Journal of Analytical Toxicology, Vol. 27, September2003
[ TechnicalNote
Use of Biosensorsto Screen Urine Samplesfor
Potentially Toxic Chemicals
Jacqui Horswell* and Stuart J. Dickson
ESR,Kenepuru Science Centre, P.O. Box 50 348, Porirua, New Zealand
[Abstract
]
Forensic toxicology laboratories are often required to implicate or
exclude poisoning as a factor in a death or unexplained illness. An
analytical tool which enables toxicologists to screen a wide variety
of common poisons would be extremely useful. In this paper, we
describe the use of a bacterial biosensor for detecting the presence
of commonly encountered potentially toxic chemicals in urine. The
biosensor responds to any chemical that causes metabolic stress to
the bacterial cell and the response is in direct proportion to the
concentration of the stressor. This allows a measure of the
concentration of a toxicant in urine, without knowing exactly what
the toxic compound(s) may be. This affords a distinct advantage
over conventional analytical techniques, which require an
extensive screening program before it is even known that a toxic
compound is present. This preliminary investigation has shown that
this biosensor can indicate the presence, in urine, of herbicides
such as glyphosate, 2,4-dichlorophenoxyacetic acid, and 2,4,5trichlorophenoxyacetic acid; the biocide pentachlorophenol; or
inorganic poisons such as arsenic, mercury, and cyanide. The
biosensor was also shown to be sensitive to a concentration range
of these toxicants likely to be found in samples submitted for
toxicological analysis.
Introduction
Forensic toxicology laboratories are frequently asked to implicate or exclude chemical poisoning as a factor in a death or
unexplained illness. A comprehensive toxicological screen includes a large number of tests for an array of potential poisons
with greatly differing chemical structures. Despite the increasing range and sophistication of modern instrumental techniques, such screens can be very costly, and, consequently, only
limited screening, dictated by individual case circumstances, is
performed.
Biosensors are defined as analytical devices incorporating a biological sensor, such as a microorganism or an enzyme, that
can provide an information linked response to a specific challenge, that is, the presence of a toxic chemical, via a suitable
transducer (e.g., electrochemical, optical) (1). Biosensors have
applications in numerous fields, including medicine testing
9 Author to whom correspondence should be addressed.
372
(2--4), environmental testing, (5--7), and food quality testing
(8,9). Bacterial biosensors are organisms whose stress responses
can be used as tools to detect a specific stress such as that
posed by the presence of a toxic compound. Bacteria are stable,
easily maintained at a low cost, and a large number of cells can
be exposed to a toxin in one study, making them an ideal organism for rapid screening of solutions (7). Bacterial biosensors
have several advantages over classical laboratory techniques.
They are generally simple and cheap to culture and have the
ability to operate in complex matrices (1). Furthermore, the
selection of a specific microorganism can be made to ensure relevance; for example, in soil environments, the choice of a relevant biosensor would ensure representation of the correct ecological niche and accurately reflect the response of the wild
type soil organism to stress. Genetic modification can greatly
enhance the detection of the microbial biosensor response, primarily through the use of reporter genes (e.g., bacterial bioluminescence).
In this preliminary study, we used a bacterial biosensor
genetically modified with genes for bioluminescence. Luminescent bacteria are the most abundant and widespread of the
luminescent organisms found in nature. The genes responsible
for light emission in a few of these organisms have been well
characterized. The genetic system required for luminescence in
the bacterium Photobacterium (Vibiro) fischeri is the lux
operon. The lux operon contains two genes for the luciferase enzyme (A and B). This enzyme is composed of two different
polypeptide chains. The operon also contains several other
genes (C, D, and E) that are thought to code for enzymes (fatty
acid reductase complex) that produce the substrates for the
light-emitting reaction. These substrates are called luciferins
and are long chain fatty aldehydes (10). Light emission results
from the luciferase catalyzing a reaction between oxygen, a reduced flavin mononucleotide, and a long-chain fatty aldehyde.
Most micro-organisms lack the genetic blueprint for luciferase
and the fatty acid reductase, but they can supply the flavin
mononucleotide; therefore, for dark bacteria to become bioluminescent, all that is required is the genetic transfer of the lux
operon. The bacterial strain used in our assay was constructed
by transformation with plasmid pUCD607,which carries the lux
genes (11). In brief, the lux genes are joined to plasmid DNA
(small circular DNA molecules that exist apart from the chromosomes) to form a hybrid, or recombinant molecule that is
Reproduction(photocopying)of editorial contentof this journal is prohibitedwithout publisher'spermission.
Journal of Analytical Toxicology, Vol. 27, September 2003
able to replicate in bacteria. In order to prepare a recombinant
molecule, the plasmid and gene of interest are cut at precise positions by specific deoxyribonucleases (restriction endonucleases) and then the molecules are spliced together to form a new
plasmid (pUCD607). The lux genes are placed behind a constitutive promoter in the plasmid, which means they are expressed
continuously. After the hybrid plasmid molecule has been prepared, it is introduced into bacterial cells by transformation. The
hybrid plasmid replicates in the dividing bacterial cells to produce an enormous number of copies of the original gene. The
construction of the plasmid is such that under normal conditions, the bacteria will bioluminesce, that is, give out visible
light. Bioluminescence is particularly useful because it is extremely sensitive and is sufficiently rapid to allow real-time
monitoring (10,12). The light output is directly proportional to
the metabolic activity (i.e., health) of the biosensor. Any chemical that causes metabolic stress to the bacterial cell will result
in a decline in luminescence that is in direct proportion to the
concentration of toxic compound present. This gives a measure
of the overall toxicity to the organism of a sample and requires
no pre-knowledge about the specific compound present in the
sample (as is necessary when employing standard chemical
analysis), thus biosensors have the potential to provide a rapid
screening test applicable to forensic toxicology. The/u.x-marked
bacterial biosensors have successfully been used to detect a variety of pollutants in the terrestrial environment [e.g., heavy
metals (13); pesticides (14); BETX (15); organotins (16)] and are
used commercially by an environmental diagnostics company
(RernediosTM) to characterize contaminated land in the United
Kingdom (www.remedios.uk.com).
In this preliminary study, Bscherichia coli HB101 pUCD607
was used to evaluate the sensitivity of this biosensor to a variety
of potentially toxic chemicals (PTCs) commonly encountered in
forensic toxicology. E. coli was chosen because it is an ubiquitous organism and robust enough to operate in complex samples such as urine.
Experimental
Urine screen
Initially, 22 urine samples were collected from volunteers to
determine the range in pH, color, and biosensor response. The
volunteers were random "healthy" volunteers on staff at ESR.
No effort was made to exclude volunteers with concurrent medications. All the urines were collected between 9 a.m. and 12
p.m. on the same day. Urine color was compared by measuring
absorbance. An arbitrary wavelength of 500 nm at the
yellow/green end of the visible spectrum was chosen, and a
Lambda Bio UVNIS Spectrometer (PerkinEImer) was used for
measurements.
Preparation of PTC standards
The PTCs selected for initial testing represented compounds
that are either encountered relatively frequently by toxicologists
and/or can only be detected by individual tailored and time
consuming tests. The PTCs were N-(phosphonomethyl)glycine
(glyphosate) and the commercial formulation of this herbicide
"Roundup| 2,4-dichlorophenoxyacetic acid (2,4-D) (sodium
salt), 2,4,5-trichlorophenoxyacetic acid (2,4,5-T) (sodium salt),
methyl viologen dichloride (paraquat) and the commercial formulation "Gramoxone| gamma hydroxybutyrate (GHB), ethylene glycol, sodium cyanide, sodium arsenate, sodium arsenite,
mercuric chloride, sodium fluoroacetate (1080), phenol, and
pentachlorophenol (PCP) (Table I). Stock solutions were made
up as follows: 1000 mg of each of the chemicals were individually weighed out using a five-point balance, transferred to acidwashed 1000-mL glass volumetric flasks, and made up to the
mark with fresh urine. For 2,4-D and 2,4,5-T, 1000 mg was
added to approximately 800 mL of urine and then concentrated
sodium hydroxide was added drop by drop until the compounds
dissolved, then made up to 1000 mL with urine. Test solutions
were made by diluting the stock solutions to the required concentration with urine. The urine used in all the PTC screens,
was provided by one subject.
Methods
The biosensor consisted of lyophilized cells ofEscherichia coli
HB101 genetically modified with the lux CDABE genes, originally isolated from Vibrio fischeri, using the multicopy plasmid
pUCD607 (7).
The biosensor was resuscitated from freeze-dried vials for i h
in 10 mL 0.1M KC1 and used directly (7). A suspension of cells
(100 pL) was added to 900 pL of test solution and mixed by
pipetting for 5 s at 15-s intervals between each sample. Luminescence was measured in a LumiSkan TL (Labsystems) at 550
nm after a 15-min exposure. All assays were carried out in triplicate using independent vials. Data were expressed as a percentage of the luminescence measured in control urine samples.
Quality assurance of biosensors
The biosensor was kindly supplied by The University of
Table I. Potentially Toxic Chemicals Used in This Study
Potentially
TypicalUrineMeanand
Biosensor
Toxic
RangeConcentrations(mg/L)
Response
Compound
in FatalPoisonings
Reference (ECs0)
Glyphosate
Roundup
2,4-D
2,4,5-T
Paraquat
Gramoxone
GHB
Ethyleneglycol
Cyanide
Arsenate
Arsenite
Mercuric
18
18
19
20
21
21
22
19
23
24
24
19
316.3
93.1
23.7
100.4
> 500
> 500
> 500
> 500
3.9
0.2
6.9
0.4
1080
Phenol
11,200
11,200 (as glyphosate)
314 (111-670)
53
(I-I0)
(I-I0) (as paraquat)
(0-I 513)
5700(600-10,800)
37 (0.4-230)
0.2 (as arsenic)
0.2 (as arsenic)
(0.5-I .6) for
vapor inhalation (as mercury)
368
214 (208-220)
19
19
> 500
5.2
(EC150)
PCP
153 (28-520)
19
0.0077
373
Journal of Analytical Toxicology, Vol. 27, September 2003
Aberdeen. To assure the quality of the response, the biosensor
was exposed to test solutions containing serial dilutions of
TCP TM liquid antiseptic (Pfizer Ltd., Sandwich, U.K.), an
aqueous glycerol solution of halogenated phenols (0.68% w/v)
and phenol (0.175% w/v). The effect of urine on the biosensor
response was determined by comparing the TCP TM ECs0 in both
urine and water.
Data interpretation
The bioluminescence was plotted as a percentage of the response of the biosensor to the unadulterated urine (control)
against PTC concentration. Negative exponential curves were
fitted to the data points by non-linear regression techniques for
most of the PTCs. The decline in bioluminescence was described by the negative exponential curve, Y = a + be ~, where
Y = % bioluminescence, X = PTC concentration, a = lower
asymptote, b = range of Yvalues, K = rate parameter (13). For
phenol, a positive exponential curve was fitted to the data, described by the equation Y = a + b (1 -e-k). Curves were fitted
using Prism 3.0 (San Diego, CA) and ECs0 values were calculated by solving the equation for values of Y = 50% for each of
the PTCs (Y = 150% for phenol).
Results
Urine screen
Urine pH and color varied considerably among the 22 samples
analyzed (data not shown). The pH ranged from 5.5 to 7.3 with
an average of 5.6. The color of urine was also highly variable,
ranging from very dark (absorbance of 0.82) to very pale (absorbance of 0.02). Bioluminesence of the bacterial biosensor
ranged from 0.03 to 1.08 relative light units (RLUs).A multiple
regression analysis was carried out to determine if there was a
relationship between color, pIi, and luminescence. A log trans120 1
6
O
~
~
form was applied to the luminescence variable prior to the
analysis. Results indicated that there was a linear relationship
between these three variables (adjusted R 2 for the model was
0.64, p < 0.0001) and that pH and color could be used to predict
luminescence. It should be noted that the analysis was carried
out on a limited number of samples provided by volunteers. To
improve the prediction of the model a more extensive screen
should be undertaken. Results also indicated that at pH values
below 6, luminesence is low. The urine used in all the PTC
screens was provided by one subject and had a very consistent
pH (6.9 + 0.14), well within the pH range of urines from healthy
individuals [typical range between 4.8 and 7.5 (17)]. This individual's urine was selected for use in the PTC screens because
the subject was available on call to provide urine and because
the urine had a relatively high luminescence thus providing a
greater range for test samples (0.8 RLUs).
Response of biosensor to PTCs
The plotted graphs of bioluminescence as a percentage of
the unadulterated urine (control) against PTC concentration,
provide a rapid method of assessing the effect of the toxin on the
metabolism of the organism. Examples of dose-response curves
obtained are shown for the 2,4-D; 2,4,5-T; mercury (Figure 1);
and phenol (Figure 2). Declining bioluminescence with increasing concentration of PTC provide an indicator of the toxicity of the element on the metabolism of the organism enabling
an EChoto be calculated. The EChovalues for TCP test solutions
in urine and water were very similar, 0.11 and 0.17, respectively
(Figure 3).
Results from toxicity testing of aqueous standards of PTCs
showed the use of lux-marked bacteria provides appropriate
sensitivity for testing of a broad spectrum of PTCs. The ECs0
values for the PTCs are shown in Table I. Exposure of the
biosensor to phenol, resulted in a stimulation in bioluminescence with increasing concentration of the toxic compound
(Figure 2).
For paraquat (pure chemical formulation and commercial
formulation), GHB, 1080, and ethylene glycol, an ECs0was not
reached, even at concentrations up to 500 mg/L.
8060-
Discussion
4020-
A bioluminescent biosensor was shown to be a useful tool for
-@
0
0
50
100
mg/L
150
200
200
9
40
~
20
0
,
0.5
-20
~
e
1
1.5
2
2.5 3
mg/L Hg~
3.5
4
4.5
5
Figure 1. The effect of varying concentrationof 2,4-D; 2,4,5-T; and
mercury, on bioluminescence of E. coil HB101 pUCD607.
374
12o !
IO0
i~
"6=
16of
i 14~
18o t / %
120
80 4
6O4
40
20 4
OI
0
--..e
100 200 300 400 500 600 700 800 900 1000
phenol
Figure2. The effect of varying concentration of phenol on bioluminescence of E. coil HB101 pUCD607.
Journal of Analytical Toxicology, Vol. 27, September 2003
120
100
uu-~
80 9 ~
60-
['-~
in waterl
L--m--TCPin urine J
4020.
0
0
0.1
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
% TCP
=
1
1.1
Figure 3. The effect of varying concentrations of TCP test solutions of
urine and water on bioluminescenceof E. coil HB101 pUCD607.
estimating the toxicity of urine samples. The biosensor was
able to indicate the presence of a toxic substance in aqueous
standards at concentrations likely to be found in toxicological
samples. In some instances (e.g., phenol), presence of a PTC
may cause an increase in bioluminescence. This is still a
biosensor "response" to the presence of a compound and it
may be possible to calculate an effective concentration that
causes a 50% increase in luminescence (EC150). It is probable
that the biosensor is using phenol as a carbon source, hence increasing the organism's metabolism and resulting in an increase in bioluminescence. However, caution must be taken
when interpreting these results as urine may contain utilisable
substrates which may increase the bioluminescence response of
the organism. Indeed, the effect of common over-the-counter
compounds such as ibuprofen, or compounds such as caffeine
and oral contraceptives were not measured in this preliminary
work. Thus, the rapid screening for PTCs in urine must be interpreted as a balance between the effect of toxins and the response to substrates. For some PTCs, paraquat (pure chemical
formulation and commercial formulation), GHB, 1080, and
ethylene glycol, the biosensor was not sensitive to the significant toxicological concentrations in urine (Table I); therefore,
this biosensor is unlikely to be usefuI in detecting these chemicals in forensic toxicological samples.
To date, the E. coli biosensor has primarily been used in environmental testing (6,14,25), and validation of the biosensor
response to PTCs has been carried out in water based media
(e.g., soil solution). Thus, it was necessary to determine if the
biosensor response could also be validated in urine. The QA test
solution (TCP) results showed that urine does not significantly
effect the biosensor response or the ECs0 for this compound.
Thus, the protocol developed for environmental testing can
also be used for the testing of urine samples.
The urine screen suggested that luminescence was significantly reduced in samples with a pH < 6. Luminescence response from the biosensor has been shown to be stable across
the pH range 4.5-7.0 (7). Consequently, one of the limitations
of the technique could be the presence of compounds in urine
[e.g., ammonium (26)] that affect both its color and pH and reduce the light output of the biosensor in a manner similar to a
toxicity response. Results for the bacterial biosensor response
are normally compared to those of a non-contaminated reference sample of similar physical make-up; thus, variations in luminescence caused by pH and color are removed. However, ob-
taining a "control" forensic sample will not usually be possible.
It may, however, be possible to use a "control" urine that has
been matched in both colour and pH to the toxicological
sample, as statistical analysis of the urine screen results suggest
that both of these variables can be used to predict luminescence.
An alternative control may be obtained after processing of the
toxicological sample to remove any compounds present (e.g.,
metals removed by passing the sample through activated
carbon) (27). A second and more promising method used in environmental testing is the "flash assay". In the "flash assay", the
control reading is replaced with the peak luminescence observed immediately after addition of the bacteria into the actual
sample. Luminescence is recorded 20 times/s for the first 30 s
after dispensing and this becomes your reference or control
value. The luminescence of the sample is then measured in
the usual manner after a 15-min exposure time. The color and
pH effects are therefore taken into account throughout the
whole measurement period. The ratio of these two luminescence values is used as a measure of the toxicity of the sample
(28). Constraints on the use of the "flash assay" would be that
light-output of the biosensor is reduced to background by pH
and/or urine color; hence, the biosensor would be unable to detect the presence of toxins over and above this effect. Buffering
of samples may be an option for future optimization of the
assay. However, adjustment of the pH of a sample may not be a
simple matter as it may alter the form and availability of toxins
to the biosensor. In addition, results from the urine screen
suggested that both pH and urine color affected light-output of
the biosensor. More work is needed to optimize the assay and to
determine the best "control". Another limitation of the tool is
that forensic toxicology urine samples are normally preserved
by the addition of sodium fluoride (1-2% w/v); this is added primarily to inhibit microbial growth and changes in concentrations of alcohol and other chemicals in the sample (29,30). We
have shown that the bacterial biosensor is inhibited by the
presence of sodium fluoride, and therefore we were unable to
validate the biosensor using actual forensic toxicological samples. For routine use of the biosensor in screening toxicological
samples a urine sample should be collected without the sodium
fluoride preservative.
Future application of such biosensor technology in forensic
toxicology analysis may lie with the design of more specific
biosensors. The reporter gene (lux) can be fused to genes involved with a particular response. For example, Rod et al. (31)
used a bacterial biosensor with a luciferase gene fused to a
gene encoding resistance to mercury. The highly sensitive mercury biosensor was used to detect urinary mercury from
amalgam fillings. For many herbicides, bacterial resistance
genes have been identified and it may be possible to design a
whole suite of specific bacterial biosensors.
Conclusions
Bacterial biosensors are not as specific, sensitive or accurate
as chemical testing but are relatively cheap, easy to use, reproducible, and rapid (incubation time is 15 min for most chemicals). The applicability of this system is the rapid broad spec375
Journal of Analytical Toxicology,Vol. 27, September2003
trum screening of forensic toxicological samples for the presence or absence of PTCs, as a pointer for further confirmatory
chemical testing.
Acknowledgments
This work was funded by The Institute of Environmental Science and Research Limited and the New Zealand Police under
their joint Capability Development Programme. The help of
Bjorn Sutherland is gratefully acknowledged.We would also like
to thank The University of Aberdeen, U.K., for provision of the
lux-marked biosensor.
References
1. M.J. Dennison and RE Turner. Biosensors for environmental monitoring. Biotech Adv. 13:1-12 (1995).
2. D.G. Georganopoulo, R. Carley, D.A. Jones, and M.G. Boutelle.
Development and comparison of biosensors for in vivo applications. Faraday Discuss. 116:291-303 (2000).
3. T. Vo-Dinh and B. Cullum. Biosensors and biochips: advances in
biological and medical diagnostics. J. Anal. Chem. 366:540-551
(2001).
4. F.W. Scheller, U. Wollenberger, A. Warsinke, and F. Lisdat. Research and development in biosensors. Curr. Opin. Biotechnol. 12:
35-40 (2001).
5. S. Tombelli, M. Mascini, C. Sacco, and A.P.F. Turner. A DNA
piezoelectrical biosensor assay coupled with a polymerase chain
reaction for bacterial toxicity determination in environmental samples. Anal. Chim. Acta. 418:1-9 (2000).
6. G. Strachan, S. Preston, H. Maciel, A.J.R. Porter, and G.h Paton.
Use of bacterial biosensors to interpret the toxicity and mixture toxicity of herbicides in freshwater. War. Res. 35:3490-3495 (2001).
7. G.L Paton, E.A.S. Rattray, C.D. Campbell, M.S. Cresser, L.A. Glover,
J.C.L. Meeussen, and K. Kiltham. Use of genetically modified microbial biosensors for soil ecotoxicity testing. In Biological Indicators of Soil Health and Sustainable Productivity. C. Pankhurst,
B. Doube, and V. Gupta, Eds. CAB1 Publishing, Wallingford, U.K.,
1997, pp 397-418.
8. J. Chen and M.W. Griffiths. Salmonella detection in eggs using lux+
bacteriophages. J. Food Prot. 59:908-914 (1996).
9. M.J. Loessner,C.E.D. Rees, G.S.A.B. Stewart, and S. Scherer. Construction of a luciferase reporter bacteriophage A511: luxAB for
rapid and sensitive detection of viable Listeria cells. Appl. Environ.
Microbiol. 62:1133-1140 (1996).
10. J.I. Prosser, K. Killham, L.A. Glover, and E.A.S. Rattray. Luminescence-based systems for detection of bacteria in the environment.
Crit. Rev. Biotech. 16:157-183 (1996).
11. J.J. Shaw and C.I. Kado. Development of a Vibrio bioluminescence gene-setto monitor phytopathogenic bacteria during the ongoing disease process in a non-disruptive manner. Bio/Technology
4:560-564 (1986).
12. G.I. Paton. Lighting the way for bioremediation. Chem. Ind. Feb:
77-79 (2001).
13. G.I. Paton, G. Palmer, M. Burton, E.A.S. Rattray, S.P.McGrath, L.A.
Glover, and K. Killham. Development of an acute and chronic toxicity assay using lux-marked Rhizobium leguminosarum biovar tri-
376
folii. Lett. Appl. Microbiol. 24:296-300 (1997).
14. G. Strachen, S. Capel, H. Maciel, A.J.R. Porter, and G.I. Paton. Application of cellular and immunological biosensor techniques to assess herbicide toxicity in soils. Eur. J. Soil Sci. 53:1-8 (2002).
15. E.M. Boyd, A.A. Meharg, J. Wright, and K. Killham. Assessment of
toxicological interactions of benzene and its primary degradation
products (catechol and phenol) using a lux-modified bacterial
bioassay. Envrion. Toxicol. Chem. 16:849-856 (1997).
16. J.G. Bundy, J. Wardell, C.D. Campbell, K. Killham, and G.I. Paton.
Application of bioluminescence-based biosensors to the ecotoxicity assessmentof organotins. Lett. AppL Microbiol. 25:353-358
(1997).
17. K. Diem and C. Lentner. Scientific Tables, 7th ed. Ciba-Geigy,
Basle, Switzerland, 1971, p 662.
18. S.J.Dickson, R.H. Meinhold, I.D. Beer, and T.D. Koelmeyer. Rapid
determination of glyphosate in postmortem specimens using 31p
NMR. J. Anal Toxicol. 12:284-286 (1988).
19. R.C. Baselt. Disposition of Toxic Drugs and Chemicals in Man, 5th
ed. Chemical Toxicology Institute, Foster City, CA, 2000, pp
765-766.
20. S.J.Dickson. Environmental Science and Research Limited, Porirua,
New Zealand, personal communication, 2002.
21. J.M. Scherrmann, P. Houze, C. Bismuth, and R. Bourdon. Prognostic value of plasma and urine paraquat concentration. Hum.
ToxicoL 6:91-93 (1987).
22. C.J. Rehling. Poison residues in human tissues. In Progress in
Chemical Toxicology, Vol. 3, A. Stolman, Ed. Academic Press,
New York, NY, 1967, pp 363-386.
23. M.J. Ellenhorn and D.G. Barceloux. Medical Toxicology: Diagnosis
and Treatment of Human Poisoning, Elsevier, Amsterdam, The
Netherlands, 1988, p 1014.
24. W.C. Duet, K.L. Byers, and J.V.Martin. Application of a convenient
extraction procedure to analyze gamma hydroxybutyrate in fatalities involving gamma-hydroxybutyric acid, gamma-butyrolactone, and 1,4-butanediol. J. Anal. Toxicol. 25:576-582 (2001).
25. G.M. Sinclair, G.I. Paton, A.A. Meharg, and K. Killham. Luxbiosensor assessment of pH effects on microbial sorption and toxicity of chlorophenols. FEMSMicrobiol. Letl. 174:273-278 (1999).
26. J.J. Bray, P.A. Cragg, A.D.C. Macknight, R.G. Mills, and
D.W. Taylor. Leclure Notes on Human Physiology, 1st ed. Blackwel[ Scientific Publications, Melbourne, Australia, 1986, pp
502-507.
27. K. Kadirvelu, K. Thamaraiselvi, and C. Namasivayam. Removal of
heavy metals from industrial wastewaters by adsorption onto activated carbon prepared from an agricultural solid waste. Bioresource Technol. 76:63-65 (2001).
28. J. Lappalainen, R. Juvonen, K. Vaajasaari, and M. Karp. A new flash
method for measuring the toxicity of solid and coloured samples.
Chemosphere 38:1069-1083 (1999).
29. P.S.Lough and R. Fehn. Efficacy of 1% sodium fluoride as a preservative in urine samples containing glucose and Candida albicans.
J. Forensic Sci. 38:266-271 (1993).
30. A.W. Jones, L. Hylen, E. Svensson, and A. Helander. Storage of
specimens at 4~ or addition of sodium fluoride (1%) prevents formation of ethanol in urine inoculated with Candida albicans.
J. Anal. Toxicol. 23:333-336 (1999).
31. A. Rod, P. Pasini, M. Mirasoli, M. Guardigli, C. Russo, M. Musiani,
and M. Baraldini. Sensitive determination of urinary mercury (11)by
a bioluminescent transgenic bacterial-based biosensor. Anal. Lett.
34:29~11 (2001).
Manuscript received June 18, 2002;
revision received November 15, 2002.