Biosensors for marine pollution

Donny Edra Prastya

Telp. 083849821231

Abstrak
Measurement of ecological, climatic and anthropogenic changes underpins the formulation of
effective management strategies for sustainable use and protection of the marine environment.
Sensors are traditionally used in marine studies to determine physical parameters, but there is
increasing demand for real-time information about chemical and biological parameters. These
parameters are currently measured in samples collected at sea and subsequently analysed in the
laboratory. Biosensors fuse the exquisite sensitivity and specificity of living systems with the
processing power of microelectronics to deliver simple, inexpensive measurement systems for use in
the field or deployment in situ. While their potential for use in the marine environment is enormous,
much published work to date has focussed on applications in freshwater and wastewater. Marine
applications pose a substantial challenge in the robustness required for remote application, but recent
developments in portable medical devices and receptor design suggest that these demands can now be
realistically tackled.
Keywords: Marine biosensors; Decentralised measurements; Novel instrumentation; Analytical
techniques; Chemical pollution; Pollution monitoring

1. Introduction
Pollution of the marine environment is a complex topic of immense scientific and political
importance, as is continuously reflected by the content of this bulletin. In order to develop an
understanding of marine pollution issues, which ultimately allows effective management strategies to
be developed and implemented, we rely on both research and monitoring of the oceans. The aim is to
identify contaminant nature, sources, distribution, concentration, persistence, uptake into biota and,
arguably most importantly, effect on the ecosystem or any part thereof. To meet the need for improved
assessment of ecosystem change over appropriate temporal and spatial scales, new measurement
strategies based on continuous or semi-continuous observations are required (Kroger et al., 2000a,b).
Sensors for temperature, conductivity, depth and turbidity have already been used to good effect in
oceanography, but it is time that the measurement of chemical and biological parameters is enhanced
to accommodate the required decentralised and high frequency observations. The way in which novel
instrumentation such as biosensors could contribute to these measurements is the subject of this
review.
2. What are biosensors?
Biosensors (Fig. 1) are analytical devices incorporating a biological material, biologically
derived material or biomimetic intimately associated with or integrated within a physicochemical
transducer or transducing microsystem, which may be optical, electrochemical, thermometric,
piezoelectric or magnetic (Turner, 1999). A wide variety of receptors have been incorporated into
biosensors including biocatalysts such as enzymes, organelles, intact microorganisms and tissues, and
affinity elements such as antibodies, cell receptors, nucleic acids and biomimetic molecules (Turner,
2000). It is important to appreciate that biosensors do not necessarily measure a biological parameter,

but use a biological sensing element to detect whatever the analyte of interest isi.e. biosensors can
and have been used
to quantify chemical pollutants. Another important feature is the ability of biosensors to measure a
class of compounds and not just single chemicals. These socalled summing parameters include
effects such as toxicity (Farre et al., 2001), carcinogenicity, mutagenicity, cytotoxicity (Castillo et al.,
2001) and genotoxicity (Billinton et al., 1998; Polyak et al., 2000, 2001), which are virtually
impossible to characterise using conventional chemical analysis. Sensors developed for these
biological effects are numerous and highly relevant to marine measurements but it is beyond the scope
of this article to provide a detailed review of the field. While broad classes of pollutant can be
measured, biosensors can also offer a high degree of specificity, often distinguishing between stereo
isomers of the same compound. Sensitivity too can be high. In the extreme case of electrochemical
immunosensors as little as 1020 mol of a substance has been measured (Jenkins et al., 1991), although
this is quite exceptional and sensors more normally operate in the nanomolar to millimolar range. The
integration of the biological sensing element with the electronic transducer system leads to a compact
design, which is easy to use, inexpensive and portable.

Biosensor technology is also amenable to mass production making sensor elements disposable and
cheap enough to be deployed at multiple sites. Another feature of biosensors as opposed to bioassays
or other analytical methods is the way that the assay design is permanently fixed in the construction of
the device. This means that a biosensor is designed to perform a particular analytical task, measuring
a set analyte or class of analytes in a defined medium. While this leads to some restriction on the
breadth of use, it has the advantage of allowing the sensor to be used directly in complex matrices
without sample preparation. This in turn leads to improved reproducibility by avoiding operator error.
The close proximity of the sensing element to the transducer also enhances the response time leading
to rapid measurement due to low diffusion times. This means that automatic analysers can have high
sample throughput or that single measurements can be delivered in real time. This speed, combined
with the continuous output of a sensor means that a much more complete visualisation of dynamic
phenomena can be obtained and, in certain circumstances, automatic corrective action can be
triggered.
The development of biosensors was originally driven by medical applications. One of the
most successful devices has been an electrochemical biosensor for glucose based on the enzyme,
glucose oxidase (Cass et al., 1984; Cardosi and Turner, 1990). The convenient pensized instrument
originally developed at Cranfield University in collaboration with Oxford University has proved a

huge commercial success (Fig. 2). It also serves to illustrate some of the features and benefits
discussed

above, offering an inexpensive, simple-to-use instrument for home blood glucose monitoring, based
on a massproducible and disposable enzyme electrode. This graphic illustration of utility has
encouraged workers to seek wider applications for the technology and environmental monitoring has
proved an attractive target (Bilitewski and Turner, 2000). While the immediate commercial returns
from environmental sensors are substantially less than from medical diagnostics, public demand and
government funding has generated a major research effort. Much of this work, however, has been
directed at fresh water monitoring and adaptation of this powerful analytical technology to the marine
environment is just beginning.
3. Important topics in marine pollution research, monitoring and control
There are different pathways by which contaminants can be introduced into the marine environment:
direct inputs, riverine contributions, and drawdown from the atmosphere. Examples for direct inputs
include discharges of industrial or domestic effluents, chemicals released in conjunction with oil
exploration or the decommissioning of oil platforms, and chemicals originating from agricultural runoffsuch as fertilisers and pesticides. Anti-biofouling agents leaching from ship paints (TBT and
booster biocides) and other releases from ships, including events such as accidents involving oil
tankers or ships carrying hazardous chemicals, also contribute to the pollution of the seas. Pollutants
reaching the marine environment mainly through riverine inputs include agricultural run-offas well as
industrial discharges and substances released from sewage treatment plants, most importantly
nutrients and pollutants such as surfactants, plasticisers (nonylphenol), steroid hormones, and other
pharmacological products. Atmospheric pollutant inputs are equally diverse and include emissions
from power plants, other industrial sites and general traffic. In addition to chemicals introduced into
the marine environment from external sources, contaminants of biological origin such as algal toxins
found in shellfish cause concern and bacterial or viral contamination of waters can have pathogenic
effects either following direct exposure or by uptake through the food chain. From the above diverse
pollutant issues, it is apparent
that the range of determinants important to measure is wide, but an equally broad range of different
biosensors have been developed that measure relevant analytes. As mentioned above only a few
biosensors have been developed, to date, directly for marine applications. Therefore we will also
highlight important technology that could usefully be adapted to the marine challenge in future years.
4. Biosensors for marine determinants
4.1. Nutrients

Two key nutrients regularly measured in the marine environment are nitrate and nitrite. The
compounds are essential for phytoplankton growth and thus form the basis for marine productivity.
Considering them as pollutants is only justified in particular cases, when their input from the
anthropogenic sources, as outlined above, far exceeds requirements and leads to undesirable effects
generally described as euthrophication. Nutrient measurements can conveniently be made using
biological sensing systems, particularly with metabolic sensors. A nitrate sensor was constructed using
nitrate reductase immobilised in an electrogenerated polymer, which simultaneously entrapped the
enzyme and connected it electrically to the electrode (Cosnier et al., 1994). Enzyme- based nitrate
measurements were also made using an optical transducer and selecting a periplasmic enzyme from
the denitrifying bacteria Thiospaera pantotropha immobilised in an optically transparent solgel
matrix (Aylott et al., 1997), which increased the enzyme stability while allowing for unhindered
optical detection. The enzyme nitrite reductase isolated from Alcaligenes faecalis has been
successfully employed in the construction of an amperometric sensor (Sasaki et al., 1998). Using
whole microorganisms rather the isolated enzymes, a nitrate (NO_3 ) biosensor has been constructed
based on immobilised denitrifying bacteria (Agrobacterium radiobacter) and electrochemical
quantification of the nitrous oxide (N2O) released (Larsen et al., 1997). By careful selection of the
denitrifying strain and continuous supply of electron donors and other nutrients from a built-in
reservoir to the immobilised cells, a very selective and stable system was obtained. The described
biosensor was subsequently used in the determination of the microscale distribution of nitrate, nitrate
assimilation, nitrification and denitrification in sediment (Lorenzen et al., 1998). Nitrification kinetics
have also been elucidated using biosensors (Ficara et al., 2000).
4.2. Anti-biofouling agents
Organotin compounds are used as biocides in antifouling paints on marine structures, including the
hulls of commercial and leisure craft. Because of its high toxicity and negative effects on non-target
biota, TBT use has been carefully monitored (Waldock et al., 1993; Law et al., 1994; volume 258 of
Science of the Total Environment, 2000 and references therein). Among the best-documented effects
of organotins on biota are direct toxicity, shell thickening in oysters, a decline in recruitment of their
juvenile stages, and endocrine disruption (imposex). Much of the evidence forTBT-mediated effects
on biota is based on empirical observations in the field, but many described effects have been
reproduced in the laboratory and the link between TBT and endocrine disruption is by now generally
accepted (Matthiessen and Gibbs, 1998). With the increasingly strict legislation regarding the use of
TBT, instruments that allow efficient decentralised compliance monitoring will be required and work
on the development of a sensor system for TBT detection is in progress in the CEFAS laboratory. In
order to replace organotins, anti-biofouling paints are now largely based on copper and other biocides
such as pesticides (see section below) and it is well possible that increased monitoring efforts will be
required to elucidate the effects of these compounds on the receiving environment.
4.3. Pesticides
Many biosensors have been developed for the detection of pesticides in complex matrices (Kroger
and Turner, 2000). By far the most common biocatalysts in the development of pesticide biosensors
based on inhibition studies are cholinesterases. Acetylcholinesterase (AChE) is an enzyme involved
in signal transduction inthe central and peripheral nervous system. Organophosphorus and carbamate
pesticides inhibit cholinesterase and are therefore powerful insecticides, but at the same time exhibit
high toxicity for other animals and humans. An alternative biocatalytic approach is the use of intact
microorganisms. The most successful commercial biosensor for environmental use is the biochemical
oxygen demand sensor (Wittmann et al., 1997). This exploits immobilised microorganisms and an
oxygen electrode to rapidly determine the amount of metabolisable organic material in a given
sample. Various approaches have been taken to adapt this measurement principle to the detection of
metabolites and inhibitors in the environment including the selection of appropriate strains and
genetic manipulation to introduce specific
transducible properties. A variation has been to use photosynthetic bacteria coupled to electrochemical

transducers to monitor herbicides (Rawson et al., 1989) (an example is shown in Fig. 3). Another
group of biological sensors relevant to contaminant measurements is affinity sensors and in particular
immunosensors. Applying antibodies as their sensing elements, they can be tailored to a range of
different analytes. The affinity constants and crossreactivity patterns of the selected antibodies
determine their selectivity and specificity.

4.4. Endocrine disruptors

In addition to the endocrine disruptor effects caused by organotins, there has been increased concern
in recent years over the presence in the aquatic environment of other compounds with estrogenic
properties, since they may interfere with the normal endocrine function in animals and affect
development and reproduction (Sadik and Witt, 1999). Among these compound are the naturally
occurring steroids estrone and 17b-estradiol, and the synthetic contraceptive steroid 17a-ethinyl
estradiol, which have been detected in surface waters and sewage treatment plant effluents (Kelly,
2000; Rodgers-Gray et al., 2000). A number of publications deal with the development of steroid
sensors based on biological or biomimetic sensing elements. The natural sensing element most
commonly used is the human estrogen receptor and examples include an acoustic biosensor for
estrogen determination using a sandwich-type monitoring assay system with a piezoelectric sensor in
combination with flow injection (Mo et al., 1999), and a voltammetric sensor based on changes in
electrode properties related to binding of the affinity ligand (Murata et al., 2001). Estrogenicity has
also been determined in surfaces waters by using a biosensor to detect the key biomarker,
vitollogenin, in fish (Sole et al., 2000) and an portable immunosensor (Fig. 4) for detection of
hormone mimics based on surface plasmon resonance has been described (Sesay and Cullen, 2001).
4.5. Molecular taxonomy
Species identification is crucial to marine ecosystem monitoring. The species under investigation can
include anything from large animals such as whales or fish right down to microscopic organisms such
as zoo or phytoplankton. Particularly for the latter groupthe very small or even single cell
organismsaccurate taxonomy can be very difficult and sensor technology offers a valuable
alternative. A good example of such an application is the increasing international pressure to monitor
variations in phytoplankton ecology, as an indicator of nutrient-driven eutrophication, and to develop
better methods for managing the consequences of the presence of toxic algae. Effective monitoring for
presence, changing species composition or bloom development would preferentially be carried out
using continuous automated remote or in situ approaches because of the episodic and sporadic nature
of algal growth and our limited ability to predict bloom occurrence. At present algal taxonomy is
mostly performed by specialist in laboratories on fresh or preserved samples using optical
microscopy, but there are cases of morphologically nondistinct species with different toxicity where
these methods are inadequate. Molecular biological methods of species identification based on
characteristic nucleic acid sequences are rapidly gaining recognition as valuable additions or
alternatives to classical taxonomy based on visual inspection of morphology (Medlin et al., 1998;
Scholin, 1998; Haley et al., 1999). Some sequence information for organisms of interest is already
published and as world-wide sequencing efforts progress increasingly comprehensive databases
become available for molecular biological species identification. The group around Chris Scholin at
Monterrey Bay Research Institute has made significant advances in the development and application
of RNA probes to algae species identification and in the adaptation of these methods to user-friendly
formats such as an automated sandwich assay in microtiter plate form and whole cell assays using
fluorescently labelled probes (Scholin et al., 1999).
In order to realise the full potential of these new methods the use of nucleic acid-sensor systems is
desirable to move to automated systems and sensors. Some approaches were presented at the
International Conference on Harmful Algal Blooms and the Workshop on Harmful Algal Bloom and
Marine Biotoxin Monitoring (HAB, 2000; HAB-TECH, 2000). Much interesting work on detection

methods for algal toxin monitoring has been conducted or is in progress and a separate review of this
specific topic would be desirable. For taxonomy, nucleic acid sensors represent a subgroup of affinity
recognition-based sensors, in which the recognition event is the sequence specific hybridisation
between a nucleic acid released from the target organism and a specific, usually short probe sequence.
By strictly controlling the environmental parameters, this interactioncan be made highly selective
(thus taxonomically relevant) and the number of publication describing different types and formats of
nucleic acid sensors is increasing at great speed.

5. Emerging technique: molecular imprinting

One of the critical requirements in the development of marine biosensors is their robustness.
Antibodies, enzymes and receptors can suffer from lack of stability. One of the emerging techniques,
offering some advantages with respect to receptor stability is molecular imprinting. Molecular
imprinting is a generic technology, which introduces recognition properties into synthetic polymers
using appropriate templates. The typical recipe for MIP preparation includes mixing together the
target compound (template) with a corresponding functionalmonomer (most frequentlymethacrylic
acid) and cross-linker (e.g. ethylene glycol dimethacrylate) in appropriate solvents (chloroform,
acetonitrile) and polymerising this mixture using UV or chemical initiation (Fig. 5).

6. What is required for existing and future systems to meet the marine challenge?
The potential benefit of using biosensors and biomimetic sensors for marine monitoring have been
outlined above and hopefully will stimulate some thoughts and discussions about worthwhile
applications. Despite all the promise the technology holds for future observational strategies, some of
the factors that present considerable hurdles and can hold back application should also be considered.
In order for high-frequency measurements to be made successfully in situ, the instruments employed
need to be very robust in terms of possible physical impacts and functionality in a highly corrosive
working environment. This is mainly an engineering challenge and has been overcome for the
instruments available to date, but is important to consider when selecting and developing new
technology. When integrating new sensors into measurement systems such as the CEFAS SmartBuoy

it is important to obtain signal stability over deployment periods of minimum 46 weeks, preferably
significantly longer The advantage of this approach is not only the fast availability of the data but also
the inherent risk minimisation: even if at some point during the observation interval instrument loss
occurs, the data to that point is safe. Additionally, back-up data loggers associated with the
instrumentation are required to safeguard the user against loss of data due to telemetry failure. Ideally
sensors would be completely drift-freea specification which is rarely achievable. Sensor drift is
acceptable if it can be accurately predicted and accounted for or if frequent re-calibration is part of the
measurement schedule. Sensor calibration can involve strategies such as use of a stable on-board
standard, which is inserted between analyses and/or collection of calibration samples at intervals
during the measurement period for subsequent analysis in the laboratory. The second approach relies
on retrospective accounting for any sensor drift with the calibration data obtained. Biofouling is
always a problem for sensors submerged for prolonged periods and measures have to be taken to
eliminate its influence on sensor readings. Much thought and research is going into the prevention of
biofouling, ranging from the selection of the most appropriate surfaces, to chemical use, UV
irradiation or the application of physical shutters to minimise sensor surface exposure. It is beyond the
scope of this paper to review anti-biofouling strategies, but it is certainly a topic relevant to any sensor
developer. Power consumption is also an important point for measurements at sea. As power is still
most commonly supplied by battery packs, it can become a limiting factor. Miniaturisation, resting
periods between readings in which the sensors are not powered up, and the use of less power-hungry
technology and measurement as well as intelligent data-logging regimes all have to be considered.
Another feature of marine pollution monitoring is that the associated concentrations of pollutants are
frequently very low (micromolar to nanomolar or even less). Therefore sensors have to be optimised
not only for excellent stability but also for high sensitivity. Where this sensitivity cannot be achieved
in a primary system, sample enrichment methods have to be considered. Work is in progress at
CEFAS on the development of an automated in situ solid-phase-extraction instrument, which has the
potential to be a modular addition to any sensor system for pollutants.
7. Conclusions
From the above it should be apparent that the development of biosensors for in situ measurements is a
difficult but very worthwhile challenge and that work is underway to address some of the issues
raised. While the ability to make such observations is the ultimate goal of most developments, already
the ability to make decentralised measurements in the field rather than having to process collected
samples in specialised laboratories is a significant advance which should not be overlooked. This is an
area where biosensors are likely to find their initial significant applications: the low-cost, userfriendly test that can be carried out to pre-screen samples, possibly helping to direct the sample
collection effort for
further detailed chemical analysis.