Team:Edinburgh OG/Biosensors

Biosensors

Given the enormous volumes of water and azo dyes used in the textile factories and released without a proper treatment, azo dyes spread readily in the environment. Monitoring of the dye pollution therefore plays a major role in their bioremediation. The analytical methods for the detection and quantitation of azo dyes and their degradation products traditionally utilize various spectroscopy, chromatography and mass spectroscopy techniques [1]. While these techniques typically offer highly sensitive and specific pollutant detection, often complex and time-consuming sample preparation together with specialized and expensive laboratory equipment required for the analysis make these methods not very well-suited for the pollutant analysis in situ.

An attractive alternative to the aforementioned physicochemical methods are biosensors. Biosensors are analytical devices combining a biological sensing element with a physicochemical transducer, converting a highly sensitive and specific biological response proportional to the target stimuli into an electrical signal [2]. Compared to the rather tedious and costly traditional physicochemical methods, biosensors might offer a rapid, highly sensitive and specific, portable and inexpensive platform for real-time in situ monitoring of textile waste pollution.

We decided to develop biosensors which would allow us to target three distinctive components of the textile waste effluents – i) directly azo dyes themselves, ii) heavy metals, which are very commonly present in textile wastewater, and iii) aromatic amines as the azo dye synthesis and degradation intermediates. Such biosensors would play an integral part in our system for both the monitoring of the textile waste, as well as the control of dye-degrading enzymes and silk protein expression.

Heavy Metal Sensor

Utilising one major characteristic of textile dye waste effluent – the presence of heavy metal contamination, largely by metals such as zinc, nickel and cobalt [1] – we constructed a biosensor genetic circuit in E. coli, using genes from S. elongatus and B. subtilis. In theory, this circuit could sense the presence of heavy metals in wastewater and produce laccase in response to break down dye. This would result in a completely self-contained device that could turn off and on without external inducers.

Design

The biosensor constructed in this project consists of two genes, smtB and cotA, and the cognate promoter of smtB, PsmtA. cotA, a laccase from B. subtilis, is used as the gene to be induced by the sensor.

The SmtB protein is a repressor; in the absence of any inhibitory molecule, SmtB will bind to and repress transcription from the PsmtA promoter. The sensor is constructed with two copies of PsmtA, one for each gene (Figure 1.), so that transcription from both smtB and cotA genes is controlled by the same inducer, giving the entire sensor binary ‘on’ and ‘off’ states. This was intended to reduce metabolic load and increase the sensor’s specificity. SmtB is usually produced constitutively, however, by reducing the overall amount of SmtB present in the ‘off’ state, it was thought that this would decrease any inappropriate repression when metals were present, while still allowing repression to be maintained. It was also thought that increasing the SmtB production when repression is relieved would shorten the time taken for the sensor to switch back to the ‘off’ state when metal was no longer present. This would, in theory, result in a much more reliable sensor.

Figure 1. Genetic map of biosensor construct. The biosensor consists of two copies of the PsmtA sequence (dark green) in opposing orientations. Each promoter controls a single gene – either smtB or cotA. The cotA sequence also contains a BioBrick ribosomal binding site (orange) and the PelB leader peptide (purple).

The cotA portion of the biosensor is designed so that the coding sequence after the PelB leader peptide can be easily removed by restriction digest and replaced with another coding sequence. In this case, eiraCFP, a gene encoding a cyan fluorescent protein, was intended to replace cotA to act as a control circuit. This would act as both a positive and negative control, both showing that the circuit is functional and that any changes seen in assays was due to laccase activity. This construct is shown in Figure 2. However, this construct suffered from a deletion that caused no fluorescence to be observed.

Figure 2. Genetic map of positive control construct. The overall layout is the same as in Figure 1 above, but eiraCFP replaces cotA as the sensor’s target gene.

While many inducible protein expression systems rely on the addition of an external inducer such as IPTG, there is significant interest in more flexible biosensors that use genetic logic to self-regulate expression of a target gene. The biosensor in this project has been constructed according to these principles. In particular, this project follows the design for a heavy-metal bioremediation device described by Tay, Nguyen and Joshi [2], which consists of the target gene, a constitutive repressor gene, and the repressor’s cognate promoter sequence. The original design and how it has been adapted in this project is shown in Figure 3.

Figure 3.Design of biosensor circuit. A) The underlying biosensor circuit design, from a mercury-bioremediating cell. B) Adaptation in this project, using a zinc-sensitive repressor (smtB) to produce laccase (cotA). Adapted from Tay, Nguyen and Joshi [2].

Experiments

PCR, Cloning, Transformation and PAGE were carried out according to standard protocols available in our protocols section. Cloning was done via digestion and ligation.

Bacterial Protein Expression Induction

To induce protein expression at high levels in both experimental and control constructs, E. coli BL21-DE3 cells containing the appropriate plasmid were grown overnight (37°C, shaking) in 10ml LB broth with antibiotic but without the appropriate inducer. 1µl of antibiotic was added per ml of LB broth. Kanamycin was prepared in water at a concentration of 50mg/ml, while Chloramphenicol was prepared in absolute ethanol at a concentration of 34mg/ml. 30µl of overnight culture was then transferred to 3ml fresh LB broth and grown until the OD reached 0.3-0.5, approximately 2 hrs. The inducer was then added, and cultures allowed to grow again for 16 hours. Experimental assays were carried out with these cultures.

The inducers used were heavy metals. Zinc sulphate heptahydrate (ZnSO4.7H2O), cobalt nitrate hexahydrate (Co(NO3)2.6H2O) or nickel chloride hexahydrate (NiCl2.6H2O) solutions were prepared with water. When inducing, these were added to LB broth to give final concentrations of 1, 4, 15, 50, 200 and 800µM.

Laccase Activity Assay

Laccase activity was measured using 2,2'-azino-bis(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) as a substrate. Cell lysate was prepared by centrifugation of induced cultures (1252 g, 5 minutes). The supernatant was discarded, and cell pellets were lysed with a 1% Triton (ThermoFisher) solution in PBS and diluted 1:10 with water. 100µl cell lysate was added to 400µl 200 mM sodium acetate buffer (pH 4.5) containing 1mM ABTS and 0.2mM CuSO4 and incubated for 20 hours (37°C, shaking). Reactions were centrifuged (1252 g, 5mins) to pellet lysate and the absorbance of the supernatant measured at 420nm in a FluoSTAR Omega microplate reader (BMG Labtech). Absorbance was compared to control values, and this was converted to enzyme activity units using the equation:

Enzyme activity (mU)=((Δ420 ×cf ×1000)/(ε ×df))/1200

Where Δ420 is the change in absorbance of the sample at 420nm, cf is the correction factor to give a path length of 1cm, ε is the millimolar extinction coefficient of ABTS (36 mM-1 cm-1) [3], and df is the dilution factor of the enzyme. As enzyme activity is defined per minute, with 1 U being the amount of enzyme required to oxidise 1µmol of substrate per minute, the results are divided by the total number of minutes – 1200 – to give the average activity per minute. Significance was assessed using an unpaired two-tailed t-test.

Results

Construction and Cloning

Due to the limitations of DNA synthesis in dealing with long repetitive stretches of DNA, the biosensor was synthesised containing only a single copy of the PsmtA sequence. It was not possible to synthesise both as they would form a stem-loop during synthesis. This version, containing only a single copy of the promoter, was amplified via PCR and inserted into a pET28aa-GG plasmid backbone containing mCherry via PstI and XbaI restriction sites at either end of the construct. The construct was designed with two restriction sites – for BamHI and KpnI – between the original copy of PsmtA and the cotA gene. Primers which would add these sites to the end of PsmtA were used to amplify the sequence. This would ensure that, upon digestion and ligation, the second copy of the regulatory sequence would be inserted into the plasmid in the correct orientation. Insertion of the second copy of PsmtA was confirmed via colony PCR and restriction digestion with ApaI and NheI.

A)B)

Figure 4. A)Colony PCR confirming insertion of construct into pET28aa-GG. Colony PCR using the T7 forward and reverse primers revealed two colonies with the expected band size of 2.6kb. Water was used as a negative control and the original, mCherry-containing plasmid as a positive control. Re-ligation of the PstI and XbaI ends can clearly be observed in the other sampled colonies. B) Colony PCR to confirm insertion of second promoter sequence. Colony PCR using the T7 forward and reverse primers to confirm the insertion of the second copy of PsmtA as compared to the positive control (+), which was the initial construct. As the difference is small (111bp), care was taken to only choose colonies which were of the expected size and clearly larger than the control. Colonies 15, 32 and 33 were deemed positive and proceeded to restriction digestion.

A)B)

Figure 5. A). Restriction digest confirmation of insertion. A positive from colony PCR was tested via restriction digestion with either ApaI, a double cutter, or AvaI, a multiple cutter. In ApaI, a single 5.9kb band is observed in the control, but two bands of 5.4kb and 2.5kb are seen in the sample as expected. For AvaI, the shift from 192bp in the original pET28aa-GG plasmid to 2.4kb in the construct-containing plasmid is clearly visible, confirming the insertion. B) Restriction digestion to confirm insertion of second promoter sequence. The three chosen colonies from colony PCR were digested with ApaI and NheI and compared to the original construct digested with the same enzymes. In both cases, bands of 2.5kb and 4.9kb are expected, but there should be a shift from 349bp to 452bp visible if both promoters are present. This band was observed in colony 32 and this was used in future experiments.

Zinc Induction

An ABTS assay using whole cells was carried out to determine if there was appreciable laccase activity from the biosensor in response to a variety of concentrations of zinc sulphate heptahydrate (ZnSO4.7H2O) ranging from 1 to 800µM. In the wild type S. elongatus, SmtB response to zinc has been observed at concentrations as low as 4µM [4], while the concentration of zinc in a typical textile effluent may be between 7µM (0.51 mg/l) [1] and 91µM (6mg/l) [5].

Figure 6. Induction of laccase in response to zinc. Laccase activity was tested at a variety of concentrations of zinc to look for induction. In all tested concentrations, laccase activity from the biosensor construct was significantly higher than in either cells constitutively expressing RFP, or cells containing the biosensor construct with CFP instead of CotA. As the concentration increases, laccase activity decreases in all 3 samples, probably due to cellular toxicity of zinc. (n=3, error bars = SEM)

Cross-Reactivity

Cobalt and nickel are both commonly found in textile wastewater [1,6] and are also known to be sensed by members of the SmtB-ArsR family [7,8]. In order to test if the sensor showed any cross-activation by these metals, it was induced in the same manner as zinc and both laccase activity assays and SDS-PAGE were performed.

Figure 7. Laccase activity in cells induced by cobalt and nickel. In both cases, there does not appear to be a clear relationship between laccase activity in the biosensor and concentration, either positive or negative. In A), activity in the cobalt-induced cells drops significantly from uninduced to induced before rising, and then falling again. At the 50µM concentration, activity is no better than the CFP control circuit. In B), the nickel-induced cells, although there appears to be a large spike, this has a very high error and so cannot be said to be significant. While the laccase cells are higher than the control cell types, there does not appear to be a significant level of induction with increasing concentration, as activity begins to fall after 4µM. (n=3, error bars = SEM)

Sequencing Issues

The second copy of the promoter appeared to be completely wrong in the sequence data. Nucleotide BLAST of the 95bp sequence which appears instead of PsmtA, shown in Figure 8, shows a 100% homology to the E. coli K12 genome sequence coding for a putative ATP binding protein (GenBank CP027060.1, Region: 3284213..3284307) [9].

Figure 8. Nucleotide BLAST of sequencing results. The BLAST search of the 95bp sequence found in the sequencing results instead of PsmtA as expected is shown as Query 1. It completely matches a portion of the E. coli genome sequence at positions 3284213-4284307, the coding sequence of an ATP binding protein. It appears a recombination event has occurred.

There also appears to be a frameshift mutation in the smtB sequence in this sequencing data, caused by a single base pair deletion, shown in Figure 9. In this case, it is unclear whether the frameshift is present in the original synthesised DNA or has been introduced at a later step. The frameshift observed causes 22 random amino acids to be introduced before it encounters a stop codon at position 89, truncating the protein by 33 amino acids (27%). However, as we observed a functional induction, it is unclear if this data is correct and the frameshift is actually present.

Figure 9. Alignment of smtB sequencing data to designed construct. The data from sequencing is shown on the top row, with the intended sequence below. Below this, the translation for each sequence is shown. The frameshift observed in the sequencing data can be seen in red at amino acid position 66, causing a run of random amino acids to be translated until a stop codon is encountered at position 89, compared to 122 amino acids in the correct sequence.

Overall we appear to have constructed a partially functional heavy metal sensor, but it is possible its potential would be far greater without the sequence issues.

References

[1]Bhardwaj V, Kumar P, Singhal G. Toxicity of Heavy Metals Pollutants in Textile Mills Effluents. Int J Sci Eng Res. 2014;5: 664–666. Available: https://www.ijser.org/researchpaper/Toxicity-of-Heavy-Metals-Pollutants-in-Textile-Mills-Effluents.pdf [2] Tay PKR, Nguyen PQ, Joshi NS. A Synthetic Circuit for Mercury Bioremediation Using Self-Assembling Functional Amyloids. ACS Synth Biol. American Chemical Society; 2017;6: 1841–1850. doi:10.1021/acssynbio.7b00137 [3]Kenzom T, Srivastava P, Mishra S. Structural insights into 2,2’-azino-Bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS)-mediated degradation of reactive blue 21 by engineered Cyathus bulleri Laccase and characterization of degradation products. Appl Environ Microbiol. 2014/09/26. American Society for Microbiology; 2014;80: 7484–7495. doi:10.1128/AEM.02665-14 [4] Morita EH, Wakamatsu M, Uegaki K, Yumoto N, Kyogoku Y, Hayashi H. Zinc Ions Inhibit the Protein–DNA Complex Formation between Cyanobacterial Transcription Factor SmtB and its Recognition DNA Sequences. Plant Cell Physiol. 2002;43: 1254–1258. doi:10.1093/pcp/pcf140 [5] Hussein F. Chemical Properties of Treated Textile Dyeing Wastewater. Asian Journal of Chemistry. 2013. doi:10.14233/ajchem.2013.15909A [6] Yaseen DA, Scholz M. Textile dye wastewater characteristics and constituents of synthetic effluents: a critical review. Int J Environ Sci Technol. 2019;16: 1193–1226. doi:10.1007/s13762-018-2130-z [7] Busenlehner LS, Pennella MA, Giedroc DP. The SmtB/ArsR family of metalloregulatory transcriptional repressors: structural insights into prokaryotic metal resistance. FEMS Microbiol Rev. 2003;27: 131–143. doi:10.1016/S0168-6445(03)00054-8 [8] Cavet JS, Meng W, Pennella MA, Appelhoff RJ, Giedroc DP, Robinson NJ. A Nickel-Cobalt-sensing ArsR-SmtB Family Repressor: CONTRIBUTIONS OF CYTOSOL AND EFFECTOR BINDING SITES TO METAL SELECTIVITY . J Biol Chem . 2002;277: 38441–38448. doi:10.1074/jbc.M207677200 [9]National Center for Biotechnology Information. Escherichia coli str. K-12 substr. MG1655 strain K-12 chromosome [Internet]. 2018 [cited 29 Jul 2019]. Available: https://www.ncbi.nlm.nih.gov/nucleotide/CP027060.1?report=genbank&log$=nuclalign&blast_rank=88&RID=KWCFA324014&from=3284213&to=3284307

Aromatic Amine Sensor

Aromatic amines are the main intermediates of azo dyes (and a whole range of other chemicals) synthesis and degradation. Due to their abundance, persistence and potential carcinogenicity, aromatic amines have been designated as high-priority pollutants hazardous to the environment and human health [1]. A biosensor for detection of aromatic amines could aid their and azo dyes’ monitoring and bioremediation. Although protein receptors to a variety of aromatic organic compounds have been described and used in various biosensor circuits [2], an aromatic amines-specific receptor or biosensor has not yet been described or developed. Nevertheless, one transcription factor – the BTEX (Benzene, Toluene, Ethylbenzene, Xylene) receptor XylR – and its mutant variants have been shown to exhibit an exceptionally broad substrate spectrum specificity for many aromatic compounds [3], including aminotoluenes - toluene-like aromatic amines [4]. This makes XylR a good candidate for evolving a biosensor for the detection of broader spectrum of aromatic amines and this project aimed to do so through the directed evolution of XylR’s receptor domain (Figure 1.).

Figure 1. Schematic organization of XylR domains and their functions. Effector-binding, receptor A-domain [5], but also B-domain to a certain extent [6], have been shown to be responsible for XylR’s substrate recognition and binding properties.

Design

We constructed XylR:GFP biosensor genetic circuit in E. coli. In theory, through rounds of random mutagenesis of XylR receptor domain via error-prone PCR and selection for mutants producing GFP fluorescence signal in response to selected aromatic amines (Figure 2.), it should be possible to develop a protein receptor, and at the same time a biosensor of aromatic amines.

Figure 2. Schematic workflow of directed evolution of XylR-based biosensor for detection of aromatic amines.

The biosensor genetic circuit comprised XylR transcription factor as the sensing module and superfolder GFP (sfGFP) as the reporter module. The biosensor circuit design followed the design previously developed and shown to reliably detect toluene by Ionis Paris iGEM team 2016 [7], submitted as part BBa_K2023015. In this design, the XylR sensing module (composite part BBa_K2023002) comprises a constitutive native Pr promoter driving the expression of the XylR gene. XylR protein binds toluene-like aromatic compounds and facilitates transcription of the sfGFP reporter gene from the Pu promoter within the reporter module (composite part BBa_K2023016 with the original, wild-type GFP changed here to sfGFP) (Figure 3.). In contrast to the Ionis Paris iGEM team’s design, codon-optimized sfGFP (BBa_K2541400) was used in this design as a less expensive, more convenient and comparably sensitive reporter alternative to the originally used bioluminescent GLuc (BBa_K2023009) reporter and also as an improved, brighter version of originally used GFP (BBa_E0040) reporter. The biosensor circuit was set within the same standard, high-copy pSB1C3 plasmid and transformed into TOP10 E. coli cells.

Figure 3. xylR:sfgfp biosensor genetic circuit design.

As this initial design did not work properly and resulted in constantly fluorescent cells, presumably due to the leaky expression of sfGFP reporter gene, another two iterations of the biosensor circuit were constructed in an attempt to lower the system leakiness. First, the orientation of the XylR sensing module was inverted in order to prevent transcriptional overreading of the sfGFP reporter gene from constitutive Pr promoter. Alternatively, the high-copy pSB1C3 plasmid backbone was exchanged for the low-copy pSB3C5 and pSB4C5 plasmid backbones, in theory reducing the amounts of XylR protein and reporter gene copies and thus the chances for unspecific sfGFP expression.

Experiments

Cloning

The sensing and reporter modules were synthesized individually. Through Q5 PCRs with adapter primers, modules were equipped with corresponding standard BioBrick prefix and suffix restriction sites, respectively, and with BsaI restriction sites at their interface, enabling the cloning of biosensor circuit into vector plasmid (Figure 4.). Ligated biosensor plasmids were transformed into TOP10 cells and correct constructs were confirmed by colony PCRs and plasmid digestion.

Figure 4. Scheme of XylR:sfGFP_pSB1C3 biosensor plasmid cloning.

Screening for Toluene-sensitive Cells

The colonies with wild-type XylR:sfGFP biosensor plasmids were initially visually screened and compared on LB agar plates without or with 10 mg/L toluene inducer under the blue light.

Troubleshooting

In an attempt to reduce the presumably responsible leaky expression of sfGFP reporter gene, we recloned the biosensor genetic circuit with XylR receptor module in the opposite direction from the sfGFP reporter module. We also recloned the biosensor genetic circuit into the low-copy number pSB3C5 and pSB4C5 plasmid backbones. Again, these biosensor variants were initially visually screened on LB agar plates without or with 10 mg/L toluene inducer under the blue light.

Fluorescence Assay

The accurate performance of the biosensor variants without or with 10 mg/L toluene was assessed by fluorescence assay. Due to the all cells carrying different biosensor circuit variants exhibiting high un-induced sfGFP fluorescence and being otherwise insensitive to induction by toluene - XylR’s natural substrate - we were unable to proceed with the directed evolution of XylR receptor domain.

Results

Cloning

We successfully assembled XylR:sfGFP_pSB1C3 biosensor plasmid and transformed it into TOP10 E. coli cells. This was confirmed by both colony PCR and plasmid digestion (Figure 5.A and B, respectively).

Figure 5. (A) Colony PCRs confirming the presence of the biosensor plasmid in 13 out of 15 tested clones. xylR-specific primers amplify 745bp long region. NC – water, PC – IDT xylR module. (B) Digestion reactions confirming correct biosensor plasmids assembly. Plasmids from six different colonies were digested with XbaI and PstI-HF (DC) or with PstI-HF alone (SC). Expected bands: DC – 3480 bp (xylR:sfgfp insert) and 2044 bp (pSB1C3 vector backbone), SC – 5524 bp (linearized biosensor plasmid). PC – digested rfp:pSB1C3 with expected bands: DC - 1069 bp (rfp insert) and 2044 bp (pSB1C3 vector backbone), SC - 3139bp (linearized plasmid).

Screening For Toluene-sensitive Cells

Colonies carrying XylR:sfGFP_pSB1C3 biosensor plasmid were either white or constitutively fluorescent and visually neither of these two phenotypes reacted to the induction by toluene (Figure 6.).

Figure 6. Screening of cells carrying biosensor plasmid for the induction by toluene. No visible differences in fluorescence could be observed between induced and uninduced cells.

The white, insensitive colony phenotype could be attributed to the unintended mutations in the sfGFP gene, which were indeed confirmed by sequencing and might potentially interfere with its proper function as the reporter. The constitutively fluorescent phenotype even in the absence of any inducer was most probably a result of sfGFP reporter leaky expression obscuring the signal from the actual biosensor induction. Such leaky overexpression of the reporter gene is highly undesirable in biosensor systems, as it creates a high basal signal background and impedes the sensitivity and detection range of the biosensor.

Troubleshooting

Three main factors might contribute to such high sfGFP reporter leaky expression: i) direct orientation of constitutively expressed XylR and following inducible sfGFP genes, ii) high-copy plasmid number due to the pSB1C3 plasmid backbone used, and iii) too bright codon-optimized sfGFP variant used as a reporter. In an attempt to lower the presumable system leakiness, biosensor circuit variants with reversed XylR sensing module orientation or low-copy plasmid backbone were constructed. However, similar colony phenotypes were observed and again neither of these two variants seemingly reacted to toluene inducer.

Fluorescence Assay

In order to accurately estimate the performance and response of the three biosensor variants to the toluene induction, we carried out fluorescence assay with two colonies of each variant (Figure 7.). The fluorescence signal was increasing steadily in a majority of biosensor variants but was apparently not dependent on the induction by toluene, as both induced and uninduced variants were producing fluorescence signal with similar rates. This implies that the observed increase in fluorescence was due to the higher cell culture density over time.

Figure 7. Biosensor variants inducibility by toluene. The fluorescence was normalized to cell culture ODs and to the negative control. RFU = Relative Fluorescence Unit. Biosensor variants are plotted separately and error-bars are omitted for better clarity. (A) The original xylR:sfgfp_pSB1C3 biosensor variants. (B) xylR_inv:sfgfp_pSB1C3 biosensor variants with opposite module orientation. (C) Low-copy xylR:sfgfp_pSB3C5/pSB4C5 biosensor variants.

Considering that the used toluene concentration generously exceeded XylR’s detection limit levels and still no distinct biosensor response was triggered, we concluded that none of the constructed biosensor variants were sensitive enough and worked properly. This prevented us from proceeding with the directed evolution of XylR protein. The downfall of this subproject was most likely the use of the too strong GFP variant as the reporter in our biosensor circuit design. For the similar future projects, the original biosensor design by Ionis Paris iGEM team 2016 [7] might be a better option.

References

[1] L. Pereira, P. K. Mondal, and M. Alves, “Aromatic Amines Sources, Environmental Impact and Remediation,” 2015, pp. 297–346. [2] R. Fernandez-López, R. Ruiz, F. de la Cruz, and G. Moncalián, “Transcription factor-based biosensors enlightened by the analyte,” Front. Microbiol., vol. 6, p. 648, Jul. 2015. [3] T. C. Galvão and V. de Lorenzo, “Transcriptional regulators à la carte: engineering new effector specificities in bacterial regulatory proteins,” Curr. Opin. Biotechnol., vol. 17, no. 1, pp. 34–42, Feb. 2006. [4] R. Salto et al., “Modulation of the Function of the Signal Receptor Domain of XylR, a Member of a Family of Prokaryotic Enhancer-Like Positive Regulators,” 1998. [5] D. Tropel and J. R. van der Meer, “Bacterial transcriptional regulators for degradation pathways of aromatic compounds.,” Microbiol. Mol. Biol. Rev., vol. 68, no. 3, pp. 474–500, table of contents, Sep. 2004. [6] J. Garmendia and V. de Lorenzo, “The role of the interdomain B linker in the activation of the XylR protein of Pseudomonas putida,” Mol. Microbiol., vol. 38, no. 2, pp. 401–410, Oct. 2000. [7] “Team:Ionis Paris - 2016.igem.org.” [Online]. Available: https://2016.igem.org/Team:Ionis_Paris.

Azo Dye Sensor

In order to regulate the expression of azo dye-degrading enzymes in dependence of fluctuating levels of dyes in the wastewater, a riboswitch-based biosensor could be engineered through a randomization and screening approach. Riboswitches are sequences in untranslated regions of mRNA that can exercise control on the expression of the adjacent gene through the formation of two mutually exclusive secondary structures. Riboswitches are commonly found in bacteria, often allowing the control of the expression of an enzyme through its own substrate. Switching between the two gene expression states is catalyzed through the binding of a specific ligand to the so-called ‘aptamer’. Depending on the activation or inhibition of gene expression upon ligand binding, riboswitches are categorized into the two groups of ON- and OFF-switches (Figure 1.)

Figure 1. Schematic of a common OFF-switch riboswitch function. Binding of a ligand (red) to the aptamer leads to rearrangement of secondary structures resulting in the blocking of the ribosomal binding site (RBS, blue) through binding of the anti-RBS sequence (green). Expression from start codon (AUG) is thus prohibited.

To generate aptamers binding novel target ligands, a method called ‘Systematic Evolution of Ligands through Exponential Enrichment (SELEX)’, employing iterations of selection steps that refine an initial random library of aptamer sequences to few sequences with high affinity to the target, has been developed by Tuerk and Gold (1990). However, SELEX procedures exceed the financial scope of an iGEM project and often result in aptamers that cannot be transferred into the native context in a functioning riboswitch. Therefore, this project aimed for the reengineering of a known, theophylline-dependent riboswitch to bind a representative azo dye – aniline yellow (AY) - as a proof of concept for the development of dye-dependent riboswitches.

Design

First, the aptamer region of the theophylline-dependent hammerhead ribozyme presented by Wieland and Hartig, 2006, would be randomized. The resulting library of riboswitches with random aptamers would be then transformed into E. coli cells and screened for the sequences binding AY in vivo, through the expression of adjacent RFP as a reporter (Figure 2.). In theory, this system should yield riboswitches with aptamer sequences able to bind aniline yellow, and potentially other azo dyes. Upon sequencing, another rounds of randomization and screening could be applied to improve specificity and sensitivity of the riboswitches. Such azo dye-dependent riboswitch could then be coupled with different reporter genes, serving as a direct detector of azo dyes, or with the dye-degrading enzyme genes for the precise control of their expression.

Figure 2. Schematic of the project’s concept. The library of aptamer region-randomized riboswitch sequences controlling RFP expression would be transformed into E. coli. Plating on agar plates containing aniline yellow allows selection of colonies with a riboswitch reacting to aniline yellow.

Conclusion and Future Directions

We successfully developed a heavy metal biosensor comprising SmtB heavy metal receptor and CotA laccase. Heavy metals, such as zinc, nickel and cobalt, are commonly found in textile waste effluents and can be used as indirect indicators of textile waste pollution.

Our heavy metal biosensor responded reliably to presented zinc in medium and produced the dye-degrading CotA laccase in turn. We tested the biosensor’s performance with both ABTS and dye-decolourization assays. Our experimental results show that the biosensor responds positively well to zinc levels even lower than those typically found in textile effluents. This means that this biosensor system has a great potential to be used in industrial setting and our envisioned system to drive the expression of laccase and other dye-degrading enzymes.

However, this heavy metal biosensor system did not work reliably when nickel and cobalt were targeted, and when the concentration of heavy metals – zinc including – exceeded certain levels, presumably due to the metals having a toxic effect on cells and/or impeding laccase activity. This might pose a potential problem, as textile waste effluents typically contain a mixture of heavy metals and of higher concentrations. Therefore, further testing in authentic textile wastewater samples would be required to assess the performance of this biosensor system in an industrial setting.

Currently, there is a tendency for textile industry to move away from the excessive use of heavy metals in the manufacturing process. Therefore, other biosensor systems targeting other components of textile waste might need to be developed in the future. While we have not been able to engineer biosensors targeting azo dyes and aromatic amines, our conceptual ideas and results will hopefully aid their future development.

Contact Us

Edinburgh OG
Peter Wilson Building
University of Edinburgh
edigemmsc@ed.ac.uk