FIELD PORTABLE IMMUNOSENSORS THAT MEASURE LOW MOLECULAR WEIGHT CONTAMINANTS IN ENVIRONMENTAL WATER SAMPLES Scott J. Melton1, Ruquia B. Ahmed-Schofield2, Robert C. Blake II3, Diane A. Blake4 1 Interdisciplinary Program in Molecular and Cellular Biology

1
FIELD PORTABLE IMMUNOSENSORS THAT MEASURE LOW
MOLECULAR WEIGHT CONTAMINANTS IN ENVIRONMENTAL
WATER SAMPLESScott J. Melton1, Ruquia B.
Ahmed-Schofield2, Robert C. Blake II3, Diane A.
Blake41 Interdisciplinary Program in Molecular
and Cellular Biology and 4Department of
Biochemistry, Tulane Univ. Hlth. Sci. Ctr. New
Orleans LA, 701122Department of Chemistry and
3Division of Basic Pharmaceutical Sciences ,
Xavier University of Louisiana, New Orleans, LA
70125 Tel [phone] email
blake_at_tulane.edu, smelton_at_tulane.edu
Table 1. Results from a caffeine spike and
recovery using an alpha prototype of the field
portable instrument.
Abstract Immunosensors that employed a kinetic
exclusion method have been used to develop a
variety of assays for low molecular weight
contaminants. The prototype sensors and
associated software were provided by Sapidyne
Instruments Inc (Boise, ID). These sensors
produced assays that were 100- to 1000-fold more
sensitive than competitive microwell-based
immunoassays prepared using identical primary
antibodies and analyte-conjugates. Analysis of
operator-prepared samples in the field-portable
instrument required 165 seconds, and CVs were
10-15. The in-line sensor had the ability to
collect a sample from a process line, add
reagents, inject the incubation mixture and
compare the results to an instrument-generated
standard curve. Total analysis time, including
washing and mixing, was between 600 and 900
seconds, and CVs were 5 or less. These sensors
have been used to develop assays for heavy metals
and caffeine in environmental and food samples.
We are also in the process of developing
monoclonal antibodies to brevetoxin selected for
optimal performance in these immunosensors. These
monoclonal antibodies will be used to develop a
field portable assay for brevetoxin that will
provide results in a nearly real-time format.
A
B
Purified Water Caffeine ppb spiked in Amount
Recovered 1nM 0.911nM
91.10 2nM 2.284nM
114.20 6nM 5.578nM
92.96 Average
99.42
50 Environmental Water Caffeine ppb spiked
in Amount Recovered 1nM 0.943(SD .173)
94.30 2nM 1.94 (SD 0.095)
97.00 6nM 4.98 (SD .225)
83.00 Average
91.43
Figure 3. Two kinetic exclusion based instruments
useful for field-based studies. A. An automated
sensor that has the ability to autonomously run a
standard curve from stock reagents and to prepare
and analyze environmental samples (3). B. A beta
prototype field-portable device. Samples are
injected over a flow cell containing microbeads
coated with ligand-conjugate complex. After the
generation of a standard curve, data can be
obtained within minutes. The device interfaces
with a laptop computer through a wireless
connection and weighs less than 12 kilograms.
Both instruments promise a wide range of utility
for the near real-time analysis of environmental
samples

• Conclusions
• These data demonstrate that the format of an
  immunoassay may significantly influence its
  performance characteristics, even when identical
  reagents are employed. Immunosensors based on the
  principal of kinetic exclusion appear to be
  uniquely suited for the analysis of low molecular
  weight contaminants.
• The total time required for sample mixing,
  incubation and analysis is 5-10 minutes.
• Modifying the sensor for analysis of a new
  analyte is as easy as changing the beads in the
  flow cell and changing the antibody in the
  incubation mixture. Kinetic exclusion assays have
  been formulated for caffeine and heavy metals in
  our laboratory (3,4) and for PAHs and TNT in
  aquatic samples at sites near Virginia Institute
  of Marine Science (5). Assays for marine algal
  toxins are in development at Tulane.
• These sensors are expected to fill an important
  niche in analysis of low molecular weight
  environmental contaminants.

Figure 1. Principle Behind the Immunosensor
Assay. A. Fluorescently labeled antibody is mixed
with ligand (L) and allowed to come to
equilibrium. B. The mixture is then passed over
beads containing bound ligand. Excess antibody
that did not bind to ligand in the first step is
captured by the beads. C. As more antibody binds
to the column, fluorescent signal increases. The
resultant slopes generated are inversely
proportional to the amount of ligand present in
the incubation mixture. D. A plot of slope versus
ligand concentration generates a standard curve
that can be used to assess the concentration of
the ligand in an unknown sample (1,2).
Figure 4. Results from field test of the in-line
instrument at Oak Ridge National Lab (ORNL). The
in-line device shown in Panel A above was
transported to ORNL and set up in the
laboratories there. After signal tests to insure
that the sensor was not damaged during transport,
the instrument was loaded with reagents. The
sensor autonomously mixed samples with known
quantities of uranium (VI) (0-21 nM) to generate
a standard curve (?), then mixed two samples from
a bioremediated site with reagents (buffer,
chelator, an antibody that recognized a chelated
form of U(VI) and Cy5-labeled anti-mouse Fab) for
comparison to the instrument-generated standard
curve.

• Please contact us to discuss development of
  additional assays for these biosensors
• The in-line sensor is suitable for deployment in
  a buoy or jetty.
• The field portable sensor is ideal for use in a
  boat or along the shore.
• Both sensors provide near real-time data about
  contaminant levels.
• We are actively trying to identify new marker
  compounds that could assist investigators in the
  study of the processes that lead to hypoxia.
• The ideal marker for these sensors would be a
  compound that has a molecular weight gt400, is
  structurally complex, and is difficult to measure
  with current technologies. It should be present
  in environmental samples at concentrations of at
  least 10 nM.

Kinetic Exclusion
cELISA
A
B
References 1. Blake, D.A. et al. (1996) Metal
binding properties of a monoclonal antibody
directed towards metal-chelate complexes. J.
Biol. Chem. 27127677-27685. 2. Blake, R.C. II
et al. (1999) Automated kinetic exclusion assays
to quantify protein binding interactions in
homogeneous solution. Anal. Biochem. Anal.
Biochem. 272123-134. 3. Yu, H. et al. (2005) An
immunosensor for autonomous in-line detection of
heavy metals Validation for hexavalent uranium.
Int. J. Env. Anal. Chem. 85817. 4. Blake, D.A.
et al. (2001) Antibody-based sensors for heavy
metal ions. Biosens. Bioelectron., 16799-809. 5.
Bromage, E.S. et al. (2007) The development of a
real-time biosensor for the detection of trace
levels of trinitrotoluene in aquatic
environments. Biosens. Bioelectron, in press.
Figure 2. Comparison of a cELISA with a kinetic
exclusion-based assay for caffeine. A.
Competitive ELISA for Caffeine. Because caffeine
concentrations are relatively high (30 nM) in
water contaminated with human waste, caffeine has
been proposed as an anthropogenic marker for
untreated wastewater. A competitive ELISA
developed in our lab using commercially available
reagents (anti-caffeine and a caffeine-BSA
conjugate) provided an assay with a minimum
detection limit of 500 nM. B. Kineitc Exclusion
Assay for Caffeine. This assay used the same
anti-caffeine antibody and caffeine-BSA conjugate
employed in the cELISA. Because of the very short
contact time of the antibody with the
caffeine-BSA coated beads, the antibody bound to
soluble caffeine is does not have time to
dissociate and re-bind to the caffeine-BSA
conjugate. This kinetic exclusion results in
an assay with a limit of detection approximately
4900-fold more sensitive than the cELISA.
Figure 5. Analysis of caffeine in the
field-portable sensor. Soluble caffeine (0-6 nM)
was mixed with buffer, anti-caffeine antibody and
Cy5-labeled anti-mouse Fab. After 10 minutes at
room temperature, the reaction mixture was
applied to a flow cell containing beads coated
with caffeine-BSA. Delta fluorescence (calculated
by subtracting the average the signal from
seconds 5-10 of the trace from the signal at
155-160 seconds of the trace) was plotted versus
caffeine concentration. The figure compares the
dose-response for caffeine spiked into reagent
grade water (red) and into a sample that
contained 50 of a caffeine-free environmental
water sample (green).