In general, heavy metals are the metals with relatively high mass densities (above 5 g/cm³), such as platinum, gold, mercury, and lead. The definition of heavy metals varies from discipline to discipline due to different concerns and perspectives. For physiology and pathology, heavy metals mainly refer to heavy elements with obvious biotoxicity. In terms of toxicity, the most common heavy metals include lead (Pb), mercury (Hg), arsenic (As), cadmium (Cd), and chromium (Cr) ^[1][2][3][4][1,2,3,4]. These heavy metal ions (HMIs) accumulate in the human body and cause chronic poisoning, resulting in damage to the immune, gastrointestinal, reproductive, and nervous systems ^[5][6][7][5,6,7]. Long-term exposure to heavy metals may even cause cancer in humans [8]. Being non-biodegradable, HMIs will continue to exist for decades and even hundreds of years if they are released into the environment [9]. With increasing human activities in mining and manufacturing, etc., HMI contamination and pollution pose great risks to ecological and human health ^[10][11][10,11]. Therefore, it is of great importance to be able to effectively screen for and quantitatively determine HMIs in a range of situations, including food safety, environmental pollution, and clinical diagnosis. Many international organizations and regulatory bodies, such as the World Health Organization (WHO) and the U.S. Environmental Protection Agency (EPA), have listed heavy metals as priority substances for intensive monitoring [1]. In practice, most countries have set standards on the permissible limits for HMI concentrations in water or other mediums to protect the environment and public health.

Traditional HMI measurements are mostly performed using spectroscopic techniques, including atomic absorption spectroscopy (AAS) ^[4][12][4,12], X-ray fluorescence spectrometry (XRF) ^[4][13][4,13], and inductively coupled plasma mass spectrometry (ICP-MS) ^[4][14][4,14], which are gold standards in HMI detection. However, they are expensive, difficult to perform, require complicated pretreatment for the samples, and need to be operated in a central laboratory setting. Thus, traditional spectroscopic techniques cannot meet the needs for pervasive HMI monitoring and screening. In contrast, surface-based biosensors are a good candidate to realize low-cost, sensitive HMI detection with a quick turn-around time. Similar to other biochemical sensors, there are three important components in an HMI sensor: analyte recognition, physical support for probe immobilization, and signal transduction. Various types of sensors have been reported, including nanoparticles ^[15][16][17][15,16,17], field effect transistors (FETs) ^[18][19][20][18,19,20], and working electrodes for voltammetric analysis [21]. In terms of signal transduction, there are mainly optical [15] and electrochemical methods [4]. Specific target recognition is always critically important to analyte detection. Antibody [22], nuclei acids ^[23][24][23,24], and peptide/amino acids ^[25][26][25,26] are the most commonly used probes to recognize and bind with HMIs. An antibody as a probe has the best specificity, but faces challenges when detecting HMIs. HMIs are too small to be directly detected with good sensitivity by an antibody, and antibody binding reactions may require a fluid environment different from that HMIs naturally exist in. Two types of functional nuclei acid probes have been developed for metal sensing: aptamers and DNAzymes. It is well-known that screening aptamer sequences to find one that specifically binds with a certain HMI is difficult, in addition to being laborious and time-consuming. Very few metal ions have successfully found their specific aptamers ^[27][28][27,28]. DNAzymes can selectively detect HMIs based on their catalytic activity; however, another nuclei acid sequence is required to be added to work as the substrate–enzyme pair. During metal detection by DNAzyme, additional steps and time are needed to allow for catalytic reactions between the DNAzyme and the substrate. Besides antibodies, aptamers, and DNAzyme, another option is to use a peptide or an amino acid as a probe to bind with HMIs, such as cysteine [29]. The most widely reported peptide probe for HMI detection is glutathione (GSH) [30]. Because there are six possible coordination sites in GSH for binding with metal ions, it has distinct advantages in capturing HMIs when compared with cysteine, especially in case of lead ion (Pb²⁺) detection. Furthermore, it is highly stable, cost-effective, and easy to immobilize on a sensor. Therefore, GSH, as a promising probe for HMI detection, has received much attention in recent years.

Glutathione (GSH) is a ubiquitous tripeptide biomolecule (γ-Glu-Cys-Gly) widely found in animals and plants with a concentration from 0.1 mM to 10 mM [31]. GSH is known as an antidote to prevent cells from toxicosis of heavy metal ions by forming complexes as a key step in biological detoxification processes [30].

The chemical structure of GSH is shown in Figure 1. If oxidized, GSH will form disulfide bonds between themselves, and become what is known as glutathione disulphide (GSSG). GSSG lacks the ability to bind with electrodes or nanoparticles. In practice, GSH in its reduced form is used to bind with HMIs. Reduced GSH is a linear tripeptide of L-glutamine, L-cysteine, and glycine, so reduced GSH is also labeled as L-GSH. GSH contains one amino (-NH2), one sulfhydryl (-SH), and two carboxyl (-COOH) groups, all with the ability to bind with HMIs. Therefore, GSH is a versatile ligand with several binding modes.

GSH has a sulfhydryl group on the cysteinyl portion. The sulfhydryl group is known to coordinate with a wide range of metal ions. For surface-based or nanoparticle-based HMI sensors, GSH is usually immobilized on gold nanoparticles or electrodes. Numerous experiments indicate that there is a strong binding force between gold and the (-SH) group. It is expected that, during probe immobilization, Au–S bonds are formed to link GSH molecules onto the gold surface ^{[32][33][34][35]}[32,33,34,35]. As a result, only one amino (-NH₂) and two carboxyls (-COOH) are still available to bind with metal ions.

A number of published papers have demonstrated that via the two free (-COOH) groups, strong binding with Pb²⁺ is preferred over that with other HMIs ^[35][36][37][35,36,37]. It was found that the solution pH has a strong effect on the affinity of the GSH–Pb²⁺ complex. At around pH 7, (-COOH) is deprotonated to (-COO-). Known as an oxyphilic ion, Pb²⁺ can combine with four acetate molecules or eight oxygen atoms [33]. Within a pH range of 4–8, binding between Pb²⁺ and (-COO-) will predominate over other reactions. Under the same pH condition, (-NH₂) is protonated to (-NH₃⁺), which is hypothesized to prevent GSH from combining with other ions due to its charges ^[33][37][33,37]. Therefore, good affinity and selectivity can be achieved for GSH–Pb²⁺ binding. There exist a number of reports on using GSH to specifically detect Pb²⁺ from a mixture of multiple HMIs.

Nevertheless, detection of metal ions other than Pb²⁺ can also be achieved using GSH-based sensors. The GSH support material and the solution pH value play an important role in the HMI–GSH complexation pattern. Research has shown that GSH’s conformational structure is strongly influenced by the solution pH [38]. Various functional groups in GSH can participate in complexation under different working conditions, leading to a certain selectivity. For example, GSH-capped quantum dots (QDs) are often used to detect HMIs other than Pb²⁺. In [39], GSH immobilized on sulfur- and nitrogen-co-doped carbon QDs presented free (-SH) groups for binding, which showed strong affinity to Hg²⁺ as opposed to other metal ions (Ag⁺, Al³⁺, Ba²⁺, Cu²⁺, Pb²⁺, etc.) and realized specific detection of Hg²⁺. Simultaneous detection of Cr²⁺ and Pb²⁺ was reported using GSH-coated magnetic nanoparticles (NPs) [40], as GSH molecules were immobilized onto Fe₃O₄ NPs through a hydrogen bond. In this detection scheme, NH⁴⁺, Na⁺, Mg²⁺, Hg²⁺ and other MIs caused no obvious responses, except for Cd²⁺ and Pb²⁺. This was attributed in part to the negative charges on the functional groups of GSH [40], since the testing was done at a pH of 4.5. Simultaneous detection of multiple HMIs is thus possible by adjusting the solution pH and the GSH immobilization strategy. However, complexation reactions between GSH and HMIs are complicated and difficult to predict, and the detection outcome under particular conditions usually needs to be verified through experimental results.

The foregoing introduces the interactions and formation of complexes between GSH and HMIs, which are the basis for the specific recognition of certain HMIs by GSH-based detection methods. To convert specific binding between GSH and HMIs into a quantifiable signal, a transducer is needed, which can take many different shapes and forms. Transduction schemes may have effects on the sensor performance as well, especially the plasmon-based detection that depends on the interaction between metal ions and nanoparticles. In next section, various GSH-based detection platforms are presented and discussed, with a focus on optical and electrochemical methods.

Based on a signal transduction strategy, most of the GSH-based HMI detection methods can be categorized into two groups: optical methods and electrochemical methods. Then, each method is further divided into subtypes based on the support materials for GSH.