Heavy Metals in the Environment - Chapter 19 (end) pptx

19
Bacterial Metal-Responsive Elements
and Their Use in Biosensors for
Monitoring of Heavy Metals
Ibolya Bontidean and Elisabeth Cso
¨
regi
Lund University, Lund, Sweden
Philippe Corbisier
Institute for Reference Materials and Measurements,
Geel, Belgium
Jonathan R. Lloyd and Nigel L. Brown
The University of Birmingham, Edgbaston, Birmingham, United Kingdom
1. INTRODUCTION
Society is learning to adapt to pollution by heavy metals in the environment, and
is now attempting to remediate, control, and minimize such pollution wherever
possible. To do this, there is a need for methods of assessing the amount of
heavy metal pollution in the natural and industrial environments. Although it is
relatively straightforward to use the techniques of analytical chemistry to detect
total amounts of heavy metal in a given location, this rarely tells you how much
Copyright © 2002 Marcel Dekker, Inc.
of this metal is a biological hazard. To achieve this, biological methods may
offer distinct advantages over chemical methods.
It is really only since the industrial revolution that large numbers of peo-
ple have been exposed to significant levels of toxic metals, although at least
since Roman times heavy metals have been used in medicine and cosmetics (1).
In contrast, microorganisms have always lived with ‘‘pollution’’ by heavy
metals, as they have evolved to occupy ecological niches in which these toxic
metals naturally occur and in which they may be released by geochemical
processes. Consequently, bacteria in particular, but also yeasts, fungi, and
many plants, have developed specific mechanisms to tolerate or detoxify heavy

metals. This chapter describes some of the ways in which we and others have
begun to exploit these biological mechanisms to determine the amount of ‘‘bio-
available’’ heavy metal in natural and industrial environments. These methods
are still in their infancy compared with the techniques of analytical chemistry,
but may offer some advantages in ease of use as well as in biological relevance.
In particular, we are attempting to couple the high specificity of biological sys-
tems with the high sensitivity of modern microelectronics in the development of
biosensors.
Elsewhere in this book, you will find information on the occurrence of
heavy metals in the natural environment, and we will not repeat it here. However,
knowledge of the occurrence and amount of heavy metal ions is important in
many fields, such as environmental monitoring, clinical toxicology, wastewater
treatment, and industrial process monitoring. Therefore, many spectroscopic
methods, including atomic absorption and emission spectroscopy (2), flame
atomic absorption spectrometry (3), and inductively coupled plasma mass spec-
troscopy (2,4), have been developed and are commercially available. These meth-
ods exhibit good sensitivity, selectivity, reliability, and accuracy, but they often
require sophisticated instrumentation and trained personnel. Electrochemical
methods like ion selective electrodes, polarography, and other voltammetric
methods (5) are much simpler and require less complex instrumentation, but are
often unable to monitor at very low concentrations. None of these techniques
can define or quantify the amount of heavy metal that is bioavailable and therefore
likely to be a risk to living organisms. To achieve that, one needs a measurement
that is biologically relevant, and the development of biosensors offers consider-
able promise in this respect.
A biosensor is a combination of a highly selective biological recognition
element, responsible for the selectivity of the device, and a detection system (the
transducer) for quantifying the reaction between the biological component and
the target substance (analyte) to be monitored. In this chapter, we describe the
background of bacterial interactions with heavy metals, and illustrate how that

information is being used in the development of biosensors for heavy metals.
Copyright © 2002 Marcel Dekker, Inc.
2. BACTERIAL RESISTANCES TO HEAVY METALS
Heavy metals interact with living organisms in a variety of ways. A number of
metals (e.g., Cu, Fe, Zn, V, Ni) are essential components of metalloenzymes (6).
Others (e.g., Hg, Pb, Cd) are highly toxic with no known beneficial function.
Metal-binding proteins are synthesized by many cell types in response to the
presence of specific metals (7). Both prokaryotic and eukaryotic cells have mech-
anisms to transport essential metals to the sites of synthesis of metalloproteins,
and many bacterial cells have specific systems for conferring resistance to heavy
metals. These transport or resistance systems may be inducible by the metal, and
therefore gene regulatory systems may be required that recognize the metal (8).
The best understood of these systems at present are those responsible for confer-
ring resistance to heavy metals in bacterial systems.
A list of some of the determinants of resistance to heavy metals found in
bacteria is given in Table 1. These include resistances to cations and to oxyanions
of metals in their most common physiological forms, and many of these resistance
determinants have been described in recent reviews (9–12). These resistance de-
terminants confer specificity to one or a few related metal ions, unlike most eukar-
yotic systems, where resistance is due to sequestration by relatively broad-range
determinants, such as metallothioneins or phytochelatins. Metallothioneins have
rarely been identified in bacterial systems (13,14). For all the resistance determi-
nants in Table 1, the genes have been sequenced and the identities of the proteins
conferring resistance have been predicted.
The mechanisms of metal tolerance and resistance vary, but the majority
are due to efflux of the toxic metal from the cell (15). Some of these efflux
systems are part of the normal metal homeostasis systems of the bacterial cell
and the efflux pumps are encoded on the bacterial chromosome. These can be
considered as proteins that confer the normal metal tolerance of the bacterial cells
in which they occur. Examples of such proteins are the ZntA zinc transporter

(16) and the CopA copper transporter (17) in Escherichia coli, or the CopA and
CopB copper transporters in Enterococcus hirae (18). These proteins and some
of the metal resistance proteins [e.g., CadA from Staphylococcus aureus, which
confers Cd(II) resistance (19), or PbrA from Ralstonia metallidurans, which is
part of the lead resistance determinant (20)] have similar structures. They are
P-type ATPases, with eight transmembrane helices, one of which contains the
amino acid sequence Cys-Pro-(Cys/His/Ser), and are known as CPx-ATPases
(21). The N-terminus of about 100 amino acids shows some sequence similarity
to the periplasmic mercury-resistance protein, MerP (see below), and contains a
Cys-X-X-Cys motif associated with heavy metal binding (where X is any amino
acid) (21). This N-terminal MerP-like region may be repeated and was thought
to confer metal specificity on the transporter.
Copyright © 2002 Marcel Dekker, Inc.
T
ABLE
1 Some Heavy Metal Resistance Determinants in Bacteria
Metal Organism Mechanism Location Ref.
Hg Ps. aeruginosa (and large number Uptake of Hg
II
and reduction to Plasmids and transposons 25
of other gram-negative and Hg
0
by mercuric reductase
gram-positive genera)
Cd Staph. aureus Efflux CPx-ATPase Chromosome 19
Ralstonia sp. Efflux pumps (czc, Cd, Zn, and Co; Plasmid 70
cnr, Cd and Ni)
Ps. aeruginosa CMG103 Efflux pumps (czr, Cd, Zn) Plasmid 71
Zn Ralstonia sp. See Cd, czc system Plasmid 70
E. coli Efflux CPx-ATPase Chromosome 16

Ps. aeruginosa CMG103 See Cd, czr system Plasmid 71
Cu Ps. syringae Surface sequestration (cop sys- Plasmid 72
tem)
E. coli ? Surface sequestration/efflux Plasmid 73
(pco system);
efflux CPx-ATPase (copA) Chromosome 17
E. hirae Efflux CPx-ATPase (copA/B) Chromosome 74
Ralstonia sp. Efflux CPx-ATPase Plasmid
Co Ralstonia sp. Efflux pumps (czc Cd, Zn, and Co) Plasmid 70
Synechocystis Efflux pump (coaT) Chromosome 75
Ni Ralstonia sp. Efflux pump (cnr Cd and Ni) Plasmid 76
Pb Ralstonia sp. Possible efflux and sequestration Plasmid 20
As Staphylococcus aureus Arsenate reductase and arsenite ef- Plasmid 77
flux
Cr Ralstonia sp. Efflux Plasmid 78
Ps. aeruginosa Efflux Plasmid 79
Bacillus sp. Efflux ? 80
Ag Salmonella Sequestration and efflux Plasmid 81
Copyright © 2002 Marcel Dekker, Inc.
Expression of the metal homeostasis proteins is usually regulated as part
of the mechanisms whereby the bacterial cell adjusts the intracellular concentra-
tion of the individual metal. ZntA and CopA are regulated by activator proteins
(ZntR and CueR), which respond to the specific metal (22; 22a), and the E. hirae
system involves a complex interaction of the regulatory proteins CopY and CopZ
(23).
Resistance determinants per se are frequently plasmid-borne and may inter-
act with the chromosomally encoded systems for metal homeostasis if the metal
is also an essential nutrient. For example, the copper resistance determinant of
E. coli appears to require proteins that are part of the normal homeostasis system
of E. coli (24) and are therefore encoded on the chromosome. Determinants of

mercuric ion resistance, on the other hand, appear to require only those genes
carried on the mercury-resistance plasmid (25).
The proteins required for metal homeostasis or metal resistance are often
expressed by the bacteria in response to metal ion concentration (26). For ex-
ample, many resistance determinants are expressed only in the presence of the
specific metal ions at high subtoxic concentration (8). This involves specific
regulatory proteins, either repressors or activators, that bind the metal ion and
alter transcription of the structural genes responsible for metal sequestration,
transport, or modification. Metal-resistance determinants and the chromosomal
determinants of metal homeostasis contain metal-responsive genetic elements
responsible for expression of structural gene products that bind and/or transport
the metal ions. These regulatory elements, the regulatory proteins, or the products
of the structural genes could be used in the construction of metal-specific biosen-
sors.
Probably the best understood of all metal resistances is the widespread
group of mercury-resistance (mer) determinants (25,27). Mercury is not an es-
sential nutrient and the resistance determinants are often found in plasmids or
transposons. Among the simplest of these mer determinants is that of transposon
Tn501 from Pseudomonas aeruginosa. This is shown in Figure 1. Three structural
genes, encoding (a) a small periplasmic protein, MerP, (b) an inner membrane
transport protein, MerT, and (c) the enzyme mercuric reductase, are expressed
under the regulation of the activator protein, MerR, which binds Hg(II) and
activates gene expression (25). Possibly because of our detailed knowledge
of this system, several different components of the mer system have been used
in the design of biosensors. These include the NADPH-dependent mercuric
reductase in an enzyme-linked biosensor (28), the mer regulatory region in a
whole cell biosensor (29), and the MerR protein in a capacitance biosensor
(30).
We believe that, as a general principle, we can use many of the bacterial
resistance determinants for other metals in the development of biosensors. Some

examples of the creation of such biosensors are given below.
Copyright © 2002 Marcel Dekker, Inc.
F
IGURE
1 Diagram of the mercuric ion resistance (mer) operon of transposon
Tn501 showing the genes, gene products, and regulatory sequences. The
regulatory region, the mercuric reductase enzyme, and the MerR protein
have all been used in the development of biosensors for Hg(II).
3. OTHER METAL-REGULATED SYSTEMS
A number of heavy-metal-regulated bacterial systems not directly related to
heavy-metal-resistance mechanisms have been described. Those systems are
mainly involved in the intracellular regulation of essential transition metals ions
such as iron, nickel, molybdenum, and magnesium.
The transport and regulation of iron concentration in bacteria has been stud-
ied in detail. This essential metal for cellular metabolism is needed as a cofactor
for a large number of enzymes, but is not easily available to microorganisms in
aerobic environments. Therefore, most aerobic bacteria produce and secrete low-
molecular-weight compounds termed siderophores to capture Fe
3ϩ
from the extra-
cellular medium. The iron uptake has to be very well regulated to maintain the
intracellular concentration of the metal between desirable limits, since too high
an intracellular concentration of iron can catalyze Fenton reactions and generate
toxic species of oxygen. An understanding of how bacteria regulate iron transport
(31) through the Fur protein (for ferric uptake regulation) was gained by mapping,
cloning, and eventually sequencing the fur gene (32). The Fur protein has been
purified (33), and recently the abundance of the Fur protein, the form of interac-
tion with target DNA sequences, and the involvement of Fur in many cell func-
Copyright © 2002 Marcel Dekker, Inc.
tions indicate that the Fur protein performs more like a general regulator than a

specific repressor (34). The cooperative binding of the Fur protein in extended
promoter regions would explain how a relatively simple protein controls a com-
plex regulon in a gradual fashion.
The type of regulation described for Fur appears to be very similar to that
of other metal-dependent repressors. Zinc is also an essential element that, de-
pending on the concentration, becomes a potent toxin. In addition to the regula-
tion of zinc efflux by ZntR (22), the regulation of zinc uptake by the Zur protein
has been described in E. coli (35). The genes involved were named znuACB (for
zinc uptake) and localized at 42 min on the genetic map of E. coli.AznuA-lacZ
operon fusion was repressed by 5 µM zinc and showed a more than 20-fold
increase in β-galactosidase activity when zinc was bound to a zinc chelator; this
was under the control of the zur (zinc uptake regulator) gene. High-affinity
65
Zn
transport of the constitutive zur mutant was 10-fold higher than that of the unin-
duced parental strain. An in vivo titration assay suggested that Zur binds to the
bidirectional promoter region of znuA and znuCB. The Zur protein showed 27%
sequence identity with the iron regulator Fur and is very similar to the Bacillus
subtilis (36) and Listeria monocytogenes (37) homologs.
The Zur and Fur proteins have significant sequence identities (24% in B.
subtilis and 27% in E. coli), and Zur-binding sequences have been described for
promoters of genes related to zinc uptake that are similar to the Fur box. More-
over, the fact that Fur has recently been defined as a zinc metalloprotein con-
taining one structural ion of zinc per polypeptide (38) makes the relation between
these proteins even more complex.
Another example, in addition to Fur and Zur, is SirR—a novel iron-depen-
dent repressor in Staphylococcus epidermidis with homology to the DtxR family
of metal-dependent repressor proteins (39). SirR functions as a divalent metal-
cation-dependent transcriptional repressor and is widespread among the staphylo-
cocci. In B. subtilis the PerR regulon (40) has been shown to respond to iron as

well as the genes involved in the response to oxidative stress such as katA (encod-
ing catalase A) and aphC (alkyl hydroperoxide reductase). However, the Per
boxes are associated with oxidative stress genes in several gram-positive bacteria
rather than with iron transport. The Fur, Zur, SirR, and PerR are all proteins that
can be used as potential biological components of biosensor devices.
Other nonspecific metal-regulated genes related to global stress responses
have also been described and used as biological components of biosensor devices
(41). Heat shock gene expression is induced by a variety of environmental
stresses, including the presence of metal ions. Escherichia coli heat shock pro-
moters for dnaK and grpE were fused to the lux genes of Vibrio fischeri, and it
has been suggested that biosensors constructed in this manner have potential for
environmental monitoring (42).
Copyright © 2002 Marcel Dekker, Inc.
4. BIOSENSORS FOR HEAVY METALS
Biosensors are often cheap analytical devices in which a simple biological event
is transduced into an electronic signal in a quantitative fashion, and ideally these
should show high sensitivity and high specificity, and should work robustly in
complex matrices, such as soil, water, and biological material. The selectivity of
a biosensor depends on the biological component, and its sensitivity depends on
the response of that component and the ease with which this can be transduced
into a measurable signal. A large variety of biological components and transduc-
ers that can be used for heavy metal sensing are summarized in Table 2, showing
their main analytical characteristics, such as limit of detection (LOD), dynamic
range (DR), and selectivity. As can be seen, these characteristics are highly de-
pendent on the type of biological molecule and the transducer used for biosensor
design and construction. The various biosensors also display different stability,
and those based on immobilized enzymes are characterized by a low operating
period.
Recently we have been involved in the development of two different types
of biosensor (43). In one, bacterial cells are genetically modified to respond to

the presence of a heavy metal by the emission of light (44). These whole-cell
biosensors (or in vivo biosensors) are now commercially available. The other
biosensor uses immobilized bacterial proteins that bind heavy metals and alter
the surface properties of an electrode in response to metal binding (30). Such
capacitance electrodes show high sensitivity and some selectivity, but are at an
early stage of development.
Some publications use ‘‘biosensor’’ in the context of detection of toxic
compounds by viability assays, of varying types, on whole cells. We eschew such
a definition, and use ‘‘biosensor’’ in the context of detection of a specific analyte
or a small range of chemically related compounds.
4.1 Whole-Cell Biosensors
A significant area of research in bacterial molecular genetics has been the study
of the control of gene expression (8,26). As part of these studies, many ‘‘reporter
systems’’ have been developed that allow a transcriptional regulatory element to
be placed such that it regulates expression of a gene that has a quantifiable prod-
uct. Two of the most commonly used systems are the lacZ gene of E. coli and
the lux genes of V. fischeri (for example, see ref. 45). The former encodes β-
galactosidase, the production of which can be determined in a simple enzyme
assay using the chromogenic substrate o-nitrophenol-β-d-galactose (46). This is
of some use in the laboratory, but of little use in making a biosensor. The lux
genes of V. fischeri are much more useful in the construction of biosensors, as
they produce and oxidize long-chain aldehydes and generate photons as part of
the reaction (47). The light that is emitted can be measured.
Copyright © 2002 Marcel Dekker, Inc.
T
ABLE
2 Heavy Metal Biosensors and Their Properties
Transducer Operating
Biological molecule type conditions M
2ϩ

LOD DR Ref
Whole cell Mosses Sphagnum sp. Stripping differ- Acetate pH 6.0, Pb
2ϩ
2 ng/ml 5–125 ng/ml 82
ential pulse IS 0.7, 10%
voltammetry moss, carbon
paste elec-
trodes
E. coli ϩ mer pro- Optical detection Bioluminescence Hg
2ϩ
0.1 µM20nM–4µM83
moter ϩ lux is measured at
genes from V. fi- 28°C, 30 min
scheri response time
under aeration
E. coli ϩ lux genes Optical detection Hg
2ϩ
0.1 µM84
from V. fischeri Cu
2ϩ
0.1µM
R. silverii ϩ lux op- Optical detection 23°C, 0.2% ace- Cu
2ϩ
2 µM2–40µM 43,50,
eron from V. fi- tate, 20 mM Zn
2ϩ
5 µM 5–250 µM 55,85
scheri MOPS, pH 7.0, Cd
2ϩ
5 µM 5–200 µM

20 µg/ml tet- Cr
6ϩ
1 µM1µM–40 µM
racycline Pb
2ϩ
1 µM1µM–40 µM
Tl
ϩ
Ni
2ϩ
Bacteria R. silverii ϩ lux op- Optical detection Microorganisms Cu
2ϩ
1 µM1 86
eron from V. fi- immobilized in
scheri polymer matri-
ces, 25°C
E. coli ϩ lux operon Optical detection 30°C, M9 Hg
2ϩ
10 nM 52
medium Cu
2ϩ
1 µM
E. coli ϩ firefly lucif- Optical detection Luminescence is Hg
2ϩ
0.1 fM 0.1 fM–0.1 µM87
erase gene measured in
microtiter
plate after 60
min at 30°C
Copyright © 2002 Marcel Dekker, Inc.

T
ABLE
2 Continued
Transducer Operating
Biological molecule type conditions M
2ϩ
LOD DR Ref
Staph. aureus ϩ Optical detection Luminescence is AsO
43
1 µM1–5µM49
firefly luciferase measured in Cd
2ϩ
1 µM1–20µM
gene scintillation
counter after
60 min
Staphy. aureus ϩ Optical detection Luminescence is Cd
2ϩ
10 nM–1 µM88
firefly luciferase measured in Pb
2ϩ
33 nM–330 µM
gene microtiter Hg
2ϩ
33–100 nM
plates, 30°C
B. subtilis ϩ firefly Optical detection Luminescence is Cd
2ϩ
3.3 nM–1 µm88
luciferase measured in Zn

2ϩ
1–33 µM
microtiter
plates, 30°C
Enzyme Urease ISFET Inhibition of ure- Cu
2ϩ
1–10 mg/L 89
ase immobi- Hg
2ϩ
0.25–5 mg/L
lized on differ- Cd
2ϩ
3–10 mg/L
ent membranes, Pb
2ϩ
2–10 mg/L
0.02 M HEPES,
25°C, batch
mode
ISFET Inhibition of ure- Hg
2ϩ
1 µM90
ase immobi- Cu
2ϩ
3 µM
lized in a Nafion
film, 20°C
Ammonia Inhibition of ure- Cu
2ϩ
0.25 ppm 0.4–0.7 ppm 91

sensor ase, cuvette test Hg
2ϩ
0.07 ppm 0.07–1 ppm
with ammonia- Zn
2ϩ
50 ppm 50–70 ppm
sensitive coat- Pb
2ϩ
100 ppm 100–350 ppm
ing on the wall,
0.1 N maleate
buffer pH 6
Copyright © 2002 Marcel Dekker, Inc.
Ammonia Enzyme reactor Hg
2ϩ
0–15 nM 92
sensor with urease in-
hibited by mer-
cury, enzyme
immobilized
on glass
beads
pH sensor Inhibition of ure- Cu
2ϩ
93
ase immobi- Hg
2ϩ
2 ppb
lized with thy-
mol blue

covalently
bound to
aminopropyl
glass at the tip of
an optical fiber
Conductometric Enzyme on inter- Hg
2ϩ
Cu
2ϩ
1–50 µM94
detection digitated gold Cd
2ϩ
2–100 µM
electrodes, re- P b
2ϩ
5–200 µM
sidual activity of Co
2ϩ
0.02–5 mM
urease is mea- 10–500 µM
sured, 5 mM
Tris-HNO
3
pH
7.4, 50 mM urea
Conductometric Inhibition of ure- Hg
2ϩ
20 ppb 95
detection ase is moni-
tored with a

standing
acoustic wave
device
Fluorimetric de- Flow system, en- Hg
2ϩ
0.5–100 ng/ml 96
tection at 340/ zyme immobi-
485 nm lized on con-
trolled pore
glass, 0.005 M
phosphate
buffer pH 6.5
Copyright © 2002 Marcel Dekker, Inc.
T
ABLE
2 Continued
Transducer Operating
Biological molecule type conditions M
2ϩ
LOD DR Ref
Fluorescence de- Flow system, in- Hg
2ϩ
2 ppb 97
tection at 340/ hibition of ure-
455 nm ase detected
using o-phtha-
ladehyde
Urease IrTMOS Ammonia detec- Hg
2ϩ
0.005 µM98

tion by
IrTMOS, 0.05
M Tris-HCl, pH
8.3
Carbonic anhydrase Fluorescence an- Enzyme labeled Cu
2ϩ
pM 99
isotropy de- with deriva- Co
2ϩ
pM
tection tives of benzo- Zn
2ϩ
pM
xadiazole sul-
fonamide
L
-Lactate dehydro- Amperometric Enzyme coim- Hg
2ϩ
1 µM 100
genase detection mobilized with Cu
2ϩ
10 µM
L
-lactate oxi- Zn
2ϩ
25 µM
dase on the
top of an oxy-
gen electrode
Glycerophosphate Oxygen elec- Inactivation of Hg

2ϩ
µmolar 20–500 µM 101
oxidase trode enzyme by
metal ions, en-
zyme immobi-
lized by reticu-
lation in
gelatin film or
covalent bind-
ing on a mem-
brane
Copyright © 2002 Marcel Dekker, Inc.
Pyruvate oxidase Oxygen elec- Hg
2ϩ
10 nM 101
trode
Cholinesterase Voltammetric de- Flow system, en- Pb
2ϩ
5 µM 102
tection zyme immobi- Cu
2ϩ
50 nM
lized on nitro- Cd
2ϩ
5 µM
cellulose film
with glutaral-
dehyde
Alkaline phospha- Spectrophoto- Chemilumines- Zn
2ϩ

0.17 ppm 103
tase metric de- cence from en-
tection zyme-cata-
lyzed
hydrolysis of a
phosphate
derivative of
1,2-dioxetane
is measured
Horseradish peroxi- Spectrophoto- Inhibition of en- Hg
2ϩ
0.1 pptr 4 orders of 104
dase metric de- zyme immobi- magnitude
tection lized on solid
supports is
measured
Invertase Amperometric Inhibition of en- Hg
2ϩ
1 ng/ml 105
detection zyme immobi-
lized on a
membrane is
measured
Acetylcho- Amperometric Inhibition of en- Cu
2ϩ
0.01 pM 106
linesterase detection zyme by metal Cd
2ϩ
1pM
ions is mea- Fe

2ϩ
10 pM
sured Mn
2ϩ
100 pM
Apoenzyme Alkaline phospha- Calorimetric de- Enzyme immobi- Zn
2ϩ
0.01–1.0 mM 107
tase tection lized on epox- Co
2ϩ
0.04–1.0 mM
ide acrylic
beads, 100
mM TRIS-HCl
pH 8.0
Copyright © 2002 Marcel Dekker, Inc.
T
ABLE
2 Continued
Transducer Operating
Biological molecule type conditions M
2ϩ
LOD DR Ref
Spectrophoto- Flow injection Zn
2ϩ
sub-µ 0.1–10 µM 108, 109
metric de- system, Co
2ϩ
1–200 µM
tection change in ab-

sorbance at
405 nm is mea-
sured
Potentiometric Flowthrough IS- Zn
2ϩ
0.01–1.0 mM 110
detection FET, pH shift
detected
Optical detection Chemilumines- Zn
2ϩ
0.5 ppb 0.5– 50 ppb 103
cence from en-
zyme-cata-
lyzed
hydrolysis of a
phosphate
derivative of
1,2-dioxetane
is measured
Ascorbate oxidase Calorimetric de- Flow system, en- Cu
2ϩ
1–50 µM 111
tection zyme immobi-
lized on po-
rous glass
beads
Spectrophoto- Absorbance at Cu
2ϩ
0.1–10 µM 112
metric de- 265 nm is mea-

tection sured
Amperometric Polarographic Cu
2ϩ
0.5–2 µM 113
detection oxygen elec-
trode is used
Copyright © 2002 Marcel Dekker, Inc.
Carbonic anhydrase Calorimetric de- Flow system, en- Zn
2ϩ
25–250 µM 114, 115
tection zyme immobi- Co
2ϩ
50–200 µM
lized on po-
rous glass
beads
Optical detection Recognition of Zn
2ϩ
40–1000 nM 99,116
at 326/460 and metal ion by
560 nm apoenzyme
transduced by
the dansylam-
ide fluorescent
probe
Galactose oxidase Calorimetric de- Cu
2ϩ
5–20 mM 117
tection
Amperometric Detection with Cu

2ϩ
0.1–10 mM 113
detection oxygen elec-
trode
Alkaline phospha- Amperometric Enzymes coim- Cu
2ϩ
2–100 µM 118
tase ϩ ascorbate detection mobilized on a Zn
2ϩ
2–200 µM
oxidase polymer
membrane
attached to a
polarographic
oxygen elec-
trode
Tyrosinase Amperometric Flow system Up to 0.05 119
detection with oxygen mM
electrode
Protein Apophytochelatin UV- 270 mM apophy- Cd
2ϩ
1–6 ppm 120
spectrophoto- tochelatin is
metric used
detection at
215 nm
Copyright © 2002 Marcel Dekker, Inc.
T
ABLE
2 Continued

Transducer Operating
Biological molecule type conditions M
2ϩ
LOD DR Ref
MerR-LacZα:M15 Spectrophoto- Microtiter plates Hg
2ϩ
ppb level 121
complex metric de- coated with
tection BSA-divinyls-
ulfone- glu-
tathion were
treated with
Hg
2ϩ
concen-
trations and
after washing
the protein
was bound to
it
Glutathione UV- 160 mM glutathi- Cd
2ϩ
1–8 ppm 120
spectrophoto- one is used
metric
detection at
215 nm
pH mea- Protein cross- Cd
2ϩ
10–80 ppm 120

surement linked with glu-
taraldehyde
and en-
trapped be-
hind a dialysis
membrane
Antibody Antibody against Spectrophoto- Microtiter plates Cd
2ϩ
7 ppb 10–2000 ppb 122
Cd-EDTA com- metric de- were coated
plex tection with Cd-EDTA-
BSA conju-
gate and then
the antibody
was added,
HEPES buffer
pH 7.0–7.2
Copyright © 2002 Marcel Dekker, Inc.
The principle of whole-cell biosensors is simple (43,44). The biological
component is a viable bacterial cell that has been modified to contain, say, the
lux genes under the control of a metal-responsive promoter, together with the
regulatory proteins required to express that promoter in the presence of metal.
A stylized system is shown in Figure 2. The lux ‘‘reporter gene’’ is only ex-
pressed, and therefore, the lux gene products are only produced, in the presence
of metal. Therefore, a calibrated system can detect the presence of metal by the
emission of light. As virtually all metal homeostasis genes and all metal resistance
F
IGURE
2 Diagram of the lux transposon in whole-cell biosensors for heavy
metals. Luciferase is expressed from the gene fusion with a metal-regulated

promoter (Pr) under the control of the regulatory gene (Re).
Copyright © 2002 Marcel Dekker, Inc.
proteins appear to be expressed only in the presence of the cognate metal, a wide
variety of metals could be detected by variants of this simple system.
A patent has been published on an early biosensor constructed in such a
fashion (29,48) in which the mercury resistance operon of Serratia marcescens
was coupled to the lux genes of Vibrio to detect in quantitative fashion the pres-
ence of mercuric salts in aqueous solution. However, as pointed out by Rouch
et al. (45), the regulatory region of mercury resistance determinants is not a good
system for quantitative estimates of Hg(II), as it responds in a hypersensitive
manner across a very narrow concentration range. Other systems, such as the
copper resistance (pco) determinant of E. coli, respond across a wide range of
metal ion concentration (45) and may be better suited to the development of
quantitative biosensors. Corbisier and co-workers (49) have linked the regulatory
regions of the cadmium (cad) and arsenate (ars) resistance determinants of Staph-
ylococcus aureus to the lux genes of Vibrio harveyi in a shuttle vector that allows
expression in Staph. aureus and E. coli.
More recent developments exploit the unusual properties of Ralstonia met-
allidurans (formerly Alcaligenes eutrophus) CH34. This organism was isolated
from the soil downwind of a zinc-smelting plant in Belgium. This area is so
heavily polluted with heavy metals from the smelting plant that little grows. Rals-
tonia metallidurans CH34 contains two large plasmids, which between them en-
code a large number of metal-resistance determinants conferring resistance to
copper, cadmium, cobalt, copper, lead, mercury, thallium, and zinc, among others
(50). A transposable element has been constructed that places the lux genes of
V. harveyi under the control of adjacent promoters outside the transposon. This
allows the selection by genetic techniques of different strains in which light is
emitted in response to different metals. A panel of such strains is now commer-
cially available under the name BIOMET.
The use of these bacterial sensors to detect the biologically available metal

fraction in polluted environmental samples has also been demonstrated (51). A
bacterial copper sensor, AE1239 (based on Ralstonia silverii DS185), and a zinc,
cadmium sensor, AE1433 (based on R. metallidurans CH34), have been used to
assess the quality of incinerator fly ashes (52) contaminated by heavy metals and
high concentrations of inorganic salts. The analysis of this type of sample was
made possible only because the sensor was inserted into a soil bacterium able
to grow even in the presence of a high concentration of salts. Those sensors have
also been very efficient for quick evaluation of the efficacy of bioremediation
techniques such as in situ metal inactivation in contaminated soils (53,54). Van-
gronsveld et al. have combined the use of sequential extraction procedures, the
BIOMET microbial heavy metal biosensors, phytotoxicity tests, and a zootoxicity
test as a test system for the evaluation and monitoring of the efficacy and durabil-
ity of in situ immobilization of metals in contaminated soils (53). Good agreement
was found between the different evaluation criteria, and the BIOMET sensors
Copyright © 2002 Marcel Dekker, Inc.
were recommended by the main competent authority, the Public Waste Agency
of Flanders (OVAM), as a relevant tool to detect bioavailable heavy metals in
polluted soils.
A R. metallidurans nickel sensor, AE2515, was also used to quantify nickel
bioavailability in various nickel-enriched soils that had been treated with addi-
tives for in situ metal immobilization (55). The nickel bioavailability measured
with the biosensor correlated linearly with data on the biological accumulation
of nickel in specific parts of important agricultural crops. Therefore, the biosensor
could be used to assess risk associated with the potential transfer of nickel to
organisms in the food chain, in this case maize and potato plants grown on nickel-
enriched soils (55).
The analysis of lead-polluted soils with the bacterial sensor responsive to
lead ions is given below as an example. The lead bacterial biosensor was based
on a genetically engineered R. metallidurans, AE2450. This bacterium bears a
plasmid with a pbrRpbrAϻluxCDABE fusion (43), and a bioluminescent signal

is produced when the bacteria are in contact with lead ions, as illustrated in
Figure 2. Those ions activate the lead-sensitive promoter within the cell, which
controls the production of luciferase enzyme responsible for the bioluminescence
signal. The bioluminescence was recorded with a luminometer and was propor-
tional to the biologically available fraction of lead present in the sample.
4.2 Enzyme-Based Biosensors
Enzyme-based biosensors are by far the most common class of biosensor yet
constructed and constitute the majority of biosensors in commercial use. These
are commonly used for detecting metabolites in clinical chemistry, owing to the
specificity and variety of enzymes available that act on biochemical compounds
(56). However, heavy metals participate in relatively fewer specific biochemical
reactions and any such reaction may be relatively nonspecific. Mercuric reductase
is a member of the dithiol oxidoreductase class of proteins and its enzymology
is well understood (57–59). The enzyme is a flavoprotein that catalyzes the two-
electron reduction of Hg(II) to Hg(0) in an NADPH-dependent manner. The
likely physiological substrate is the dimercaptan rather than the divalent cation
(Equation 1).
RS-Hg
II
-SR′ ϩ NADPH ϩ H
ϩ
→ Hg
0
ϩ NADP
ϩ
ϩ RSH ϩ R′SH (1)
As was pointed out some years ago (60), this reaction can be followed
using an enzyme electrode to follow the consumption of H
ϩ
. Alternatively, the

NADPH can be regenerated in a linked second reaction, which can be followed.
Recently a biosensor has been patented (28) in which the NADPH regeneration
is linked to the production of a long-chain aldehyde, such as octanal or decanal,
which are substrates for luciferase. As the aldehyde is produced it is oxidized
Copyright © 2002 Marcel Dekker, Inc.
by luciferase to the corresponding long-chain acid with the concomitant produc-
tion of light, which is detected. This biosensor has not yet been described in the
peer-reviewed literature, but the patent also covers the use of organomercury
lyase to convert organomercurials to Hg(II), thus allowing the detection of orga-
nomercurials, and the patent extends to cover other metal ions that may be re-
duced in biochemical reactions. Hg, Cr, As, Tc, Cu, Ag, Se, V, Mo, and U are
specifically mentioned, although suitable enzymes for reduction of all of these
have yet to be purified.
For sake of completeness, we should mention that apoenzymes are some-
time used to detect the presence of the metal ion cofactor normally found in the
corresponding holoenzyme (61). The metal ion can be detected in quantitative
fashion owing to gain of enzyme activity on binding the metal ion.
4.3 Capacitance Biosensors
Recently, a new heavy metal biosensor was designed, based on heavy-metal-
binding proteins, as the biorecognition element, and a highly sensitive capacitive
transducer (30,62–64). The characteristics of two classes of metal-binding pro-
teins used for biosensor design are presented below.
The biosensor uses a capacitive transduction principle that can be briefly
outlined as follows. There is a double layer between a metal electrode (gold) and
the solution, resulting in a double-layer capacitance as well as in a Faradaic cur-
rent giving rise to a Faradaic background current in electrochemical measure-
F
IGURE
3 Principle of the capacitance biosensor, showing the basic construc-
tion and the proposed effect of metal binding.

Copyright © 2002 Marcel Dekker, Inc.
F
IGURE
4 (A) Schematic diagram of the experimental setup of the capaci-
tance biosensor. (B) Current transient obtained after the potentiostatic pulse
is applied.
Copyright © 2002 Marcel Dekker, Inc.
ments. The electrode surface can be isolated by, for example, covering it with
long-chain alkenethiols that readily self-assemble to materials like gold. A capaci-
tance still can be measured although it is not a double-layer capacitance in the
usual sense. The capacitance change is measured by applying a fast potentiostatic
pulse to the electrode and evaluating the resulting current transients, which are
altered by any changes that affect the metal-binding proteins.
Thus, monitoring of heavy metals with this type of biosensor is based on
the conformational change that occurs when the metal ions bind to the protein.
The conformational change results in change of the capacitance. A schematic
representation of the detection principle is shown in Figure 3.
4.3.1 Biosensor Preparation
Biosensors were prepared according to a previously published protocol (30). The
biosensor was inserted as the working electrode in a specially constructed three-
(four)-electrode flow cell with a dead volume of 10 µl, as shown in Figure 4A.
The electrodes were connected to a fast potentiostat described elsewhere (62).
A platinum foil and a platinum wire served as the auxiliary and the reference
electrodes, respectively. An extra reference electrode (Ag/AgCl) was placed in
the outlet stream. The buffer solution was pumped by a peristaltic pump and
samples were injected into the carrier buffer flow. The working electrode had
initially a rest potential of 0 mV versus the Ag/AgCl reference electrode. Mea-
surements were made by applying a potential pulse of 50 mV and current tran-
sients were recorded following the applied potential step according to Equation
2 (see Fig. 4B).

i
(t)
ϭ u/R
s
exp(Ϫt/R
s
C
1
) (2)
where i
(t)
is the current at time t, u is the amplitude of the potential pulse applied,
R
s
is the resistance between the gold and the reference electrodes, C
l
is the total
capacitance over the immobilized layer, and t is the time elapsed after the poten-
tial pulse was applied.
5. PRACTICAL EXAMPLES
5.1 Whole-Cell Biosensors
A bacterial sensor for lead was used to investigate levels of bioavailable metals
in soil. Fifteen soil samples from the Haarlem, Arnhem, and Rotterdam munici-
palities were analyzed with the R. metallidurans AE2450 lead sensor. Results
obtained with this Pb sensor and by chemical analysis are summarized in Table 3.
Five grams of each soil sample was mixed in a reconstitution medium (0.2%
gluconate, 20 mM MOPS, pH 7.0, 20 µg.ml
Ϫ1
tetracycline). β-Glycerophosphate
was used in the medium to avoid lead precipitation. An aliquot of 2 ml was

Copyright © 2002 Marcel Dekker, Inc.
T
ABLE
3 Characterization of the Soil Samples by the BIOMET Lead Sensor and by Chemical Analysis
Pb in
Pb in carbonate
Sample Sample Pb equivalent Total content exchangeable fraction
TNO ID municipal ID Origin (mg /kg dw) (mg/kg dw)
a
(mg/kg dw)
a
(mg/kg dw)
a
822523-001 1.3 Haarlem 2.29 Ϯ 1.27 640 0 127
822523-002 2.4 Haarlem 2.21 Ϯ 0.51 300 0.3 99
822523-003 3.4 Haarlem 1.89 Ϯ 0.51 1500 0 321
822523-004 4.1 Haarlem 1.64 Ϯ 0.15 370 0 78
822523-005 5.2 Haarlem 2.88 Ϯ 0.59 470 0 88
822526-001 003A Arnhem 1.53 Ϯ 0.29 240 0 50
822526-002 006A Arnhem 1.69 Ϯ 0.51 400 0 99
822526-003 010A Arnhem 3.50 Ϯ 1.76 460 0 126
822526-004 011A Arnhem 2.98 Ϯ 0.00 980 0 342
822526-005 014A Arnhem 2.05 Ϯ 0.73 700 0.7 160
822526-006 017A Arnhem 3.29 Ϯ 0.73 860 0 236
822526-007 020A Arnhem 1.69 Ϯ 1.03 3200 0 25
822526-008 021A Arnhem 6.69 Ϯ 0.62 4400 0 576
422278-001 Almstraat Rotterdam n.d. 1100 0 667
422278-003 Schiekade Rotterdam 13.4 Ϯ 4.57 2700 0 1420
422278-004 Strekkade Rotterdam 6.34 Ϯ 1.13 1020 0 137
n.d., not determined because of lack of soil material.

a
The chemical analyses were performed by Tauw Milieu b.v.
Copyright © 2002 Marcel Dekker, Inc.
further diluted 1:2 and 1:4, and 20 µl of the soil solution was mixed with 180
µl of the reconstituted bacterial sensor (at a concentration of 10
7
–10
8
CFU/ml).
Duplicate soil/solution/bacteria samples were incubated in a Lucy 1 luminometer
at 23°C and bioluminescent measurements were automatically performed every
30 min for 8 h. A calibration curve was generated by adding 20 µl of Pb(NO
3
)
2
solutions to the 180-µl bacterial solution. The following concentrations were
tested: 0, 0.6, 1, 2.5, 5, 10, 15, and 25 µM lead (equivalent to 0, 0.124, 0.207,
0.518, 1.04, 2.07, 3.11, and 5.18 ppm lead). The equation derived from the cali-
bration curve fitting was used to determine the concentration of lead equivalent
in the soil sample and is presented in Figure 5. The soil dry weight (dw) was
determined separately by drying the soil samples at 105°C for 18 h.
The detection limit of the lead sensor was around 0.85 mg lead/kg dw.
Most of the values obtained with the soil samples were low but not negligible,
except sample 021A from Arnhem and the two soil samples from Rotterdam.
There was no direct relationship between the total content of lead of the soils,
or the CaCl
2
-extractable fraction, and the quantity of lead measured with the
bacterial sensor (Fig. 6). Almost no lead could be detected in the CaCl
2

fraction
by conventional analytical methods. This CaCl
2
fraction is sometimes considered
the ‘‘bioavailable’’ fraction for living organism and, from our data, it is greatly
underestimated. However, the bioavailable concentration measured with our bac-
terial sensor was directly related to the quantity of lead found in the sodium-
acetate-extractable fraction (Fig. 7). This fraction corresponds to the fraction of
F
IGURE
5 Calibration curve of the lead AE2450 Pb-BIOMET sensor in the pres-
ence of increasing concentration of lead nitrate after 6.5 h contact. The aver-
age bioluminescence of two measurements was plotted as the ratio of the
bioluminescence observed in the presence of increasing lead concentrations
to the bioluminescent signal obtained in the presence of MilliQ-treated water.
Copyright © 2002 Marcel Dekker, Inc.
F
IGURE
6 Analysis of the soil with the AE2450 Pb-BIOMET sensor plotted
against the total amount of Pb after CaCl
2
extraction procedure. R
2
, coefficient
of linear regression.
metal bound to CaCO
3
. In other words, this means that the lead present in those
soils was available for the soil bacteria and able to induce a bioluminescence
signal. The soil samples from Haarlem and Arnhem (except number 021A) have

very little available lead. Sample 021A from Arnhem and both samples from
Rotterdam had a higher available lead concentration, the highest being the soil
F
IGURE
7 Relationship between the bioavailable Pb fraction measured with
the AE2450 Pb-BIOMET sensor and the amount of Pb after Na acetate, pH 5
extraction procedure. R
2
, coefficient of linear regression.
Copyright © 2002 Marcel Dekker, Inc.