Introduction

The detection and identification of bacterial species has been significant in a wide range of industries such as medicine, food technology/production, pharmacy, cosmetic, etc.1. Most of bacteria are nontoxic and beneficial for human health for instance, they can be employed for recovery of nutrients, nitrogen fixation, destruction processes, conversion of materials to energy and treat diseases. However, some of these bacteria are harmful and hence can be referred to as pathogenic bacteria2.

Having knowledge about population of bacteria in broth medium or a count of bacterial colonies in agar medium is significant for determining bacterial growth patterns and quantitative studies in different industries. Enumeration of the number of viable bacteria by formation of visible colonies on agar plates, known as colony formation, is a common method for determining bacterial viability3. On agar plates, the number of bacteria is often determined manually, which has some drawbacks such as low throughput, time-consuming, low accuracy, and it is tedious4,5. On the other hand, automatic numeration methods which are highly efficient, time-saving, with high resolution features and hence are often preferred as suitable alternatives to manual methods6. Although the resolution of major cell counting apparatuses like the hemocytometer1, coulter counter1,6,7, and flow cytometer8 can be affected by the cell population. Hence, by decreasing cell numbers to definite values, the accuracy of information is diminished9. Furthermore, these methods often need a purified population of target bacterium, often expensive equipment, and skilful operators. The alamar Blue is a designed assay to measure quantitatively the proliferation of different cell lines (human and animal), fungi, and bacteria, which has several benefits over usual cell counting assays10,11.

The colorimetric tetrazole 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay is based on the reduction of MTT to its respective formazan12. In this assay, purple dye appears which is then quantified using a photometer or a multiplate reader at the wavelength of 570 nm. Although this assay is often utilized to measure the proliferation and viability of mammalian cells, MTT and other relative materials can also be employed for measuring bacterial viability7. Another procedure to quantify bacterial growth is based on the measurement of the light scattering originating from the turbidity brought about by bacteria measured at 600 nm, referred to as optical density (OD600)13. This method is often used as a rapid and cost-effective technique to detect bacterial growth and throughout the culture in broth media. At a quick glance, however there are numerous methods for counting bacteria. Several investigations are being carried out to introduce highly selective, inexpensive, more accurate, and facile procedures. Among these, electrochemical sensing methods can be considered to overcome the limitations of culturing and biochemical procedures like sensitivity and specificity. Furthermore, they need lower technical knowledge than highly specific microbial or immunological methods.

MTB can easily be oriented towards the Earth’s magnetic field14. Magnetic and physical features of MTB make them valuable candidates in different fields including medicine like drug delivery15, treatment of infectious diseases, cancer therapy, and bioremediation15,16,17, cell separation18,19, enzyme immobilization20, biomineralization of metal ions21, waste water treatment, and for removing heavy metals22,23,24. However, although a wide variety of MTB’s applications have been reported so far, there are no reports available for using this type of bacteria in manufacturing a bacterial enumeration system.

The aim of our study is to present a highly selective, precise, and cost-effective method for enumeration of the MTB without requiring any time-consuming steps of bacterial purification. Our proposed method briefly employs an electrochemical (potentiometric) method, to detect the interaction of dissolved oxygen (DO) with the MTB. For measuring the DO parameter, iodometric titration is coupled with a potentiometric method. Based on the end point of titration (color of indicator and potential difference), the correlation of the magnetotactic bacteria population and potential differences are determined. Our results show that, the designed system can provide a simple and accurate chemical method for selective enumeration of magnetotactic bacteria without requiring their isolation from non-magnetotactic species.

Results and Discussion

Dual O2-H2 recognition biosensor based on the immobilized MTB

Potentiometry has proved in many studies to be a powerful tool for designing biosensors25,26 and evaluate the capacity of microorganisms for biosensing purposes27,28. In our study, oxygen was considered as an important factor for movement and orientation of the MTB towards the magnet-modified indicator electrode. The relationship between DO and the potential difference in presence of MTB is shown in Fig. 1A. This shows that there is a linear correlation between the potentiometric gradient (ΔE) and the DO concentration.

Figure 1
figure 1

(A) Relationship between DO (concentrations ranged between 4.0–20.0 mg L-1) and the potentiometric response of MTB during useage of a two electrode system. Experimental condition: 10.0 mL Fe2+/Fe3+ (5.0 mg L-1), as well as the same volume of MTB-containing suspension at pH = 9.5 using Tris-buffer solution (0.02 mol L-1). Condition: O2/H2 purging time: 5.0 min. with flow rate of 400.0 mL min.-1. Electrode system: Ag/AgCl (Sat’d Cl-) as the reference electrode and rod modified (with magnet) as the indicator electrode at room temperature. The data are the average of three replicate analyses. (B) Effect of hydrogen purging time on the potentiometric response (n = 3).

Full size image

Purging H2 gas for 5 min. increased the sensitivity of the electrochemical response. The effect of H2 gas was investigated by Thermogravimetric analysis (TG), which showed improvement of sensitivity. In TG analysis H2 absorbed /adsorbed by MTB indicates the main effect on performance of MTB. The maximum potentiometric response was obtained when DO concentration was set to 20.0 mg L-1, although salting-out effect has lowered this parameter in ecosystems29. Furthermore, studying the effect of different pH values demonestrated that the highest potentiometric sensitivity can be obtained at pH 9.0 prepared by Tris buffer (data not shown). Because of the presence of functional groups such as -OH and -NH2 in the structure of the Tris buffer as well as its buffer capacity (β), it behaved as the axillary complexing agent with cationic species such as Fe(II) and Fe(III). These cations are critical for the bacterial growth. They are mineralized inside the cell as Fe3O4. The optimum concentration of Fe(II) and Fe(III) to obtain a potentiometric response with maximum sensitivity was estimated to be 5.0 ppm.

Using direct potentiometry, the potentiometric response and DO concentrations showed a linear correlation, ranging from 4.0 to 20.0 ppm (Fig. 1A). Furthermore, in the optimum DO study, the effect of H2 gas on potentiometric response of the bacterium showed significant enhancement in different potentials during H2 purging. Figure 1B shows the increased sensitivity of the potentiometric response by H2 purging for 5.0 min. and the reversed effect by prolonging H2 purging time. This observation can be related to the decrease in the DO concentration of the electrolyte solution.

Cell enumeration by the immobilized MTB

The reliability of the O2 biosensing enabled us to introduce a selective MTB-based cell enumeration methodology using indirect potentiometric titration. In a typical experiment the titrant (Na2S2O3) is added to the analyte solution and the potential would begin to change until reaching titration endpoint (Fig. 2A). Two titration end points were detected belonging to free DO and MTB-interacted DO. However, only one end point was observed after deaeration and DO removal by N2 purging (Fig. 2B).

Figure 2
figure 2

Titration of (A) MTB and (B) N2-treated MTB with Na2S2O3 (0.02 N), H2 purging time: 5.0 min. with flow rate of 400.0 mL min.-1. Starch was used as indicator. Electrode condition: Ag/AgCl as a reference electrode and rod modified (with magnet) as the indicator electrode. Titrant was added with the speed of 0.1 mL per 30 s.

Full size image

Two independent endpoints were observed that correlated to i) the free DO in the electrolyte solution and ii) the DO adsorbed on the surface of MTB. Strong interaction of DO by MTB can probably lead to higher titrant consumption and subsequently observe two independent endpoints. Acceptable linear correlation was observed for the MTB population and the difference in potential (ΔE) and thus a calibration curve was obtained (Fig. 3A and Table 1).

Figure 3
figure 3

(A) Linear correlation between the average number of MTB and the potential difference during the potentiometric titration (n = 3). ΔE was defined as the potential gradient between Vend and Vinitial of the titration. (B) Correlation between zeta potential gradient and different populations of whole MTB cells (n = 3).

Full size image
Table 1 Correlation between the average population of MTB with Vend and ΔE in the potentiometric Winkler titration process.
Full size table

This method can be presented as a simple, selective, rapid, and reliable chemical technique to number MTB indirectly during following the correlation between potential difference and the MTB population. In the presence of (non-MTB) no interference was detected in the enumeration of MTB population (Table 2).

Table 2 Selectivity of MTB-based counting system in the presence of non-MTBa.
Full size table

Three independent experiments were carried out and estimated via extrapolation of the calibration curve. Average population of the MTB used as the reference sample was 92753 ± 25 in all the tested samples. This method displayed a good selectivity towards MTB so that, after adding an excess population of different non-MTB (vs. the control sample), no statistical difference was detected. The presence of magnetosomes analog with MTB flagella will enable them to swim, orient and subsequently making mass transfer towards a special magnetic pole when they are subjected to an external magnetic field. This intrinsic feature will differentiate MTB mainly from non-magnetotactic bacteria and even from the cell debris of dead magnetotactic bacteria. This property makes them potential therapeutic carriers30. In addition, the affinity of the isolated MTB for adsorbing or absorbing O2 and H2 can be considered as one of the other distinguishable characteristics of these bacteria from the non-magnetotactic ones. In our study, these features were considered as the key factors for reliable MTB enumeration by the method used. Accordingly, the designed enumeration system will be able to differentiate magnetotactic bacteria from other non-magnetotactic ones. The reliability of this method was evaluated via comparison data obtained by calibration curve and number of MTB in the real sample. Maximum and minimum percentages of relative errors range from 0.15 to 10.53% (Table 3), which is acceptable for potentiometric methods.

Table 3 Estimation of MTB number in five synthetic MTB-based samples.
Full size table

It has been reported that the oxygen charge density is remarkably enhanced after coordination with other ion species26. Therefore, the charge density of the MTB was significantly changed after adsorption of O2. It is expected that, physicochmeical absorption/adsorption on the colloidal surfaces with high active surface area may provide extra charge density31. A fixed charge (Δq) was considered during the interaction between MTB and O2. In thermodynamics32, Δq is divided in two parts including static and dynamic contributions. The static part of Δqstatic can be reported as dqstatic = idt. The potentiometric technique is operated at open circuit conditions; therefore the share of this kind of electrical charge was negligible and was experimentally equal to zero32. Whereas about the dynamic part of Δq (i.e. Δqdynamic), this term is individually correlated to the capacitance behavior of MTB. This finding can be reported as dV = dqdynamic/dC, which is correlated to the surface charge (i.e. Zeta potential) of the microorganism33. In other words, reaction between titrant and O2 interacting with MTB changed the ratio of dqdynamic/dC, which is detected via their mass transfer (swimming) towards the indicator electrode. Eventually, these results were employed to measure the charge density (zeta potential) at open circuit conditions. To confirm this, the zeta potential of fresh MTB-containing solution was measured and compared to those estimated after reaching the endpoint of titration. As a result, a linear correlation was observed between MTB population and zeta potential difference (Fig. 3B). It was concluded that, the zeta potential difference was statistically affected by increasing the population of MTB and subsequently higher adsorption of O2.

Conclusions

In the current study, a specified procedure for enumoration of magnetotactic bacteria was introduced by combining potentiometry and iodometric titration. Potentiometry is widely used as a powerful tool to assess the bio-sensing behavior of microorganisms. Using potentiometry, DO was detected as an affecting element on bacteria movement towards the magnet-modified indicator electrode and sunsequent potential difference so that a two or three-fold increase of DO concentration resulted in an enhanced accumulation of bacteria onto the elecrtode’s surface which was quantified by a two or three fold increase in potential difference. A significant increase of zeta potential difference was detected by enhancing MTB population. Furthermore, according to the TG analyses, a five minute H2-purching increased the potential difference to near two-fold displaying the effect of H2 absorption or adsorption behavior of bacteria on their movements and attachment onto the electrode’s surface. It has been concluded that, the introduced method provides a seletcive, fast, cost-effective, precise, and reliable tool for enumoration of magetotactic bacteria without the necessity of their isolation from other non-MTB.

Methods

Reagents and Materials

All reagents with purities more than 99% including starch, H2SO4 (96%) and also different salts including FeCl3.6H2O, FeCl2, Na2S2O3.5H2O, MnSO4.H2O, KI, NaOH, NaN3, trisaminomethane, different gasses (purity ≥ 99%), such as N2 and O2 were provided by gas cylinders and purchased from Linde (Germany). Reference electrode (Ag/AgCl/sat’d Cl-) and counter electrode (iron rode, diameter: 2.0 ± 0.1 mm) were all purchased from Metrohm (Switzerland).

Design of the dissolved oxygen (DO) biosensor

The electromotive force (Eemf) of the whole cell was measured using a potentiometer (Lutron Electronics Co Inc. 1986-01-22, Germany) and a two-electrode system including an iron rod (diameter: 2.0 ± 0.1 mm) as the indicator electrode and Ag/AgCl (Sat’d Cl-) as reference electrode. The indicator electrode was modified with a circular magnet (16.5 Tesla) with internal diameter, external diameter, and height of 1.5, 6.0, and 2.0 mm, respectively. High impendance noise was reduced significantly by empolying a thermal varnish around the indicator electrode. The electrolyte solution included 0.078 mM FeCl2 and 0.038 mM FeCl3.6H2O, Tris buffer (10.0 mM, optimum pH = 9.5), and the MTB solution (CFU: 3.2×107, 10.0 mL). The volume ratio of Fe(II), Fe(III), and bacterial population was optimized to 1:1:2. Effects of H2 and O2 on potentiometric response of the bacteria were investigated by purging these gases independently into the electrolyte solution. Hydrogen was provided using a H2 generator (OPGU-1500 S, Korea) with a flow rate of 440.0 mL min.-1 for 5.0 min. Different concentrations of oxygen from 4.0 to 20.0 mg L-1 were purged into the electrolyte solution and then the resulting DO concentration was recorded using a DO meter (Lutron-5510, Lutron Electronics Co Inc. 1986-01-22, Germany). Concentration of H2 and O2 was optimized based on the one-at-a time method. To record the direct potentiometric response of the MTB on the indicator electrode, the ΔE5min was monitored using the Visual Basic 6.0 (VB6) program.

Construction of cell enumeration system

The obtained data from direct potentiometry was further confiremed by using the iodometry titration as an indirect method24. The ΔE of indirect potentiometry was calculated in the titration endpoint. In this method, the DO concentration is measured indirectly through an estimation of the end point of the potentiometric titrimetry. In this titration, 3.64 g MnSO4.H2O, 0.50 g NaOH, 0.10 g NaN3, 0.15 g KI and 1 mL H2SO4 (96%) were used as reagents, which were added to the analyte solutions (Fe(II) 0.078 mM and Fe(III) 0.038 mM and the MTB solution 1:1:2, V/V, in Tris buffer; 10.0 mM, optimum pH = 9.5), Na2S2O3 (0.02 M) and starch (average molecular weight: 342.30 g.mol-1) used as titrant and color indicator. The interfering effect of Fe(II) in titration was removed using sodium azide according to the established procedure34. Indirect potentiometry was carried out based the same procedure of direct method except for the addition of iodometric titration reagents, agitation and titration of the solution by Na2S2O3 until the end point. Furthermore, the selectivity of the presented method for exclusive enumeration of magnetic bacteria cells was investigated by using non-magnetic bacteria such as the gram-negative bacteria like Escherichia coli (ATCC 10536), Salmonella enterica subsp. enterica (ATCC 13076), Pseudomonas aeruginosa (ATCC 9027), the gram-positive bacteria Staphylococcus epidermidis (ATCC 12228), Streptococcus vestibularis (ATCC 49124), and Lactobacillus amylovorus (ATCC 33620) and the probiotic bacteria Lactobacillus acidophilus (KCTC 3164) and Lactobacillus gasseri (ATCC 33323) (all provided by the National Center for Genetic and Biological Reserves, Tehran, Iran). All these bacterial strains were tested as negative controls.

The net charge of the MTB surface is negative and can be balanced by opposite ions presented in the surrounding media35. Zeta potential has been considered as a useful technique to investigate the activity and mechanism of microorganisms35,36. In the current study, to determine the charge of the MTB cell surface in the presence of oxygen (100.0 parts per million, ppm), zeta potential was measured directly and indirectly. Direct method was carried out using a zeta potential analyzer (Svarovsky, Zeta Meter Inc., USA) via a two-microelectrode stainless steel plate (type: 316, 1.0×1.0 cm) with a distance of 2 cm. The capacitance current change was monitored using a galvanometer (United Scientific MGV002 DC Galvanometer) to reach the end point using a magnet (16.5 Tesla) and adding different populations of MTB and a digital multimeter (Fluke 177 TRMS digital multimeter, UK). The capacitance current was calculated by measuring the static and dynamic currents of the RC circuit (R = 100.00 ± 0.01 Ω and C = 5.00 ± 0.01 μF) using the multimeter (Fluke 177).