Isoconductivity method to study adhesion of yeast cells to gold electrode

The dielectric behavior of biological suspensions has been widely studied. Nonlinear dielectric spectroscopy of microbiological suspensions was firstly proposed by the group of Woodward and Kell in 1990. They reported a biological effect under an applied sinusoidal electric field of less than 5 Vcm^-1 [1, 2, 3]. Such a nonlinear system should generate harmonics in the current polarisation that can be measured by Fourier analysis. Based on this fact, a nonlinear spectrometer was designed to evaluate some biological samples such as Saccharomyces cerevisiae [1]. However, the spectrometer could not prevent nonlinearity due to electrode–electrolyte interface (EEI) [4, 5]. Others groups also failed to eliminate the nonlinearity of the EEI, despite the use of high technology systems. [6, 7, 8]. Fortunately, in our previous study, we performed measurements with different cells and obtained relevant results with the three-electrode cell [9]. Our results indicated that the changes in the third harmonic were due to the presence of microorganisms near the EEI [9], where they were consistent with Blake-Coleman's results [10]. In all of these studies, it was not indicated whether the changes in the amplitude of the third harmonic originated in changes in the double-layer capacity or in charge transfer resistance. From an electrochemical point of view, it is not clear what is changing in the electrode-electrolyte interface.

Interestingly, it has also been shown that the attachment of microorganisms to the EEI could produce two effects, both of which reduce the double-layer capacity [11, 12, 13, 14]. The first and more important effect is the reduction of the effective surface of the electrode, similar to a bubble curtain on the electrode [15]. The second effect is the change in the microstructure of the double layer capacity. Muñoz-Berbel did not model the double layer capacity; instead, the confusing constant phase element (CPE) was used to fit the data.

Muñoz-Berbel describes a new method based on impedance spectroscopy to detect bacterial attachment. Attachment to the electrode surface usually occurs in two stages [16]. The first stage, called reversible attachment, is a fast initial adsorption phase governed by physical forces such as electrostatic and van der Waals forces. It is probable that the bacteria require the use of appendages called flagella to overcome these physical forces [17, 18, 19]. The second stage, called irreversible attachment, is a slow phase in which the bacteria adhere tightly to the electrode surface via pili, binding proteins and polysaccharides for the conditioning layer [20]. Measurements of the magnitude of CPE versus time for the following cell suspensions are presented: total samples containing both cells and released substances, samples only containing released substances, samples containing only cells and samples of secreted substance after filtration. In this work, the changes in the CPE magnitude were due to, not only yeast attachment but also to changes in conductivity. When the results of samples after filtration and total sample without filtration were compared, between 50% and 70% of changes in CPE were due to the suspending medium [21].

The adherence of eukaryotic cells to planar gold film electrodes has also been studied by other groups [22, 23, 24, 25, 26, 27]. Nor do these publications mention having taken into account the effect of changes in conductivity of the medium on the impedance measurements.

The present study had two objectives. The first one was to propose a new method to detect and monitor the attachment of yeast cells to the electrodes but avoiding changes in the conductivity of the suspensions. The second one was the electrochemical characterization of the yeast attachment using the proposed method. We have shown that the impedance measurements of EEI are sensitive to both attachment processes and changes in the conductivity of the suspensions. The proposed method is based on keeping constant the medium conductivity to avoid impedance changes due to suspension conductivity changes. To confirm that the changes in impedance measurements were only due to yeast attachment, we have done three things: the first was to show experimentally that the lower the difference between the supernatant and base solution conductivities, the lower the change in impedance. And so in the limit, when the supernatant and base solution conductivities are equal within the experimental error, impedance will not change. The second was to show experimentally that when the cell suspension and base solution conductivities are equal within the experimental error, it was observed an important impedance change. Finally the last one thing we did was to take SEM images of the naked electrode and after the measurements in which supernatant and cell suspension was added. Finally, a physical model explaining the observed behavior was proposed.

2.1

Electrochemical cell

A three-electrode cell is composed of a working electrode (WE) and a Ag/AgCl reference electrode (Re1). An AISI 304 stainless steel concave counter-electrode (CE) 85 mm in diameter was also used (Fig. 1). The WE was a solid cylinder (1.5 cm long) with only 1 cm² of its transversal section exposed and coated with gold; the rest was insulated with Grilon. The CE was made with a larger area than the working electrode to minimize its impedance.

Fig.1

2.4

Microbiological preparation

Different media were used in the assay:

Base Solution (BS): 20 mM of KH2PO4, 30 mM of KCl, 1 mM of MgCl2, to pH 6.5, where mM = mmol/L.

Yeast cell Suspension (YCS).

The YCS (109CFU/ml) was obtained by dissolving 9.33 g of dry yeast in 100 ml of BS and stirring for 5 min with a magnetic stirrer. Then, 65ml this solution was centrifuged for 10 min at 5000 rpm. The pellet was resuspended in 65 ml of BS and magnetically stirred. This process was repeated five times to ensure the complete removal of the biological material ions that may contain commercial yeast (Levex®) and to ensure that the conductivity of final suspension of washed yeasts (YW) had the same conductivity as BS (5.47±0.04) mS/cm. The YCS was then serially diluted thrice down to 10⁷ CFU/ml in decade steps. The centrifuge and resuspend process was repeated 4 and 2 times for samples of 10⁸ and 10⁷ CFU/ml respectively (Fig. 2). The conductivities were measured with an OAKTON® portable conductivity meter. All supernatants obtained after centrifugation were reserved. The concentrations were measured in colony forming units per milliliter (CFU ml⁻¹) by using the plate count method. Before measurements, biological samples were stored in the fridge at 4°C to slow growth. All of the manipulations were performed under sterile conditions.

Fig.2

The Argand diagram in Fig. 3 (complex plane) shows, as an example, the frequency sweeps with BS, YW (10⁹ CFU/ml) and the circular fitting. The horizontal and vertical axes show Z′(ω) and –Z″(ω), where Z′(ω) and Z″(ω) are the real and imaginary parts of the complex impedance, respectively; Z*(ω) = Z′(ω) + jZ″(ω). Z'(ω) at high frequencies tends to R_m (78 Ω), representing the small electrolytic resistance between the reference electrode and the working electrode. This may be appreciated in the left inset of Fig. 3.

Fig.3

Fig. 3 shows that the loci for YW and BS coincide, but there is a shift of the points corresponding to the same frequency. This behavior is observed for the three concentrations and in all situations. In this case, the point corresponding to the impedance at a frequency of 1 Hz is indicated with black and red arrows. This may be appreciated in the right inset of Fig. 3.

Fitted values of R_ct, before (R_ctb) and after (R_cta) addition of different media and the statistical analysis performed, are shown in Table 2.

The statistical analysis was performed with SPSS software v.17.

Fig. 4 shows the temporal evolution of C_p% for different media and concentrations. C_p values were obtained from the complex impedance using the equation (1).

Fig.4

Red arrows indicate the time in which the frequency sweeps like those shown in Fig. 3 were performed. In the first 10 min of measurement, the electrolytic solution was BS. The black arrow indicates when supernatant or YW was added. The error bars are not shown in Fig. 4 in order to clarify this figure. However were added in legends.

Table 3 shows the measured conductivities of media used and its impedance changes expressed as maximum C_p% (C_p values for 5 and 30 min in Fig. 4).

Fig. 5 and 6 were obtained from data of Table 3.

Fig.5

Fig.6

First, Fig. 3 shows that, while the loci for BS and YW coincide, there is a shift of the points corresponding to the same frequency. This shift occurs when there is a variation of the C_dl values but not R_ct values.

If the EEI is modeled electrically with a simple circuit consisting of an R_m in series with the parallel combination of R_ct and C_dl, the locus in the complex plane and the frequency dependencies of the real and imaginary parts are given by equations (2), (3) and (4), respectively:

Equation (2) reflects the characteristic curve of the parallel circuit RC, which was obtained eliminating ω from Equations (3) and (4). From equation (2), we can see that the locus in the complex plane only depends on the R_ct and R_m values but not C_dl. This result would explain the fact that there are no statistically significant variations in the R_ct measurements.

The statistical design was carried out to ensure that systematic and random differences between experimental units (electrodes) do not interfere with the comparison between means. T-tests for related samples were performed. Table 2 shows that, on average, there is not a statistically significant difference for ΔR= R_ctb-R_cta (last column).

Moreover, from equations (3) and (4), it can be seen that if C_dl decreases, equal frequency points on the circle move to the right (that is, to higher Z′ values). This result is shown in Fig. 3, where it can be seen that the black full circle corresponding to 1 Hz is displaced to the right in relation to the red full circle.

The effect over the impedance due to add supernatant or YW, is presented in Fig. 4. This figure shows that adding either supernatant or YW generated change in C_p%, but for different reasons. Impedance changes observed when supernatants (S_n) were added, were due to the different conductivity of these with respect to BS (Figure 5). These changes depend obviously on the supernatant added (S1, S2, etc.) and on the concentration of the YCS they come from. That is, if you add the S1 supernatant from an YCS of 10⁹ CFU/ml, will not produce the same change in C_p% that if S1 from an YCS of 10⁸ or 10⁷ CFU/ml were added.

From Fig. 5, it was observed that the larger the difference between the supernatant and BS conductivities, the greater the change in C_p%.

On the other hand, when it was added YW with the same conductivity than BS, the change experienced by C_p% was due exclusively to the presence of yeast cells and not to the changes of conductivity. These facts are very important because they show that to quantify the presence of yeast cells, the effect of the difference between the supernatant and BS conductivities, must be eliminated first. This means that YCS must be washed until conductivities of YW and BS are equal within the experimental error. Moreover, if the effect of the different conductivity of the supernatant is not removed, changes in C_p% will be due to these two effects together and therefore the measurements will be unreliable.

The Fig. 6 shows the strong dependence of C_p% with the cell suspension concentration, previously washed, so that the YW and BS conductivities coincide within the experimental error. To confirm that this the changes in impedance measurements were only due to yeasts attachment, SEM images of the electrode surface have been taken in three different situations: naked electrode (Fig. 7a), after a measurement in which supernatant S1 (which comes from to concentration of 10⁹ CFU/ml) was added (Fig. 7b) and after a measurement in which YW (also of 10⁹ CFU/ml) was added (Fig. 7c). The first two images were taken at a magnification of 1000x, while the third and its inset were taken at 150x and 2000x respectively. Note in Figure 7b that no yeast cells attached to gold electrode were observed.

Fig.7

From a dielectric perspective, the simplest representation of a yeast cell is a sphere surrounded by a membrane and covered by a cell wall. The interior of the cell has a net negative charge, which is necessary to establish a membrane potential. The cell wall is essentially a layer of polysaccharides, proteins, lipids and inorganic phosphate.

The yeast cell surface charge originates from the presence of carboxyl, phosphate and amino groups in either dissociated or protonated form, and the surface charge consequently depends upon the pH [28] and/or ionic concentration of the suspending medium [29]. Poortinga et al [28] suggested that, at physiological pH values (i.e., between pH 5 and 7), bacterial cells are generally negatively charged due to the excess of carboxyl and phosphate groups over amino groups.

It is known that, from an electric point of view, the cell is surrounded by a thin layer of high electrical conductivity. Most ions in this layer have the opposite charge in relation to the charge of the cell [30].

Let us consider that the net surface charge of a yeast cell is negative, while the net charge of the ions from the solution and close to the surface of the yeast cell is positive.

When the cell is near the outer Helmholtz plane (OHP), its counterion atmosphere is asymmetrically redistributed on both sides of the cell due to the presence of EEI.

The electric charges in the liquid phase (OHP and diffuse layer) are closer to the yeast cells than the electric charges in the solid phase (metallic electrode); therefore, the electric field acting on the yeast cells cannot be null.

This electric field on the yeast cell could produce a redistribution of its counterion atmosphere. As a result of this process, the cell is polarized, even in the absence of an external overpotential applied to the system. The asymmetric charge distribution can be modeled as a net capacitor C_b in series with C_dl, as shown in Fig. 8.

Fig.8

This model explains the behavior showed in Fig. 6. That is, the larger the yeast cell concentration, the greater the C_b and greater the C_p% change, in absolute value.

The fitting shown in this Fig. is given by equation (5) which has a correlation coefficient of 0.9997.

where C is the concentration.

When the external overpotential is applied to the EEI, an additional phenomenon occurs: (1) the potential on the EEI is modulated according to the external potential. (2) Consequently, a modulation of the asymmetric charge distribution is induced. (3) When performing a frequency sweep, the alpha relaxation phenomenon [31] could be distorted because the cells start from a previous DC polarization.

Finally, an important and novel aspect of this work is that the effect of the attachment of yeast cells on the electrode-electrolyte interface impedance was quantified instead of the joint effect of attachment and changes in the conductivity of the medium, as published by other authors (Muñoz-Berbel, Treo, Blake-Coleman). Consequently, this work provides the basis for future clinical and industrial applications such as the quantification of microorganism concentrations through their attachment or study of the rate of biofilm formation by impedance spectroscopy.