Elsevier

Talanta

Volume 58, Issue 1, 16 August 2002, Pages 181-188
Talanta

Evaluation of a GFP reporter gene construct for environmental arsenic detection

https://doi.org/10.1016/S0039-9140(02)00266-7Get rights and content

Abstract

Detection of arsenic and other heavy metal contaminants in the environment is critical to ensuring safe drinking water and effective cleanup of historic activities that have led to widespread contamination of soil and groundwater. Biosensors have the potential to significantly reduce the costs associated with site characterization and long term environmental monitoring. By exploiting the highly selective and sensitive natural mechanisms by which bacteria and other living organisms respond to heavy metals, and fusing transcriptionally active components of these mechanisms to reporter genes, such as β-galactosidase, bacterial luciferase (lux), or green fluorescent protein (GFP) from marine jellyfish, it is possible to produce inexpensive, yet effective biosensors. This article describes the response to submicrogram quantities of arsenite and arsenate of a whole cell arsenic biosensor utilizing a GFP reporter gene.

Introduction

Arsenic contamination of soil and groundwater is a problem worldwide, resulting from natural geologic activity (e.g. arsenic in surface water and geothermal springs in Yellowstone National Park, and arsenic in tube wells in Bangladesh) and manmade sources (forest products, mining, heavy industry and semiconductor manufacturing). The health effects of arsenic range from acute toxicity resulting from incorporation of arsenate as a toxic phosphate analog, or arsenite inactivation of sulfhydryl containing proteins, to chronic effects including CNS damage, skin pigmentation abnormalities (particularly of the palms of the hand and base of the foot), and cancer. In the United States, the Environmental Protection Agency recently reduced the enforceable Maximum Contaminant Level (MCL) for arsenic in drinking water to 10 μg l−1 from the previous limit of 50 μg l−1, reflecting revised risk calculations for bladder cancer resulting from arsenic exposure [1].

It has been estimated that the revision of U.S. and international drinking water standards to more conservative limits for arsenic will require the retesting of millions of wells (the US EPA estimates only 4000 water systems in the US will be impacted; [1]). In a recent report supporting the US EPA's decision to lower the MCL for arsenic in drinking water [2], it is reported that arsenic testing by established methods, including ICP-MS, ICP-AES, gaseous hydride atomic absorption spectrometry (GHAA), graphite furnace atomic absorption spectrometry (GFAA) or anodic stripping voltammetry (ASV), currently costs between $10 and 50 per sample. The cost of analytical instrumentation associated with these testing ranges from $30 000 to 200 000 per instrument, depending on the technique. Therefore, analytical testing may require substantial investments, if not in instrumentation, then certainly in increased sampling, to assure compliance with this new standard in the future.

Biosensors present an alternative to conventional analysis of heavy metals, such as arsenic, and other environmental contaminants. Biosensors have long been touted as a viable solution, owing to the selectivity and sensitivity of biomolecular interactions, and their relative simplicity and low cost. The simplicity of a biosensor is particularly attractive when considering fieldwork or even in situ placement of such sensors, especially when compared to the size, bulk and fragility of many analytical instruments. Biosensors also provide a means of directly measuring bioavailable contaminant concentrations, which can accelerate the determination of a successful environmental restoration campaign. On the other hand, widespread deployment of biosensors is limited due to limited operational lifetimes and deleterious effects on biosensor performance due to interfering substances or the nature of the operational environment itself (factors that may include temperature, pH, TDS, etc.) Regulatory agencies and other stakeholders are unlikely to accept biosensors as a replacement for conventional instrumental analysis in establishing site contamination levels or remediation endpoints, but biosensor technology does provide a means to augment more expensive contaminant quantification, particularly where they might be applied to increase the sampling frequency, or to identify hot spots for validated conventional analysis during site characterization.

A commercial test kit for arsenic utilizing a reporter gene (firefly luciferase, lucFF, rather than the bacterial lux used in other research discussed here) is available [3]. The literature provided for this test describes a whole microbe-based assay requiring 2.5 h that can be measured in a luminometer to detect total arsenic and arsenite. This reporter gene has also been used to construct whole cell biosensors for cadmium and zinc [4].

A variety of heavy metal biosensors have been described in the literature. In simplest form, the toxic effects of heavy metals and other contaminants have been measured with whole microbial cells. For example, inhibition of O2 generation has been measured with an amperometric gas diffusion electrode in Spirulina subsalsa [5]. The decrease in natural luminescence emitted by photoluminescent bacteria has been used either in suspension [6] or immobilized on solid substrates [7] to quantify toxicants in environmental samples, and are the basis of the commercial Microtox™ system. The genetic determinants of light production in the photoluminescent marine bacterium, Vibrio fisheri, and the green fluorescent protein of the marine jellyfish, Aequorea victoria have been particularly useful in the construction of reporter genes that function in whole cell biosensors. One such sensor and its construction will be described below.

Most recently, immobilized proteins have been used to produce potentiometric or capacitance-based biosensors. Urease has been immobilized on iridium oxide electrodes to produce biosensors for mercury and other metals [8]. A potentiometric biosensor has been prepared by immobilizing the mercury-binding regulatory protein associated with bacterial mercury resistance, MerR, onto a gold surface [9]. Likewise, similar sensors have been assembled using the metal-binding protein, metallothionein, or MerR [10]. In that study, metallothionein-based sensors provided a more generalized response to heavy metals compared to the relatively specific response of MerR biosensors, as would be expected given our knowledge of the metal binding properties of these two different proteins.

Bacteria have evolved resistance mechanisms to many toxic metals [11], [12], [13]. These mechanisms typically involve enzymatic redox transformations of metal ions once they have gained access to the cell. Oxidoreduction of a particular metal usually leads to a less toxic form of the metal (for example, the reduction of mercuric ion to elemental mercury [14], [15]), but in some cases, produces a more toxic intermediate which is actively exported (as in the case of arsenic detoxification, where arsenate ion is reduced to arsenite, which is then exported by a specific arsenite transporter protein; see below for details). As these mechanisms have been elucidated, it has been surprising that no mechanisms to prevent entry of toxic metals into the cell have been discovered to date.

A variety of heavy metal resistance systems have been dissected down to their genetic determinants, including mercury, cadmium, zinc, cobalt, copper, chromium, silver, and arsenic (several excellent reviews have been written recently on this topic; [11], [12], [13]). These genetic determinants are typically organized in an operon, and are frequently carried on extrachromosomal elements, including plasmids and transposons, which suggests horizontal transfer of heavy metal resistance to a wide range of bacteria in the environment. Localization of these determinants on mobile genetic elements may also provide a mechanism for increased resistance within an organism by increasing the copy number of the element carrying the metal resistance traits.

Genetic control of metal resistance mechanisms is regulated to prevent unnecessary production of nonessential proteins in the absence of heavy metals. Primary control occurs at the transcriptional level, by preventing RNA polymerase from binding to the operator/promoter for the operon. A repressor protein also encoded by the metal resistance operon binds to the operator/promoter region upstream of the structural genes in the absence of the requisite heavy metal (for example, MerR in the case of mercury resistance [16], [17], and ArsR for arsenic resistance [18], [19], [20], [21]). Upon binding of the metal ion, the repressor protein undergoes a conformational change and dissociates from the transcriptional control sequences. At this point, efficient transcription of the metal resistance system occurs. For bacterial arsenic resistance (in Gram negative microorganisms), the interaction of arsenite (as well as bismuth and antimonite) with the ars operon repressor protein, ArsR, leads to derepression of transcription and the production of ArsA (an arsenite-specific ATPase; [22], [23]), ArsB (an arsenite-translocating, transmembrane efflux channel driven by ATP hydrolysis in Gram negative bacteria; [24], [25]), ArsC (an arsenate reductase, which takes any arsenate within the cell and converts it to arsenite that is pumped out of the microbial cell through the ArsA/ArsB anion pump; [26], [27]), and finally, ArsD, a regulatory protein which provides additional control over expression of the arsenic resistance system [28]. The genetic organization of plasmid borne arsenic resistance determinants is somewhat different in the Gram positive Staphylococcus aureus plasmid, pI258, where arsA and arsD are absent [29], and the ArsB protein alone directs arsenite efflux out to the cell [30] without an ATPase as in the case for Gram negative bacteria. Both genetic systems have been used to construct reporter genes for arsenic [31], [32], [33], [34].

The specificity and sensitivity of the genetic elements and their corresponding proteins to heavy metal exposure are key to any resulting biosensor. In the case of the arsenic biosensor we describe here, we have constructed an arsenic specific reporter gene construct, in which the arsenic-binding regulatory protein gene, arsR (and a small portion of arsD) and its upstream promoter has been fused to the structural gene for green fluorescent protein from the marine jellyfish, Aequorea victoria [35]. When expressed in recombinant Escherichia coli, this reporter creates a bacterial strain that produces GFP in response to arsenic exposure. While there have been a number of publications describing similar reporter gene constructs wherein arsR has been fused to the bacterial luciferase genes (lux; [32], [33]), or alternatively, to firefly luciferase (lucFF; [34]) we sought to determine the potential advantages of a fluorescent reporter which does not rely on internal reducing equivalents being produced by the cell, and which might therefore be less sensitive to growth and/or nutritional status, and potential inhibitory materials in the sample.

Section snippets

Bacterial strains and plasmids

The O/P region, as well as the entire arsR gene and initial region of arsD, were amplified by PCR from pIRC120 [36] using oligonucleotide primers incorporating XbaI and KpnI restriction endonuclease cut sites (restriction sites in bold: arsRF:5′-GGTTCTAGATCAACCGCATAGTGACACCTGC-3′; arsRR:5′-TCAGGTACCCGACATCTGAACCACAAACGCC-3′). The primers used to amplify this fragment were designed based on the published sequence of the R46 ars operon [37]. This operon is similar in gene organization and DNA

Results and discussion

We used a 96-well microtiter plate spectrofluorimeter for measurements of GFP fluorescence to provide an efficient format for monitoring fluorescence produced in E. coli upon exposure to a range of arsenic concentrations. Some background fluorescence was observed from both the media used (M-9, a minimal bacterial medium) and the bacterial cells themselves. In order to detect GFP fluorescence, we determined that a cell density on the order of 109 cells ml−1 was necessary to exceed the background

Uncited reference

[46]

Acknowledgements

The authors thank Kevin Dolbeare for primer design and amplification of the arsRD fragment. This work was supported by funding provided by the U.S. Department of Energy, Office of Environmental Management, to the Idaho National Engineering and Environmental Laboratory, operated by Bechtel BWXT Idaho, LLC under contract DE-AC07-99ID13727.

References (48)

  • S. Tauriainen et al.

    Biosens. Bioelectron.

    (1998)
  • L. Campanella et al.

    Water Res.

    (2001)
  • S. Lee et al.

    Biosens. Bioelectron.

    (1992)
  • I. Karube et al.

    Curr. Opin. Biotechnol.

    (1994)
  • I. Bontidean et al.

    J. Inorg. Biochem.

    (2000)
  • T.K. Misra

    Plasmid

    (1992)
  • J. Wu et al.

    J. Biol. Chem.

    (1993)
  • W. Shi et al.

    J. Biol. Chem.

    (1994)
  • B.P. Rosen et al.

    J. Biol. Chem.

    (1988)
  • S. Dey et al.

    Arch. Biochem. Biophys.

    (1994)
  • J. Wu et al.

    J. Biol. Chem.

    (1992)
  • C. Chen et al.

    J. Biol. Chem.

    (1986)
  • A.M.C. Brown et al.

    Plasmid

    (1981)
  • J.M. Collard et al.

    FEMS Microbiol. Rev.

    (1994)
  • P. Corbisier et al.

    Anal. Chim. Acta

    (1999)
  • C. Tibazarwa et al.

    Environ. Poll.

    (2001)
  • B.P. Cormack et al.

    Gene

    (1996)
  • United States Environmental Protection Agency, EPA 815-F-01-010...
  • United States Environmental Protection Agency, EPA-815-R-00-010...
  • ...
  • T. Krawczynski et al.

    Biosens. Bioelectron.

    (2000)
  • I. Bontidean et al.

    Anal. Chem.

    (1998)
  • S. Silver et al.

    Environ. Health Perspect.

    (1994)
  • S. Silver et al.

    Ann. Rev. Microbiol.

    (1996)
  • Cited by (0)

    View full text