Bacteria that thrive in environments contaminated with toxic compounds are usually endowed with diverse molecular mechanisms related to tolerance to these chemicals (Top and Springael, 2003; Permina et al., 2006). From a physiological point of view, mechanisms for resistance to stressors are usually energy dependent and should be tightly regulated in order to avoid wasting valuable resources when stressors are absent (Nojiri et al., 2004; Diaz et al., 2013). Many bacteria have evolved molecular mechanisms to handle exposure to toxic metals and metalloids. The genes encoding such systems are usually transcriptionally controlled by a plethora of regulatory systems that coordinate gene expression in response to the presence of the cognate chemical.

Several regulatory systems dedicated to the sensing of toxic metals and metalloids have been characterized in bacteria (Busenlehner et al., 2003; Kaur et al., 2006; Permina et al., 2006; Reyes-Caballero et al., 2011). Genomic studies have demonstrated that most of these systems are broadly distributed among different bacterial phyla (Paez-Espino et al., 2009). For instance, transcription factors belonging to the SmtB/ArsR family have been expensively characterized for their role in sensing and controlling gene expression in response to divalent metals (e.g., zinc, nickel, and cadmium) or toxic metalloids (e.g., arsenic and antimonite) (Busenlehner et al., 2003; Eicken et al., 2003; Qin et al., 2007). Members of this family are usually small (~100 aa) regulatory proteins that repress gene expression by blocking access of RNA polymerase to target promoters in the absence of the cognate inducer (VanZile et al., 2000; Kar et al., 2001; Morita et al., 2001; Cavet et al., 2002). These proteins usually act as dimers that bind to target DNA sequences in the apo form. Once in complex with their specific target metal/metalloid, the regulators strongly decrease their affinity for DNA allowing dissociation from the promoter and subsequently, gene expression activation (Kar et al., 1997; Morita et al., 2002, 2003; Eicken et al., 2003; Chauhan et al., 2009).

Understanding the molecular mechanisms behind the transcriptional response to toxic metals and metalloids in bacteria has led to a growing interest in repurposing these systems to construct biosensors for detection of these chemicals in the environment (Stocker et al., 2003; Siddiki et al., 2011; Cortés-Salazar et al., 2013; Chen et al., 2014). Such efforts have been made for the detection of arsenic, a highly abundant and extremely toxic metalloid released to the environment as a result of anthropogenic activity (Matschullat, 2000; Mandal and Suzuki, 2002). In comparison to analytical chemistry methods, biosensors (which encompass a biological sensing component and easily detectable output) would provide reliable, specific, and an inexpensive means for the in situ detection of target compounds in environmental samples (Baeumner, 2003).

Constructed biosensors for environmental purposes have generally coupled well-characterized components (usually from bacteria) that are responsive to the target compound with a reporter gene that gives rise to a colorimetric, luminescent, or fluorescent output. In this sense, developed biosensors proved to be useful tools for arsenite and arsenate detection in groundwater (Siegfried et al., 2012) and river water (Siegfried et al., 2015) as a low cost, suitable and transportable alternative to detect the metalloid to prevent or diminish arsenic exposure. Furthermore, the advent of biological circuit design approaches in the field of synthetic biology has allowed the re-wiring of basic molecular components (regulators, DNA binding sites, and operators), which can be reinserted into the host cell to give rise to sensors with enhanced performance (Stocker et al., 2003; Trang et al., 2005; Fernandez et al., 2016; Merulla and van der Meer, 2016). In fact, the addition of ArsR operator downstream of its target promoter generates reduction in the background noise, which reduces the detection limits to as lower as one microgram per liter (Merulla and van der Meer, 2016).

Although changing the circuit design can improve the efficiency of the biosensor, the utilization of molecular components with intrinsically enhanced transcriptional performance when induced with the target compound could lead to a system with superior behavior. With this reasoning in mind, we aimed to identify a natural system that displayed enhanced transcriptional response to arsenite (As^III). We focused on the ars system from Chromobacterium violaceum, a gram-negative bacterium with a low arsenic tolerance level but endowed with a fully functional Pars and ArsR regulatory system (Azevedo et al., 2008; Silva-Rocha et al., 2013a). We demonstrate that the ars system from C. violaceum has superior arsenic induction performance when compared to the chromosomal prototype system in Escherichia coli and that these differences can be traced to the binding affinity of ArsR regulators to their DNA targets and to the occurrence of a stronger transcriptional response under inducing conditions in the C. violaceum system. Taken together, the results shown here demonstrate the potential of environmental bacteria as a reservoir of molecular components with enhanced performance for biosensor design, as well as a characterization of a novel arsenic ArsR transcription factor in bacteria.

Bacteria able to thrive in contaminated environments are endowed with very efficient detoxification mechanisms controlled at the gene expression level. In this context, different metalloregulatory proteins have evolved to control gene expression in response to different metals and metalloids. In vivo functional characterization of the ars system from several bacteria has revealed strong variation in the number of resistance genes, operon organization and, final resistance levels (Carlin et al., 1995; Paez-Espino et al., 2009). However, a conserved feature of the ars system is that the first gene of the operon encodes the ArsR regulator, which repress its own expression in a feedback loop (Wu and Rosen, 1993; Xu and Rosen, 1997). The ArsR from both E. coli and C. violaceum are members of the SmtB/ArsR family. This class of transcription factors include small proteins that share a core secondary structure formed by five alpha helixes and two anti-parallel beta strands in the form ααααββα (Cook et al., 1998; Busenlehner et al., 2003). Although DNA binding occurs through the helix-turn-helix domain formed by helix three and four, members of this family have a more degenerate inducer-binding site that could be located in helix three (Type 1 regulators – ArsR_eco) or in helix five at the dimerization interface of the protein [Type 2 regulators – ArsR_cvi (Qin et al., 2007)]. Considering the metal binding site, ArsR proteins have a conserved metalloid-protein interaction interface, in which cysteine residues are required to coordinate the ligand (Busenlehner et al., 2003). ArsR_eco has an As^III binding site formed by Cys residues located at helix three (Wu and Rosen, 1993), whereas ArsR_cvi belongs to the class of regulators where binding is formed by Cys residues at helix five (Qin et al., 2007; Azevedo et al., 2008). These differences between the two classes of regulators also imposes differences on the intramolecular allosteric switch upon inducer binding because type 1 regulators have the ligand-binding site close to the DNA binding domain, whereas signal transmission in type 2 regulators must occur from a distance (Qin et al., 2007).

While differences in the localization of metalloid biding site of the regulatory proteins might account for the different expression behaviors observed, other parameters such as protein-DNA interaction might play an important role in the process. For instance, Li et al. (2015) have used in vitro mutagenesis to tune the arsenic response of the ars system from the R773 plasmid of E. coli. By analyzing the variants with enhanced performance, they found several mutations into the ArsR binding site (ABS) at the DNA. This found is particularly relevant since, while members of the SmtB/ArsR family recognize palindromic DNA sequences, the ABS sequence found in the ars promoter of E. coli has an imperfect palindromic sequence (Wu and Rosen, 1993). As the ars systems from E. coli and C. violaceum have no conserved ABS sequence (Carepo et al., 2004; Azevedo et al., 2008), this difference on protein-DNA interface may explain the stronger binding affinity of ArsR_cvi and the difference in the dynamics of gene expression. Altogether, these structural differences between the two regulators analyzed here could explain the discrepant behavior (i.e., gradient vs. steeper) during expression of the ars systems.

While the ArsR from E. coli proved to be more sensible to lower arsenic concentrations when assayed in the decoupled system (Figures 4 and 5), the C. violaceum system displayed a better triggering system at higher concentrations, with a steeper slope promoter induction. Although the particular features of the two classes of regulators have not been systematically investigated, different molecular mechanisms for an allosteric switch induced by As^III binding could be the reason for the different induction profiles observed for ArsR regulators of C. violaceum and E. coli. Uncovering these mechanisms should be a target of future research in order to further understand the evolution of SmtB/ArsR protein family members and for biosensor development.

Microorganisms are a valuable source of molecular components for the construction of biological circuits for biotechnological applications (Cheng and Lu, 2012; Brophy and Voigt, 2014). From a historical perspective, most synthetic circuits constructed in bacteria have been implemented using components from the model organism E. coli (Church et al., 2014). This has indeed been the case for biosensors designed to detect arsenic in environmental samples with several designs having been constructed by shuffling the ars components of E. coli (Siegfried et al., 2012, 2015; Wang et al., 2015; Merulla and van der Meer, 2016). Since the World Health Organization (WHO) recommended acceptable limit for drinking water of 10 μg/L arsenic, any modification of natural arsenic sensing systems for biosensing purposes should aim increasing responsivity of the systems to concentrations close to this limit. The characterization of the ars components of an environmentally relevant bacterium, C. violaceum, shows that orthologous systems dedicated to the same function (in this case, arsenic detoxification from the cell) could have different dynamic properties. In the case of the ars system from C. violaceum, increased ArsR binding affinity by the target DNA sequence together with a stronger induction profile makes these components very attractive for novel designs of arsenic biosensing. Additionally, to the best of our knowledge, this is the first report of a naturally occurring arsenic response system in bacteria with such an atypical induction profile, which provides new venues for the investigation of ars evolution and for the application of this pathway in biosensor design.

RS-R and LA designed the experimental strategy. LA and LM performed the experiments. LA, LM, and RS-R analyzed and interpreted the data. LA and RS-R wrote the manuscript. All authors have read and approved the final version of the manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.