**A Sensitive DNA Enzyme-Based Fluorescent Assay for Bacterial Detection**

#### **Sergio D. Aguirre, M. Monsur Ali, Bruno J. Salena and Yingfu Li**

**Abstract:** Bacterial detection plays an important role in protecting public health and safety, and thus, substantial research efforts have been directed at developing bacterial sensing methods that are sensitive, specific, inexpensive, and easy to use. We have recently reported a novel "mix-and-read" assay where a fluorogenic DNAzyme probe was used to detect model bacterium *E. coli*. In this work, we carried out a series of optimization experiments in order to improve the performance of this assay. The optimized assay can achieve a detection limit of 1000 colony-forming units (CFU) without a culturing step and is able to detect 1 CFU following as short as 4 h of bacterial culturing in a growth medium. Overall, our effort has led to the development of a highly sensitive and easy-to-use fluorescent bacterial detection assay that employs a catalytic DNA.

Reprinted from *Biomolecules*. Cite as: Aguirre, S.D.; Ali, M.M.; Salena, B.J.; Li, Y. A Sensitive DNA Enzyme-Based Fluorescent Assay for Bacterial Detection. *Biomolecules* **2013**, *3*, 563-577.

#### **1. Introduction**

Infectious agents, such as foodborne pathogens, have caused numerous large-scale and costly outbreaks in the human history and will continue to be a major public health threat and financial burden for our society [1–4]. Early detection of pathogens, as the first step to prevent such outbreaks, has become increasingly more important today because the globalization of commerce and speedy travel have significantly increased the rate and breadth of the spread of infectious agents. Thus, the demand for faster, simpler, less expensive and more reliable pathogen testing methods has become ever greater.

Although the traditional culture method continues to be the "gold standard" for bacterial detection, it is time-consuming and requires days or even weeks to complete (depending on the specific pathogen in question) [5]. Modern methods take advantage of well-established biomolecular techniques, such as polymerase chain reaction (PCR) and immunoassay (where an antibody is used as molecular recognition element), to achieve faster and more sensitive pathogen detection [5–11]. Despite the popularity of these techniques, they also come with certain drawbacks, such as the need for costly instrumentation and highly trained personnel to isolate or purify relevant targets (DNA for PCR and proteins for immunoassays). Thus, the entire test using such methods often still needs one or more days to complete. Detection sensitivity (for immunoassay) and tendency to generate false-positive results (for PCR) are also issues of concerns. For these considerations, we recently began to examine the utility of RNA-cleaving fluorogenic DNAzyme (RFD) probes for bacterial detection [12–14]. RFDs can be isolated from random-sequence DNA pools to perform three linked functions: ligand binding, catalysis and fluorescence generation. Each RFD cleaves a synthetic nucleic acid substrate containing a single ribonucleotide as the cleavage site embedded in a DNA sequence, and the

cleavage site is located between two nucleotides modified with a matching pair of fluorophore and quencher [12–21]. Because of these two features, these reporter molecules emit an increasing level of fluorescence when they carry out the catalytic cleavage of the RNA linkage. In other words, the cleavage event results in separation of the fluorophore from the quencher, accompanied by the increase of fluorescence intensity in real time.

More recently, we developed a method of isolating novel DNAzyme probes against the crude extracellular mixture (CEM) left behind by a specific type of bacteria in their environment or in the media they are cultured [12]. The CEM is rich in diverse targets, including small molecules and proteins. Thus the use of the crude mixture as the complex target to conduct *in vitro* selection [22–24] experiment circumvents the tedious process of purifying and identifying a suitable target from the microbe of interest for biosensor development, and provides a subsequent assaying procedure that is simple because it does not require steps to purify a target of interest. Using this approach, we have isolated an RFD that cleaves its substrate only in the presence of the CEM produced by *E. coli* (CEM-EC) [12]. This *E. coli*-sensing RFD, named RFD-EC1, was found to be highly selective to CEM-EC but nonresponsive to CEMs from many other Gram-negative and Gram-positive bacteria. We have also shown that the DNAzyme-based assay is capable of reporting the presence of a single *E. coli* cell after 12 h of culturing. These experiments have illustrated the utility of RFDs as fluorogenic bacterial indicators. In this work we carried out a thorough investigation to characterize this bacterial detection system with a goal to further improve the detection sensitivity.

#### **2. Results and Discussion**

#### *2.1. Establishing a* Trans*-Acting DNAzyme*

Our previously reported RFD-EC1 is a *cis*-acting DNAzyme that cleaves a covalently attached substrate. However, a *trans*-acting DNAzyme where the DNAzyme cleaves a detached substrate has an additional advantage such as ease-of-synthesis, thus lowering the cost and labor. Synthesis of long DNA chain modified with fluorophore, quencher and ribonucleotide is associated with lower yields and higher costs. Therefore, in this study, we first examined the possibility of converting it into a *trans*-acting catalyst by detaching the substrate portion of the sequence, FS1, from the DNAzyme portion, EC1 (Figure 1A). We found that EC1 was indeed able to cleave FS1 in *trans*, even at 1:1 ratio (50 nM each of EC1 and FS1), in a CEM-EC dependent manner (Figure 1B). Note that the reaction mixtures were analyzed by denaturing polyacrylamide gel electrophoresis (dPAGE).

We next tested a second *trans* construct, named EC1T (Figure 1A), by truncating 28 nucleotides from the two ends of EC1 (italic letters, Figure 1A) that were used as the primer-binding sites for polymerase chain reaction during the original *in vitro* selection experiment. Interestingly, EC1T was found to be considerably more active than EC1 (comparing Lanes 3 and 5, Figure 1B; *i.e.*, 45% *vs*. 72%). As a control, we also tested a mutant sequence, EC1TM, with 10 nucleotides (lower-case letters, Figure 1A) mutated from EC1T. These mutations rendered EC1TM completely inactive in the presence of CEM-EC (Figure 1B).

**Figure 1.** Design of *trans*-acting DNAzymes. (**A**) The sequences of EC1, EC1T, EC1TM and FS1. EC1 is the full length DNAzyme including two primer binding sites (nucleotides in italic) for polymerase chain reaction used in the original *in vitro* selection experiment. EC1T is the shortened version of EC1 with deleted primer binding sites. EC1TM is a mutant of EC1T wherein the nucleotides shown as lower-case letters are altered. The substrate FS1 contains an adenosine ribonucleotide (R) flanked by a fluorescein-dT (F) and a DABCYL-dT (Q); (**B**) dPAGE analysis of the cleavage reaction mixtures of FS1 with EC1, EC1T, or EC1TM in the absence (í) and presence (+) of CEM-EC. P1 represents the 5ƍ-cleavage product, which can be observed by fluorescence scan as it contains the F unit. MK (marker) is a sample of FS1 fully cleaved by NaOH. Clv% for each sample was calculated following our previously reported method [20].

*2.2. Comparing DNAzyme Activity Using Crude Extracellular Mixture (CEM) and Crude Intracellular Mixture (CIM) of E. coli* 

The original DNAzyme RFD-EC1 was isolated to cleave in the presence of CEM of *E. coli*. We hypothesized that the target that activates the DNAzyme might be more abundant inside the cellular environment. To test this idea, we made an *E. coli* culture and used it to prepare the CEM-EC and CIM-EC as follows: the cells were precipitated by centrifugation and the supernatant was taken as the CEM-EC. The cell pellet was re-suspended in the reaction buffer, heat-treated, and then centrifuged; the remaining supernatant was taken as the CIM-EC (see experimental section for details). The CEM-EC and CIM-EC were then used to induce the cleavage activity of EC1T towards FS1, and the results are illustrated in Figure 2A. It is clear that the CIM-EC indeed contained a much higher amount of the target than the CEM-EC as it induced much stronger cleavage of FS1 by EC1T (45% *vs.* 1%). Note that much lower cleavage in this experiment with CEM-EC is due to the shorter culture time (7 h) with low number of *E. coli* cells (50,000 colony forming units). For the remaining experiments, the CIM-EC was used as the target of interest.

#### *2.3. Searching for an Optimal Culture Broth*

We next investigated the effect of bacterial growth media on the quality of CIM (as measured by the cleavage activity of EC1T/FS1) in order to establish an optimal culture broth. Seven common growth media were chosen for this analysis and they were: Luria Bertani (LB), Terrific Broth (TB), Todd-Hewitt (TH), Lysogeny Broth Miller (LBM), Tryptic Soy Broth (TSB), Super Optimal Broth (SOB) and Super Optimal Broth with Catabolic repressor (SOC). 250 *E. coli* cells were allowed to grow in 1 mL of each broth for 7 h at 37 °C, from which CIM was prepared and used to induce

the cleavage of EC1T/FS1; the results are illustrated as Figure 2B. The CIMs from SOB and SOC produced the highest activity (~26% cleavage), followed by those from LB, LBM, and THB (10%–16%). The CEMs from TSB and TB were least effective (5%). Based on these results, SOB was chosen as the broth for the remaining experiments.

**Figure 2.** Cleavage reactions of EC1T/FS1 with (**A**) crude extracellular mixture (CEM)-EC and CIM-EC and (**B**) crude intracellular mixture (CIM)-EC collected from *E. coli* cells grown in various culture broths. NC is a negative control where the reaction was conducted in the absence of CEM-EC and CIM-EC. Each reaction mixture was analyzed by 10% dPAGE, followed by fluorimaging. NC: negative control where the reaction was conducted in RB without CEM-EC or CIM-EC.


#### *2.4. Effects of Divalent Metal Ions*

Divalent metal ions play crucial roles in catalytic functions of DNAzymes and it has been shown that different metal ions can significantly affect the catalytic activity of a DNAzyme [25–28]. For example, 8–17, a well-studied RNA-cleaving DNAzyme, exhibits the highest activity in presence of lead ions even though it was originally isolated using Mg2+ [29] or Zn2+ [30]. A recent study has revealed that Pb2+ promotes the most favorable folding of 8–17 [31]. Therefore, we sought to compare the effects of various divalent metal ions on the activity of our *E. coli*-sensing DNAzyme although the original DNAzyme RFD-EC1 was obtained by *in vitro* selection in the presence of 15 mM MgCl2 [12]. Nine different divalent metal ions were tested and they were: Ba2+, Cd2+, Co2+, Mg2+, Mn2+, Ni2+, Cu2+, Zn2+, and Ca2+; the results are given in Figure 3A. We found that Ba2+, Ca2+, Mg2+ and Mn2+ all induced a robust cleavage activity of the DNAzyme (causing 56%–68% of cleavage). In contrast, Cd2+, Co2+, Ni2+, Cu2+, and Zn2+ resulted in weak cleavage (1%–2%). It is possible that Ba2+ Mg2+, Mn2+ and Ca2+ fit into the catalytic core better than the other divalent metal ions. However, this should be experimentally verified.

It is noteworthy that we have previously shown that Mn2+ exhibits potent fluorescence quenching effect, resulting in significantly reduced signal magnitude when the fluorescence intensity is measured in a fluorimeter [32]. We also found a similar effect of Mn2+ in our assay (data not shown). In contrast, Ba2+ produced no quenching effect. This observation indicates that Ba2+ is a more suitable divalent metal ion for our assay. Thus, Ba2+ was chosen for further experiments. In order to establish the optimal Ba2+ concentration we investigated the effect of Ba2+ concentration on EC1T's activity. The data presented in Figure 3B indicates that the catalytic activity of EC1T reaches a plateau at 15 mM Ba2+.

**Figure 3.** (**A**) Cleavage activity of EC1T/FS1 in the presence of CEM-EC and various divalent metal ions; (**B**) Effect of the Ba2+ concentration.

#### *2.5. Varying Reaction Temperature*

We examined the cleavage activity of EC1T/FS1 at different temperatures and the results are provided in Figure 4A. A robust cleavage activity was observed at both 15 and 23 °C. In contrast, reduced activity was observed when the reaction temperature was decreased to 4 °C or increased to 37 °C and 50 °C. Interestingly, although CIM was absolutely required to induce the cleavage at 4, 15 and 23 °C, EC1T can cleave FS1 in the absence of CIM at both 37 °C and 50 °C (grey bars in Figure 4A). Since room temperature is the most ideal condition to conduct assays avoiding the requirements of heating and cooling system, we chose 23 °C as the reaction temperature for the remaining experiments.

**Figure 4.** Cleavage activity of EC1T/FS1 with varying temperature (**A**); pH (**B**); and EC1T/FS1 ratio (**C**). The data are the average of two independent experiments.

*2.6. pH Effect* 

We next examined the activity of EC1T/FS1 when the reaction pH was varied between 5.0 and 9.0; the results are shown in Figure 4B. Although EC1T was able to cleave FS1 in the entire pH range tested, the highest activity was observed at pH 7.5–8.0. Since the original DNAzyme was isolated at pH 7.5, it is not surprising that EC1T exhibits such a narrow pH preference.

#### *2.7. DNAzyme/Substrate Ratio*

We also examined the cleavage activity at different ratios of EC1T/FS1. For this experiment, the concentration of FS1 was kept at 50 nM while the DNAzyme concentration was changed from 0 to 5 M; the results are shown in Figure 4C. The cleavage activity reached the plateau at a ratio of 50:1. Thus, this ratio was used for the remaining experiments.

#### *2.8. Specificity*

With the significant changes of the reaction conditions, we wondered if EC1T was still able to maintain its specificity for *E. coli*. Four other gram-negative bacteria and four gram-positive bacteria were arbitrarily chosen for comparison. Each bacterium was cultured in SOB for a different period of time until the OD600 (optical density at 600 nm) of each culture reached ~1. The CIM was then prepared and tested with EC1T/FS1 under the optimal reaction buffer (50 mM HEPES, pH 7.5, 150 mM NaCl and 15 mM BaCl2, room temperature, EC1T/FS1 = 50/1). None of the CIMs from other bacteria was able to induce cleavage (Figure 5), indicating that EC1T/FS1 retained the specificity for *E. coli*.

**Figure 5.** Specificity of EC1T/FS1 for various gram-negative and gram-positive bacteria. PP: *Pseudomonas peli*, YR: *Yersinia rukeri*, HA: *Hafnea alvei*, AX: *Achromobacter xylosoxidans*, EC: *Escherichia coli*, BS: *Bacillus subtilis*, LM: *Leuconostoc mesenteroides*, LP: *Lactobacillus planturum*, PA: *Pediococcus acidilactici*.


#### *2.9. Detection Sensitivity*

To test the detection sensitivity of EC1T/FS1, we prepared a series of *E. coli* stock solutions from which CIM samples were prepared as described in experimental section. These samples were then assessed for inducing the cleavage of EC1T/FS1 under the optimal reaction condition established above. These reactions were monitored in a fluorimeter in real time for 60 min (Figure 6A). The reaction mixtures were also analyzed by dPAGE (Figure 6B). We found that the fluorimeter method was able to detect 10<sup>5</sup> cells while the dPAGE method can detect 10<sup>4</sup> cells.

We also tested the detection sensitivity of the original *cis*-acting DNAzyme RFD-EC1 using the optimal reaction condition. Interestingly, RFD-EC1 showed better sensitivities: the fluorimeter method can detect 104 cells (Figure 6C) while the dPAGE method was able to detect 103 cells (Figure 6D).

**Figure 6.** Sensitivity test. (**A**) Real-time fluorescence monitoring and (**B**) dPAGE analysis of EC1T/FS1 in the presence of CIMs prepared from 103 –10<sup>7</sup> *E. coli* cells. (**C**) and (**D**) Similar experiments using RNA-cleaving fluorogenic DNAzyme (RFD-EC1) with CIMs prepared from 10<sup>2</sup> –107 *E. coli* cells. The data in (A) and (C) are the average of two independent experiments.

#### *2.10. Detection of a Single Cell via Culturing*

Finally we determined the time required to enrich a single live bacterium (*i.e.*, one colony forming unit or 1 CFU) via culturing in SOB. Following a previous protocol [12], we inoculated a single *E. coli* cell in SOB and cultured for 2, 4, 6, 8 and 10 h at 37 °C. CIMs were prepared for the samples collected at each time point and tested with both *trans* and *cis* constructs. These samples were then assessed for inducing the cleavage of EC1T/FS1 under the optimal reaction condition. Each reaction was examined both in a fluorimeter (Figure 7A) and by dPAGE (Figure 7B). Using EC1T/FS1, 8 h of culturing was sufficient for detection by the fluorimeter method (Figure 7A) and 6 h by dPAGE method (Figure 7B). Using RFD-EC1, however, only 6 h and 4 h of culturing were required to achieve the detection by the fluorimeter (Figure 7C) and dPAGE (Figure 7D) method, respectively. The lower activity of EC1T/FS1 in comparison to the *cis*-acting RFD-EC1 might be due to the weakened interaction between enzyme and substrate strands when they were separated from each other.

#### **3. Experimental Section**

#### *3.1. Synthesis and Purification of Oligonucleotides*

The standard DNA oligonucleotides (EC1, EC1T, EC1TM and EC1LT) were purchased from Integrated DNA Technologies (IDT, Coralville, IA, USA) and purified by 10% denaturing polyacrylamide gel electrophoresis (dPAGE). The modified oligonucleotide FS1 was acquired from W. M. Keck Oligonucleotide Synthesis Facilities (Yale University, New Haven, CT, USA), deprotected and purified by 10% dPAGE following a previously reported protocol [15].

**Figure 7.** Culturing time required to detect a single *E. coli* cell (1 CFU). (**A**) Monitoring fluorescence of EC1T/FS1 with CIMs prepared from samples taken after a culturing time of 2, 4, 6, 8 and 10 h; (**B**) dPAGE analysis of the reaction mixtures in (A). (**C**) and (**D**) are equivalent experiments in which RFD-EC1 was used to replace EC1T/FS1. The data in (A) and (C) are the average of two independent experiments.

#### *3.2. Enzymes and Chemical Reagents*

T4 DNA ligase and T4 polynucleotide kinase (PNK) were purchased from MBI Fermentas (Burlington, ON, Canada). Tryptone and yeast extract was acquired from BD Biosciences (Mississauga, ON, Canada). All other chemical reagents were purchased from Sigma-Aldrich (Oakville, ON, Canada) and were used without further purification.

#### *3.3. Growth Media*

Luria Bertani (LB), Terrific Broth (TB), and Todd-Hewitt (TH) were purchased from Sigma-Aldrich. Lysogeny Broth Miller (LBM) was obtained from EMD Canada (Mississauga, ON, Canada). Tryptic Soy Broth (TSB) was acquired from BD Biosciences. Super Optimal Broth (SOB) and Super Optimal Broth with Catabolic repressor (SOC) were made in house. SOB contains 2% (w/v) tryptone, 0.5% (w/v) yeast extract, 10 mM NaCl and 2.5 mM KCl. SOC has the same ingredients as SOB but also contains 20 mM glucose and 10 mM MgCl2.

#### *3.4. Preparation of* Cis*-Acting RFD-EC1*

RFD-EC1 was generated by template-mediated ligation of FS1 to EC1. In brief, 200 pmol of FS1 were treated with 1× PNK buffer A (MBI Fermentas), 1 mM ATP and 20 U (units) of PNK for 30 min at 37 °C (reaction volume = 50 ȝL). The reaction was quenched by heating at 90 °C for 5 min. Equimolar EC1 and EC1LT (5ƍ-CTAGG AAGAG TCGGA CGGAG CTG; the ligation template) were then added to this solution and was heated at 90 °C for 30 s and cooled to room temperature for 10 min. Afterwards, 10 ȝL of 10× T4 DNA ligase buffer (MBI Fermentas), 39 ȝL of deionized distilled water (ddH2O) and 1 ȝL of T4 DNA ligase (10 U/ȝL) were added. After incubation at room temperature (RT) for 2 h, the ligated EC1-FS1 was purified by 10% dPAGE.

#### *3.5. Bacterial Cells*

Gram-negative bacteria *Pseudomonas peli*, *Yersinia rukeri*, *Hafnea alvei*, and *Achromobacter xylosoxidans* were donated by Dr. Gerard Wright (Micheal G. DeGroote Institute for Infectious Disease Research, McMaster University). Gram-positive bacteria *Leuconostoc mesenteroides*, *Lactobacillus planturum* and *Pediococcus acidilactici* (PA) were gifts from Dr. Brian Coombes and Dr. Russel Bishop (Department of Biochemistry and Biomedical Sciences, McMaster University). *E. coli* K12 (MG1655) and *Bacillus subtilis* 168 are regularly maintained in our laboratory.

#### *3.6. Comparison of the Cleavage Activity of EC1, EC1T and EC1TM in the Presence of CEM-EC*

*E. coli* was plated onto a TSB agar (1.5%) plate and grown for 14 h at 37 °C. A single colony was taken and inoculated into 2 mL of TSB and grown for 14 h at 37 °C with shaking at 250 rpm. A 1% fresh culture was made by re-inoculating 20 L of the above culture into 2 mL of TSB. The re-inoculation was allowed to grow at 37 °C with shaking at 250 rpm until the culture reached an OD600 of ~1. 1 mL of this culture was centrifuged at 11,000 g for 5 min at room temperature; the supernatant was taken as the crude extracellular mixture (CEM-EC) and stored at í20 °C.

For each candidate DNAzyme construct, two reactions were set up, a control and a test. For the test, 25 ȝL of 2× reaction buffer (2× RB; 100 mM HEPES, 300 mM NaCl, 30 mM MgCl2, pH 7.5) was mixed with 23 ȝL of the CEM-EC prepared above, 1 ȝL of 2.5 ȝM FS1 and 1 ȝL of 2.5 ȝM EC1, EC1T or EC1TM. For the control, TSB was used to substitute the CEM-EC. Each reaction mixture was incubated at RT for 60 min, followed by quenching with 5 ȝL of 3 M NaOAc (pH 5.5) and 135 ȝL of cold ethanol. DNA was recovered by centrifugation and analyzed by 10% dPAGE. DNA bands in the gel were visualized by Typhoon 9200 (GE Healthcare) and quantified by ImageQuant software (Molecular Dynamics).

#### *3.7. Comparison of the Cleavage Activity of EC1T in the Presence of CEM-EC and CIM-EC*

100 ȝL of 50,000 CFU/mL glycerol stock of *E. coli* was inoculated into 2 mL of TSB and grown at 37 °C for 7 h with shaking at 250 rpm. 1 mL of this culture was centrifuged at 11,000 g for 5 min at room temperature; the supernatant was taken as the CEM-EC for this experiment. The cell pellet was suspended in 200 L of 1× RB and heated at 50 °C for 15 min. The heat-treated cell suspension was then centrifuged at 11,000 g for 5 min at RT. The clear supernatant was taken as the CIM-EC for the experiment.

The cleavage reaction with the CEM-EC was carried out by mixing 25 ȝL of 2× RB, 23 ȝL of the CEM-EC prepared above, 1 ȝL of 2.5 ȝM FS1 and 1 ȝL of 2.5 ȝM EC1T. The reaction concerning the CIM-EC was conducted by mixing 41 ȝL of 1× RB, 5 ȝL of CIM-EC, 1 ȝL of 2.5 ȝM FS1, 1 ȝL of 2.5 ȝM EC1T and 2 ȝL of 2× RB (note that the CIM-EC was made by suspending the cell pellet from originally 1 mL of *E. coli* culture in 200 ȝL of 1× RB, which translates into a concentrating factor of 5). A control experiment without the CEM-EC and CIM-EC was also conducted. Each reaction mixture was incubated at RT for 60 min, followed by 10% dPAGE analysis as described above.
