minimized any non-specific binding of for-
eign molecules to the surface of the fiber, which would result
in background fluorescence. The treated fiber tips were then
placed in the anti-STAT3 antibody solution at a concentra-
tion of 2.5 ␮g/ml for about 10–15 min and exposed to long
wave UV light in order to initiate the photo activation of the
cross-linker. This leads to the immobilization of anti-STAT3
antibody onto the probe surface. Anti-STAT3 has the same
binding affinity for both the non-activated STAT3 protein
and the activated STAT3 (PY-STAT3) protein.
4. Instrumentation
A typical diagram of our setup is shown in Fig. 1. A
405 nm laser diode (Nichia, Japan) is used for excitation. A
dichroic band pass filter (405 nm, band width 10 nm) was
placed in front of the diode laser to block any red tail emis-
sion. The diode light passes through a dichroic beam com-
biner/splitter. The beam combiner/splitter is a long pass filter
with high reflectivity for the excitation wavelength at 405 nm
and high transmission at wavelengths longer than 480 nm.
We used Alexa 430 as a fluorescence label with emission
peak at 530 nm. Laser output is a collimated beam and a
 R. Kapoor et al. / Biosensors and Bioelectronics 20 (2004) 345–349 347

Fig. 1. A schematic diagram of the experimental setup: ND, neutral

density; BS, beam splitter; BF, bandpass filter; and LD, laser diode.

short focal length lens is used to focus the laser beam into
a 600 ␮m core fiber. The fiber-probe was connected to this
fiber with the help of a detachable SMA connector (Thorlabs
Inc., NJ, USA). This arrangement helped us in changing the
probe fiber for different experiments. The analyte fluores-
cence couples back into the probe fiber and gets transmitted
into the collection fiber of a miniature charged coupled de-
vice (CCD) based fiber-optic spectrometer (Ocean Optics,
model HR2000). The signal from the spectrometer is cou-
pled to a computer (IBM). All the spectra were collected
with the help of this computer. The auto fluorescence gen-
erated in the transporting fibers can drastically reduce the
signal to background ratio, therefore the choice of fiber is an
important task. After testing several fibers it was found that a
600 ␮m core fiber model ZDF VIS/NIR from Ocean Optics
Inc. gave negligible auto fluorescence at the 405 nm excita-
tion wavelength. For better coupling efficiency, the core di-
ameter of the collection fiber was also chosen to be 600 ␮m.
5. Results and discussions
Fig. 2. Signals from two probes prepared in identical conditions. Both the
probes were placed for an hour in Alexa 430 labelled anti-STAT3 antibody
solution. Sharp peaks are mercury lines from leaked stray room light.
signal obtained from the extract of non-activated cells was
much smaller than that obtained from the extract of acti-
vated cells. This confirms the ability of our technique to
discriminate between the activated and non-activated cells.
To quantitate the phosphorylation of STAT3 protein, the
ratio of concentration of PY-STAT3 to that of total STAT3
(activated plus non-activated) was determined. A fiber probe
with immobilized anti-STAT3 antibody on its surface was
placed in protein extracts from activated cells. To saturate
the probe we placed it in the solution for 3 h. Under saturat-
ing conditions the ratio of PY-STAT3 molecules to STAT3
molecules, attached to the immobilized antibodies on the
probe surface, will correspond to their ratio in cell extract
solution. After rinsing with PBS-Tween-20 the probe was
kept in Alexa 430 labelled anti-PY-STAT3 solution for an

First, we compared the reproducibility of our probes.

Two probes were prepared under identical conditions and
were placed in for 3 h in whole cell protein extracts from
stimulated cells. The probes were then rinsed with PBS
containing 0.05% Tween-20 and placed for 1 h in a solution
of Alexa 430 labelled antibodies, specific to phosphorylated
STAT3 (anti-PY-STAT3). The fluorescence signals recorded
from these probes are shown in Fig. 2. It can be seen that the
signals obtained from the both the probes are nearly iden-
tical. Thus results obtained from the probes prepared under
identical conditions were reproducible. Next the signals ob-
tained from activated and non-activated cell extracts were
compared using two probes prepared under identical con-
ditions. The first probe was placed in extract solution from
cells activated with 1 nM CNTF and the second in extract
solution from non-activated cells, in each case for 3 h. After
Fig. 3. Fluorescence spectrum from two probes. The higher amplitude
rinsing with PBS-Tween-20, both the probes were placed
signal was obtained from the probe kept in extract solution from activated
in Alexa 430 labelled anti-PY-STAT3 solution for 1 h. The cells while the lower amplitude signal is from the probe kept in extract
spectra recorded from the probes are shown in Fig. 3. The solution from non-activated cells.
348 R. Kapoor et al. / Biosensors and Bioelectronics 20 (2004) 345–349
hour. It was found that 1 h was sufficient to saturate the
signal. After rinsing the probe with PBS-Tween-20, the
fluorescence spectrum was recorded. The amplitude of this
spectrum is proportional to the total PY-STAT3 concentra-
tion. Next the same probe is placed in Alexa 430 labelled
second antibody anti-STAT3 solution for an hour. After
rinsing the probe with PBS-Tween-20, the fluorescence
spectrum was recorded again. The amplitude of this spectra
is proportional to total concentration of STAT3 proten (acti-
vated plus non-activated). Since the same optical probe was
used, the spectral recording conditions for both the signals
were identical. Under identical conditions, if SPY is the sig-
nal corresponding to the PY-STAT3 protein and ST is the
signal corresponding to the total STAT3 protein (activated
plus non-activated), the ratio of the concentrations will be
given as
CPY SPY
= , (1)
CT [SPY + (DA /DN )(ST − SPY )]
where DA and DN are dye/antibody ratio for labelled
anti-PY-STAT3and anti-STAT3, respectively. The dye/
antibody ratio of the labelled-antibody solutions was de-
termined by measuring the absorbance at 430 and 280 nm
using the following formula:
[dye] 203 000 × [A430 ]
= , (2)
[antibody] 16 000 × [(A280 − (0.28 × A430 )]
where the A430 and A280 represent the absorbance at
wavelengths of 430 and 280 nm, respectively. The molar
excitation coefficients of typical antibody and Alexa 430
are 203 000 and 16 000 cm−1 M−1 , respectively and 0.28 is
the correction factor to account for absorption of the dye
at 280 nm. In our experiment the ratio DA /DN was 1.05.
Eq. (2) is valid if the signal corresponding to PY-STAT3 and
STAT3 are recorded either with two identical fiber-probes
or with same probe. Both the spectra recorded with our
Fig. 4. Fluorescence signal corresponding to the concentrations of
PY-STAT3 and total STAT3 (activated + non-activated) in the extract
solution from activated cells.
Table 1
Measurement of PY-STAT3 fraction determined in different dilutions of
whole cell protein extract (the cells were activated with 1nM concentration
of CNTF)
Relative dilution PY-STAT3 factor (%)
1 38
1:5 35
1:25 37
1:625 33
Average (36 ± 2)
Fig. 5. Western blotting technique: (A) unstimulated cells; (B) undiluted
stimulated cell extract; (C) 1:5 times diluted stimulated cell extract;
(D) 1:25 times diluted stimulated cell extract; (E) 1:625 times diluted
stimulated cell extract.
probe are shown in Fig. 4. The signal showing higher am-
plitude in Fig. 4 is proportional to the total STAT3 protein
(activated plus non-activated) concentration. To estimate
the fraction of PY-STAT3, all spectra were corrected for
background, and the signal amplitude was computed from
the area under the corrected spectral curve. The fraction of
PY-STAT3 protein in the cell extract was obtained by using
Eq. (1). The PY-STAT3 fraction in extract solution ob-
tained from cells activated with 1 nM of CNTF, was found
to be 38%. To further test the sensitivity of our method,
the experiment was repeated to determine the PY-STAT3
fraction in three different dilutions (1:5, 1:25, and 1:625) of
the extract solution obtained from cells activated with 1 nM
of CNTF. We could detect signals for all the dilutions and
the estimated fraction of activated STAT3 was unchanged
with increasing dilution of cell extract solution (Table 1).
The experiment was also repeated with 1:3125 dilution but
we could not detect any signal above the background. For
comparison purposes we determined PY-STAT3 signal for
different dilutions of the cell extract using Western blotting
(Fig. 5). We could detect signal corresponding only to the
1:5 dilution, but no signal was detected for 1:25 and 1:625
dilutions. This demonstrates that a single fiber optic probe
can be used to measure tyrosine phosphorylation of STAT3
in neuroblastoma cells, with sensitivity orders of magnitude
better than that provided by by a state-of-the-art Western
blotting method.
6. Conclusion
A rapid and sensitive quantitative detection of trophic
factor activated signaling molecules in cells has been devel-
oped using a TIRF based fiber-optic method. The method
is demonstrated by determining the fraction of tyrosine
 R. Kapoor et al. / Biosensors and Bioelectronics 20 (2004) 345–349
phosphorylation of STAT3 protein induced in neuroblas-
toma cells. The complete assay takes only 2–4 h with the
fiber-optic probe method, a significant improvement over
the 2–3 days required by Western blotting technique. It
is also demonstrated that the proposed fiber optic method
is two orders of magnitude more sensitive than Western
blotting. This technique can also be adapted for the quan-
titative detection of activation of other signaling proteins
such as ERK1/2 (extracellular signal regulated protein
kinase 1/2), p38/MAPK (p38/mitogen-activated protein ki-
nase), JNK/SAPK1 (Jun N-terminal kinase/stress-activated
protein kinase) and AKT/PI3K (AKT/phosphatidylinositol
3-kinase), using highly specific antibodies to activated
proteins that are commercially available.
Acknowledgements
We thank Regeneron Pharmaceuticals (Terrytown, NY)
for recombinant human CNTF. This work was supported
by Universal Technology Corporation through contract
#02-S470-020-C1, Infotonics through contract #Q550000
and Enhanced Center for Advanced Technology through
subcontract #55415-00-01A.
References
Andrade, J.D., VanWagenen, R.A., Gregonis, D.E., Newby, K., Lin, J.N.,
1985. Remote fiber-optic biosensors based on evanescent-excited fluo-
349
roimmunoassay: concept and progress. IEEE Trans. Electron Devices
ED-32 (7), 1175–1179.
Bhattacharya, S., Schindler, C., 2003. Regulation of Stat3 nuclear export.
J. Clin. Invest. 111 (4), 553–559.
Bromberg, J.F., Wrzeszezynsta, M.H., Devgan, G., Zhoo, Y., Pestell, R.G.,
Albanese, C., Darnel, J.E., 1999. Stat3 as an oncogene. Cell 98, 295–
303.
Darnell, J.E., 1997. STATs and gene regulation. Science 277, 1630–
1635.
Hirschfeld, T.E., Block, M.J., 1984. Fluorescent immunoassay employing
optical fiber in capillary tube, US Patent No. 4447546.
Kaur, N., Wohlhueter, A.L., Halvorsen, S.W., 2002. Activation and inac-
tivation of signal transducers and activators of transcription by ciliary
neurotrophic factor in neuroblastoma cells. Cell. Signaling 14, 419–
429.
Leonard, W.J., Oshea, J.J., 1998. Jaks and STATs-biological implications.
Ann. Rev. Immunol. 16, 293–322.
Mehrvar, M., Bis, C., Scharer, J.M., Young, M.M., Luong, J.H., 2000.
Fiber-optic biosensors—trends and advances. Anal. Sci. 16, 677–692.
Prasad, P.N., 2003. Introduction to Biophotonics. Wiley, New York,
pp. 327–334.
Shriver-Lake, L.C., Golden, J.P., Patonay, G., Narayanan, N., Ligler, F.S.,
1995. Use of three longer-wavelength fluorophores with the fiber-optic
biosensor. Sens. Actuators B 29, 25–30.
Stahl, N., Boulton, T.G., Farruggella, T., Ip, N.Y., Davis, S., Witthuhn,
B.A., Quelle, F.W., Silvennoinen, O., Barbieri, G., Pellegrini, S., Ihle,
J.N., Yancopoulos, G.D., 1994. Association and activation of Jak-Tyk
kinases by CNTF-LIF-OSM-IL-6 beta receptor components. Science
263 (5143), 92–95.
Wyatt, G.M., Lee, H.A., Morgan, M.R.A., 1992. Immunoassays for
Food-Poisoning Bacteria and Bacterial Toxins. Chapman & Hall, Lon-
don, pp. 29–67.
Zhou, C., Pivarnik, P., Auger, S., Rand, A., Letcher, S., 1997. A compact
fiber-optic immunosensor for Salmonella based on evanescent wave
excitation. Sens. Actuators B 42, 169–175.