Introduction

Cadmium (Cd) is a serious environmental toxicant with no known biological function. It is classified by the International Agency for Research on Cancer (IARC) as a group I carcinogen for humans and the seventh hazardous heavy metal listed by the Agency of Toxic Substances and Disease Registry [1, 2]. The worldwide recognition of toxic effects from minute concentrations of Cd has resulted in regulations to reduce its presence in the environment to very low levels. In support of this initiative, there is a clear need for reliable, efficient and cost-effective monitoring technologies for the presence of cadmium in the environment.

A biosensor is an analytical device for the detection of an analyte that combines a biological component with a physicochemical detector component. The major advantages of using biological components in sensors are their good specificity, sensitivity, and portability. Moreover, unlike chemical or physical analyses, biological systems do not require large and expensive instruments [3]. The analysis of heavy metal ions can be carried out with biosensors using both protein-based and whole-cell-based approaches [4]. In general, microbial biosensors comprise the molecular fusion of two linked genetic elements: a sensing bioelement and a reporter gene. In most cases, the sensing element is a promoter that specifically responds to the presence or absence of the target molecule, and the reporter gene, which is fused to the sensing element, encodes a quantifiable molecule [5].

Many whole-cell microbial biosensors for detection of heavy metals, including Cd, arsenate, mercury, and copper using recombinant DNA technology, have been reported and are considered promising applications in the fields of biotechnology and environmental sciences [3, 4]. Until now, only a few whole-cell microbial biosensors for Cd detection have been developed using genetically engineered microorganisms [3, 4, 6–9]. These biosensors carry a recombinant plasmid containing lacZ, luxCDABE, or gfp fused with the promoter of target genes to be activated by Cd. The β-galactosidase (lacZ) and luminescence (luxCDABE) assays require an expensive substrate and detection device. And the fluorescence (gfp) bioassay also requires an elaborate fluorescence detector. However, using natural pigments as the biosensor output may overcome these issues [10].

In this study, we used a unique red pigment (deinoxanthin) producing bacterium, Deinococcus radiodurans, for the sensor strain, and the crtI gene, which is responsible for the carotenoid synthesis, as a reporter gene, to construct a simple and cost-effective whole-cell biosensor independent from the specific reagents and instruments for Cd detection. A sensor strain consisting of a crtI-deleted white host strain of D. radiodurans, in which the crtI gene was reintroduced downstream of Cd-inducible promoter, was constructed to detect Cd based on the conversion of the carotenoid colors.

Materials and methods

Strain, growth conditions, and medium

The bacterial strains used are listed in Table 1. D. radiodurans strains were grown in TGY (0.5% tryptone, 0.3% yeast extract, and 0.1% glucose) broth or on TGY plate supplemented with agar (1.5%). Escherichia coli strains were grown in LB (1% tryptone, 0.5% yeast extract, 0.5% NaCl) broth or on LB plate at 37 °C with aeration. Where required, antibiotics were used at the following concentrations: ampicillin (100 μg/ml) or kanamycin (50 μg/ml) for E. coli; kanamycin (8 μg/ml) or chloramphenicol (3 μg/ml) for D. radiodurans.

Table 1 Bacterial strains and plasmids used in this study
Full size table

Enzymes and chemicals

All chemicals including heavy metals (CdCl2, CrCl3, PbCl2, NiCl2, ZnCl2, FeCl2, CuCl2, AsCl3, MnCl2, and HgCl2) were purchased from Sigma-Aldrich (St Louis, MO, USA). Stock solutions of each metal were prepared in distilled water. Restriction enzymes and reagents for PCR were purchased from Fermentas (Hanover, MD, USA) and Takara (Otsu, Shiga, Japan), respectively.

Construction of β-galactosidase (lacZ) and phytoene dehydrogenase (crtI) reporter cassettes

Chromosomal DNA was isolated from D. radiodurans as described by Udupa et al. [11]. The putative promoters of DR_0070 (267 bp), DR_0659 (592 bp), DR_0745 (270 bp), and DR_2626 (318 bp) were PCR amplified from the genomic DNA of D. radiodurans R1 using the each primer pairs (Table 2). The PCR products were digested with BglII and SpeI and ligated with the BglII-SpeI digested larger fragment of pRADZ3 to replace the putative groESL promoter. This procedure yielded pRADZ-P0070, pRADZ-P0659, pRADZ-P0745, and pRADZ-P2626, respectively. A set of deleted DR_0659 promoter fragments was amplified from the chromosomal DNA using each primer pairs (Table 2), digested with BglII and SpeI, and exchanged with the DR_0659 promoter fragment of pRADZ-P0659, generating pRADZ-P0659-1, pRADZ-P0659-2, and pRADZ-P0659-3, respectively.

Table 2 Primers used in this study
Full size table

To use crtI as a reporter gene, crtI was amplified from the chromosomal DNA using crtI-F/R primer pair (Table 2), digested with SpeI and XbaI, and cloned into the lacZ removed pRADZ3 fragment by digestion with SpeI and XbaI, generating pRADI. The promoter fragment of pRADZ-P0659-1 was obtained by digestion with BglII and SpeI, exchanged with the putative groESL promoter of pRADI, and the recombinant plasmid designated as pRADI-P0659-1.

Assay for expression of the DR_0659::lacZ reporter gene

A single colony of D. radiodurans strains was suspended in 5 ml of TGY broth and grown as a pre-culture in a shaker (200 rpm) at 30 °C for 24 h. For β-galactosidase activity measurements, exponentially growing recombinant D. radiodurans strains harboring pRADZ3 [13] derivatives were treated with metals to the indicated concentrations and time intervals. After exposure, cells (1 ml) were harvested and β-galactosidase activities of the pellets were measured as described by Bonacossa de Almeida et al. and Sommer et al. [14, 15].

Construction of crtI mutant

A fusion PCR product (3,187 bp) for crtI deletion was constructed in two steps. In the first step, two different asymmetric PCRs were used to generate fragments to the left (1,629 bp) and right (1,604 bp) of crtI sequence using crtI-DD-UF/UR and crtI-DD-DF/DR primer pairs (Table 2), respectively. In the second step, the left and right fragments were annealed at their overlapping region and amplified by PCR as a single fragment using the outer primers (crtI-DD-UF/DR). Specifically, 1 μl of each of the two asymmetric PCR mixtures and 500 μM of each of the two outside primers were mixed together and PCR amplified. The fusion product was purified (QIAGEN, Valencia, CA, USA) and cloned into a pGEM®-T easy vector (Promega, Madison, WI, USA). The resulting plasmid was designated pT-ΔcrtI. A 975-bp HincII fragment (kan cassette) from pKatAPH3 [12] was ligated to SmaI-digested pT-ΔcrtI and the resulting plasmid was designated pT-ΔcrtI::kan. A strain containing a ΔcrtI::kan mutation was constructed by transformation of the wild-type D. radiodurans strain R1 [11] with PCR product (4,162 bp) amplified from plasmid pT-ΔcrtI::kan (Table 1) using the crtI-DD-UF/DR outer primers (Table 2). Deletion of crtI was confirmed by the loss of red pigment and PCR using crtI-Dia-F/R primers (Table 2).

Results and discussion

Evaluation of the specificity of the putative promoters to Cd treatment

From previous microarray results [16], we selected putative promoters of the four Cd-inducible genes (DR_0070, DR_0659, DR_0745, and DR_2626) for further confirmation by β-galactosidase (lacZ) reporter gene assay. To examine the specificity of promoters derived from these Cd-responsive genes, putative promoter region was fused with lacZ gene. The resultant reporter cassettes, pRADZ-P0070, pRADZ-P0659, pRADZ-P0745, and pRADZ-P2626, were introduced into D. radiodurans and analyzed for their responses to metals (Fig. 1).

Fig. 1
figure 1

Analysis of Cd-inducible activities of DR_0070, DR_0659, DR_0745, and DR_2626 promoters. a Response of the promoters to several metal ions. Recombinant D. radiodurans cells, KDH041, KDH045, KDH049, and KDH055, were grown to an OD600 of 1.0, and the metals were treated at 5 μM. The cells were incubated for an hour and β-galactosidase activity was measured. b Analysis of Cd concentration-dependent activities of the promoters. The exponentially growing recombinant cells were treated with 0.1, 1, or 5 μM of Cd. After Cd treatment for an hour, β-galactosidase activity was measured. c Time course analysis of the promoter activities. The exponentially growing recombinant cells were treated with 5 μM of Cd, harvested after 15, 30, 60, and 120 min of postincubation, and β-galactosidase activity was measured. These results are a representative of three independent experiments

Full size image

The β-galactosidase activity of all the recombinant D. radiodurans was highly specific to Cd when compared with other metals (Fig. 1a). However, the recombinant strain KDH055 carrying pRADZ-P0659 showed the highest increase in β-galactosidase activity upon Cd exposure when compared with other recombinant strains. Next, we examined Cd dose and exposure time-dependent strength of the promoters. Although β-galactosidase activity of KDH041 carrying pRADZ-P0745 showed proportional increase to Cd concentration, KDH055 had the highest sensitivity among the recombinant strains (Fig. 1b). Expression pattern of lacZ upon Cd exposure time was similar to each other, except for KDH045 carrying pRADZ-P2626 (Fig. 1c). In conclusion, the putative promoter of DR_0659 was the most suitable response element for Cd biosensor development. Therefore, we analyzed the putative promoter region of DR_0659 and used it for colorimetric Cd biosensing.

Promoter analysis of DR_0659

Smaller-sized promoters are advantageous in the construction of expression cassettes with various reporter genes. Thus, we roughly identified the minimal promoter fragment required for the Cd-inducible response by simple stepwise deletion analysis of the DR_0659 promoter (Fig. 2). The 393 bp (P0659-1), 236 bp (P0659-2), and 131 bp (P0659-3) fragments of the putative DR_0659 promoter were fused with the lacZ gene to construct the expression vectors pRADZ-P0659-1, pRADZ-P0659-2, and pRADZ-P0659-3, respectively (Fig. 2a). Recombinant D. radiodurans strains, designated KDH055 and KDH084 to KDH086, were analyzed for their responses to 5 μM Cd (Fig. 2b). Although leaky LacZ expression was observed in pRADZ-P0659-3, the LacZ expression pattern of pRADZ-P0659, pRADZ-P0659-1, and pRADZ-P0659-3 in response to Cd was similar to each other. However, derepression of DR_0659 promoter activity occurred in truncated P0659-2 construct. One possible explanation for this behavior of pRADZ-P0659-2 and others can be provided by regulation mechanisms of metal resistance. Bacterial metal resistance systems are regulated by transcriptional factors from the MerR family [17]. It is likely that a major component of the activation mechanism, as in Mer, will involve the torsional distortion of the operator/promoter region to create an improved substrate for RNA polymerase action [18]. Some Mer members have broad specificity and have been reported to react with more than one type of metal ions; e.g., CueR reacts with Cu [I], Ag [I], and Au [I], whereas ZntR is mainly regulated by Zn [II], but also responds to Cd [II] and Pb [II] [17]. However, DR_0659 was highly specific to Cd, which indicates that it might be tightly regulated, other than Mer family. Reportedly, DNA sequences upstream of the rrnB P1 core promoter (−10, −35 region) increase transcription more than 300-fold in vivo and in vitro [19]. This stimulation results from a cis-acting DNA sequence, the UP element, which interacts directly with the alpha subunit of RNA polymerase and from a positively acting transcription factor, FIS. Likewise, upstream region (−237 to −394) of DR_0659 might be required for its transcriptional regulation by intrinsic DNA curvature.

Fig. 2
figure 2

Promoter analysis of the putative D. radiodurans DR_0659 promoter. a Stepwise deletion of D. radiodurans DR_0659 promoter. P0659, 592-bp promoter fragment; P0659-1, deleted 393-bp promoter fragment; P0659-2, deleted 236-bp promoter fragment; P0659-3, deleted 131-bp promoter fragment. b LacZ expression of recombinant D. radiodurans harboring each pRADZ-P0659 derivatives. The exponentially growing recombinant cells were treated with 5 μM of Cd, harvested after an hour of post-incubation, and β-galactosidase activity was measured. These results are a representative of three independent experiments. c Detection ranges of the P0659-1 promoter to Cd. The exponentially growing recombinant cell harboring pRADZ-P0659-1 was treated with various concentrations of Cd for an hour and β-galactosidase activity was measured. These results are a representative of three independent experiments

Full size image

We further analyzed Cd detection limit of KDH084 (Fig. 2c). The LacZ expression of KDH084 was logarithmically increased up to 200 nM of Cd, and the expression was not significantly changed at the concentration from 100 nM to 100 μM of Cd. However, at higher concentrations, there was a sharp decrease in the LacZ expression that might be attributed to the toxicity of Cd. The detectable minimum Cd concentration was 10 nM. A previously reported, Cd sensor using recombinant Bacillus subtilis has the lowest detection limit of 3.3 nM [20], until now. However, luminescent intensity decreased sharply at higher Cd concentration due to cadmium toxicity and it also responded to lead. The lowest detection limit of Cd biosensors developed using E. coli and Staphylococcus aureus was 10 nM, but they also responded to other metals and expression of reporter gene was highly affected by Cd concentration [20, 21]. Recently, a Cd sensing system using Pseudomonas putida has been reported [7]. The sensor showed high sensitivity to Cd and could stably detect relatively high concentration of Cd because of inherent cadmium resistance in P. putida. However, this system required further steps for stable measurement and showed relatively low specificity to Cd when compared with our sensor.

Detection of Cd by red color development

Various biological systems have been used to construct whole-cell bacterial biosensors [3–5]. Generally, genetically engineered sensor bacteria contain a recombinant plasmid that carries target chemical responsive elements and/or a reporter gene(s), such as firefly (luc) or bacterial luciferase (lux), green fluorescent protein (gfp), and β-galactosidase (lacZ). These systems require special agent or substrates or expensive equipment for target chemical detection [22]. To overcome these drawbacks, bacterial whole-cell biosensors using natural pigment as the biosensor output have been reported. A red pigment biosynthetic gene (crtA) of Rhodovulum sulfidophilum was used for arsenite and dimethyl sulfide detection [23, 24]. A blue-green pigment synthesis of Pseudomonas aeruginosa was used for N-butyryl homoserine lactone detection [22]. A whole-cell arsenite biosensor using red pigment synthetic gene (crtI) of Rhodopseudomonas palustris was developed [25]. In this study, kanamycin resistance gene, aph, replaced crtI of D. radiodurans to use the red pigment (deinoxanthin) as a Cd detection marker by the naked eye. CrtI is a phytoene dehydrogenase converting colorless phytoene to reddish lycopene. Therefore, crtI deletion mutant (KDH018) was selected by its colony color and further confirmed by PCR (Fig. 3).

Fig. 3
figure 3

Verification of D. radiodurans crtI gene deletion by PCR. PCRs were done with primers crtI-Dia-F and crtI-Dia-R locating outside from the mutant cassette using genomic DNA isolated from either wild-type (R1) or ΔcrtI::kan r mutant (KDH018) cells. The PCR product of R1 revealed a band length of 4,983 bp, whereas that of KDH018 resulted in a 4,316 bp. M denotes molecular standards

Full size image

Sensor strain (KDH081) was created by transforming KDH018 with the sensor plasmid, pRADI-P0659-1. To evaluate Cd specificity of KDH081, exponentially growing cells were treated with various metals. The color of the culture treated with Cd changed to red indicating that the KDH081 responds to Cd by changing its color (Fig. 4a). However, the color of other cultures treated with metals other than Cd did not change. This result clearly indicates the Cd specificity of the sensor strain reflecting the LacZ reporter assay results.

Fig. 4
figure 4

Color development of the sensor strain. a Change of color in the Cd sensor strain in response to various metals. The exponentially growing sensor strain, KDH081, was treated with 5 μM of each metal. A photograph was taken 24 h after the addition of metals. b Color change of the sensor strain in response to various concentrations of Cd. The exponentially growing sensor strain was treated with various concentrations of Cd. The photograph was taken 24 h after the addition of Cd

Full size image

To examine the macroscopic detection limit of KDH081, the strain was exposed to different concentrations of Cd ranging from 1 nM to 1 mM (Fig. 4b). The bacterial color change could not be recognized macroscopically at the concentrations of 1 and 10 nM. Slight color change was recognized at the concentration of 50 nM. Clear color change could be detected in concentrations from 100 nM to 200 μM. However, dose-dependent increase of chromaticity could not be differentiated macroscopically. Although reduction of chromaticity occurred when the sensor strain was exposed to Cd at concentrations of more than 500 μM, the highest macroscopic detectable concentration was 1 mM. When compared with results from KDH084 employing LacZ reporter assay for Cd detection, KDH081 showed relatively narrow Cd detection range (50 nM–1 mM) that might be caused from the color of the medium. However, the results clearly demonstrated that KDH081 could specifically indicate the presence of Cd by red color development enabling macroscopic monitoring of Cd in samples.

Conclusions

Very few attempts have been made to construct colorimetric whole-cell biosensors using pigment synthetic genes. To our knowledge, a colorimetric whole-cell biosensor for Cd, in which the signal is generated based on the conversion of intrinsic pigments of a host strain, has not been developed elsewhere. In this study, a pigment-based colorimetric bacterial biosensor has been successfully developed for easy and rapid macroscopic detection of Cd. A highly sensitive and specific promoter region was screened using LacZ reporter assay. Using β-galactosidase activity, Cd could be detected at nanomolar levels. The red pigment synthetic gene, crtI, was deleted from the host genome as an output signal. When exposed to Cd, the genetically engineered sensor strain harboring the recombinant plasmid containing the promoter fused to crtI changing color from light yellow to red with high specificity. The detection range of the sensor strain was from 50 nM to 1 mM. This new colorimetric whole-cell biosensing system provides a simple, stable, and cost-effective for Cd detection. Other biosensors using this pigment-based vector and host system could be constructed by inserting a promoter region that responds to a specific substance. Thus, this colorimetric system could be applied in a variety of fields.