Introduction
The heavy metals pollution, as consequence of the industrial progress and the human activities, is one of the important issues in the environmental field representing an appealing research. Heavy metals are involved in many pathologies from cancer, neurodegenerative and metabolic diseases. One of the most contaminated natural resource is water, since metals are often spilled in soils through the industrial and consumer wastes and accumulate in streams, lakes, rivers and water intended for direct consumption bringing a series of implications for the human safety (Masindi, Muedi, 2018; Verma, Dwivedi, 2013). The water poisoning by arsenic is one of the most studied subjects in this field. Arsenic is a natural metalloid element diffused at different levels, from soil to water and air. Within this distribution, it can react with oxygen or other molecules producing various inorganic compounds, among which the toxic arsenite (AsIII) (ATSDR, 2007). About 140 million people in 50 countries are under the risk to use drinking, food preparation and irrigation water that is poisoned by AsIII at levels above the World Health Organization (WHO) provisional guideline value of 10 μg/L, becoming very dangerous for human safety (World Health Organization, 2020; Ravenscroft et al., 2009). When high concentrations of AsIII are present, the accumulation inside the human organism by ingestion causes different effects from simple skin lesions to more dangerous systemic disorders. Prolonged exposures to AsIII, in fact, have been linked to cardiovascular diseases, diabetes and liver, prostate and bladder cancer, while other studies suggest a relationship with neurological effects and reproductive organs (Hong, Song, Chung, 2014). Moreover, it seems to influence the cognitive development and the incidence of youthful deaths (Mandal, Suzuki, 2002).
Therefore, considering the arsenic toxicity it is extremely important the availability of effective sensing technologies able to detect and quantify AsIII in water. Unfortunately, many of the available detection systems have several limitations mainly due to the complexity of the analysis. Spectrometric (CV-AAS, AEF, CV-AFS), chromatographic and potentiometric technologies even if they are the advantages to be consolidate methods with low detection limits (LoD) of about 1 µgL-1 or 1 part per billion (ppb) and a wide range of linearity in the determination of arsenic, however, present some drawbacks in terms of time/cost consuming, long procedure, expensive reagents, the need for a sample pre-concentration, lab constraints referred to the bulky instrumentations and the high trained personnel required to perform the analysis (United States Environmental Protection Agency Office of Water, 1999; Yu, Wang, 2013; Bose, Rahman, Alamgir, 2011).
In this context, recently ultrasensitive method for dual-mode detection of arsenic by colorimetric and surface-enhanced Raman SERS using glutathione functionalized Au-nanoparticles has been described reaching LoD of 0.14 ppb (Lia et al., 2020).Colorimetric assay on Paper-based microfluidic device are has been developed allowing a rapid and low-cost detection through a direct observation of a colour change induced by the reaction of arsenic with specific dyes (Morita, Kaneko, 2006; Martinez et al., 2010). This detection method is performed on paper strips or microfluidic system so that the reagents are automatically mixed and give a colorimetric signal. However, the main drawback of the procedure is the specificity, since other molecules can cross-react with the sensing dyes (phosphates and silicates, for example, compete to react with molybdenum blue) giving a false positive and the sensitivity reaching a LoD value of 10 ppb (Lace et al., 2019; Yogarajah, Tsai, 2015) .
Among the heavy metals pollution, also Hg represent a very dangerous analyte. Actually, in its divalent inorganic form (HgII), it is frequently spilled in water as waste of industrial activities and accumulates over 1 ppb (defined by the WHO as the threshold value for human safety) becoming toxic and bringing a high risk to contract severe diseases as neurological disturbances, skin rash and kidney failure. Therefore, a punctual quantification of mercury traces in water is necessary and is, usually, performed by combined gas/liquid chromatographic-spectrometric methods (GC-AFS, HPLC-AAS, HPLC-ICP-OES, etc.) and surface-enhanced Raman scattering (SERS) techniques (Kodamatani et al., 2011; Hashemi-Moghaddam, Saber-Tehrani, 2008; Guerrini et al., 2014). These approaches share a good reliability, high sensitivity (LoD of 0.14-0.17 ppb) and accuracy discriminating among multiple mercury species (as methylmercury, ethylmercury, phenylmercury and the mercury ions) with extreme precision. However, they have some limitations in terms of lab constrain, referring to the bulky instruments, and time and cost of analysis, considering the reactants and procedure of sample preparation required for the mercury speciation, which make these methods unsuitable for on-site measurements. In this sense, many attempts have been made towards the miniaturization of the entire detection system, as for the sensing platform based on optical (both colorimetric and fluorometric) and electrochemical detection (Santangelo et al., 2014). Optical sensors for the HgII detection in water provide an improvement in terms of time and miniaturization of analysis but are limited by the low sensitivity (most of the colorimetric sensors report a LoD of 10 ppb) and the risk of cross-sensitivity towards other metal ions (Chen et al., 2015; Yang et al., 2018).
The above reported limitations can be overcome by using quantitative transduction technologies based on electrical signal, as those based on the Anodic Stripping Voltammetry or the Screen-Printed Electrodes, where the measurement of the analyte concentration is not influenced by the size of the sample used (Renedo, Alonso-Lomillo, Martínez, 2007). In this context, the continuous development of electrochemical silicon-based technologies together with miniaturization of biotechnologies can give an opportunity to develop portable and easy-to-use sensing devices able to solve the lab constraints above described particularly those related to the need of specialized personnel. The effectiveness of this approach has been already demonstrated in several areas of medical field such us nucleic acids analysis (Petralia, Sosentino, Sinatra et al., 2017; Petralia, Sciuto and Conoci, 2017; Petralia, Sciuto, Di Pietro, et al., 2017) glucose and biomolecules sensing (Petralia et al., 2018). Electrochemical systems share the advantages of a high miniaturization degree, especially for those improved with silicon technology (Petralia, Sciuto, di Pietro et al et al., 2017), and a rapid and high-sensitivity analysis. Actually, it has been found in case of AsIII detection reaching a LoD >1 ppb that is well below the WHO value. However, the sample preparation requires a pre-dilution of real samples before analysis that can introduce additional variability into the measurement and interferences (Toor, Sharma, Bansod, 2015). However, electrochemical systems often require an expensive fabrication, considering the functionalization and electrodes modification, and imply weak sensing elements, as for the enzyme-based sensors (Pujol et al., 2014; Fu et al., 2011).
To overpass these limitations, electrochemical sensors based on Noble Metal nanoparticle-modified electrodes (Pd, Pt, Ag etc.) have been developed for a wide series of applications (Petralia, et al, 2012; Majid et al., 2006). Although they show excellent performances in terms of sensitivity, however few data for specificity are reported (Kumar et al., 2017; Babar et al., 2019). In this context, sensing strategies based on specific enzyme are very appealing since they can offer selective detection. In the case of arsenic direct enzymatic recognition by arsenite oxidase Male et al., 2007) or indirect recognition trough β-galactosidase have been reported (Stocker et al., 2003), reducing the risk of cross-reactivity and increasing both the sensitivity and selectivity. However, both these technological approaches present some issues related to the specificity of the nanostructures-based sensor and to the costs production and stability of enzyme-based device which degenerate after a long usage inhibiting the recognition activity.
More recently, whole-cell based biosensors received a widespread attention for their properties. Most of them are based on microorganisms, especially bacteria, that have been genetically modified in order to be sensitive towards the specific analyte of interest (Sciuto et al., 2019; Gu, Mitchell, Kim, 2004). These biosensors use engineered bacterium (generally an Escherichia coli ) as sensing element that, each time it interacts and internalises the analyte to detect, produces a reporter protein that express a detection signal (optical, electrical and/or electrochemical). The genetic modification of the sensing element makes the whole-cell-based biosensor extremely customizable. Moreover, thanks to its intrinsic properties, the sensing bacterium is physiologically more robust than enzymes, in case of prolonged usage allowing a high degree of miniaturization and portability, due to the small size and the survival at different environmental conditions (Gilchrist et al., 2005). Finally, the genetic recognition of arsenic, based on the modified expression of thears operon sequences, β-galactosidase or luciferase gene, guarantees a strong selectivity excluding the risk of cross-reactivity (Gui et al., 2017). However, the biotechnology has the main limitation of the sample management since the sensing bacteria are strictly dependant on the nutrients availability and the environmental parameters (pH, temperature, or ionic strength) and, usually, require a long time period to complete the genetic recognition of target, from the arsenic internalization to the reporter protein transcription and translation.
In this work we propose an innovative miniaturized electrochemical biosensing platform for the specific and high-sensitive quantification of metal ions in water sample. The sensor system exploits the synergy between two interfaced sensing modules: (a) a whole-cell-based module using an engineered Escherichia coli as whole-cell sensing element; (b) an electrochemical module based on a silicon chip, integrating electrochemical cells (EC-cell) composed by three planar microelectrodes, and a portable EC-reader, performing a cyclic voltammetry (CV) analysis. The whole-cell sensing element has been genetically modified to produce a redox active 4-aminophenol, as mediator, each time it interacts with the metal target in a highly specific manner. Thanks to the electrochemical detection of the mediator, the metal is indirectly detected and quantified. Sensitivity, robustness and selectivity of the sensing system have been fully studied for AsIII and preliminary investigated for HgII proving the fully versatility of the system towards multiplex heavy metals detection.