Changes on the physico-chemical surface properties and adhesion behaviour of Enterococcus faecalis by the addition of serum or urine to the growth medium

The initial microbial adhesion of bacteria to different surfaces seems to be mediated by physico-chemical forces and this is the reason why the physico-chemical surface characterisation of bacteria has recently gained interest.

In this context, the adhesion of different microorganisms to biological substrata has been described from a physico-chemical point of view, aiming to simulate, as closely as possible, the conditions of interest.

On this basis, the objective of this work is to characterise the surface of Enterococcus faecalis ATCC29212 through hydrophobicity, surface free energy and zeta potential at 37 °C, when cells grow in Trypticase Soy Broth (TSB) and TSB supplemented with serum or urine.

These variations are used to provide a theoretical description of the bacterial adhesion to glass by interaction free energy, which is verified with experimental results employing a parallel plate flow chamber set at 37 °C, to simulate the conditions of flow inside the human body.

The results show that the addition of serum to the growth medium increases the hydrophobicity and isoelectric point (i.e.p.) of the microorganisms, and this could indicate an increase in the protein content on the cell surface.

However urine does not introduce a change in the above magnitudes.

At short separation distances between the cells and the substratum, the interaction free energy predicts a favourable adhesion for serum-grown cells while non-favourable adhesion is expected for control (TSB-grown cells) and urine-grown cells.

These results are in agreement with the experimental adhesion data, obtained with the flow chamber.


Bacterial adhesion to different surfaces is a process that is of importance in many fields.1–3

In particular, in the field of medicine, the adhesion of microorganisms to inert surfaces has gained importance in recent decades due to the increasing number of clinical practices which require the use of implants.

One of the microorganisms of current interest because of its adhesion to such biomaterials is Enterococcus faecalis.

This bacterium is responsible for a great number of nosocomial infections including urinary tract and abdominal infections, wound infections, bacteremia and endocarditis.4,5

The initial adhesion of cells to biomaterials is followed by the formation of a biofilm which, because of its structure, is very difficult to eradicate.6

In these cases, the treatment with antimicrobial agents results in a non-effective therapy.7,8

Therefore, the prevention of the formation of biofilm at its initial stages, seems to be the most promising procedure.

For this reason many papers have studied the initial interaction between microorganisms and different substrata.9–11

It is widely accepted that the initial adhesion of bacteria to biomaterials is mediated by physico-chemical forces similar to those acting in the adhesion of colloidal particles.6,10

In this sense, theoretical predictions of the adhesion process based on interaction free energy calculations have been compared with experimental adhesion results in different investigations.10,12

The positive correlation between both magnitudes could help to predict and consequently control biofilm formation.

The interaction free energy between a particle (in this case bacterium) and a surface is described, from a theoretical point of view, as the sum of the Lifshitz–van der Waals, acid–base and electrical interaction free energies.13,14

This theory, known as the extended DLVO (X-DLVO: Derjaguin, Landau, Verwey and Overbeek), provides the dependence of the interaction free energy with the separation distance.

The long-range Lifshitz–van der Waals and the short-range acid–base forces are obtained through the surface free energy of both interacting surfaces, expressed as the sum of the Lifshitz–van der Waals and acid–base surface free energy components (γLW and γAB, respectively).

Both components are calculated from the contact angles that probe liquids form on lawns of bacteria deposited on filters.

The electrical interaction free energy is obtained from the zeta potential (ζ) of bacteria and substrata.

In relation to bacterial adhesion, there are different methods to quantify the adhesion of microorganisms to biomaterials.

Static and dynamic adhesion-based tests are the most commonly used.10

However, static tests have been criticised because they neglect the conditions of flow that bacteria find in different parts of the human body and because they usually need a biochemical method to permit the count of the adhered bacteria.

This work therefore utilises a flow chamber to carry out the adhesion of E. faecalis to glass, allowing a live observation of microorganisms adhered to solid substrata.

In line with the above explanation, the goal of this work is the physico-chemical characterisation of E. faecalis ATCC29212 through measurements of water, formamide and diiodomethane contact angles and zeta potential determinations at 37 °C.

The incubation of cells is carried out in a standard culture medium (TSB) and TSB supplemented with serum or urine, in order to simulate the physiological conditions in which bacteria cause infections.

The theoretical predictions of the adhesion process of such microorganisms to glass, defined by the X-DLVO, are verified with the adhesion experiments carried out in a parallel plate flow chamber at 37 °C.


Bacteria preparation process

E. faecalis ATCC29212 bacteria were stored at −80 °C in porous beads (Microbank, Pro-Lab Diagnostics, Austin, Texas, USA).

From the frozen stock, blood agar plates were inoculated and incubated at 37 °C to obtain cultures.

Bacteria were incubated overnight in 100 ml of urine-free Trypticase Soy Broth (TSB) (BBL, Becton Dickinson, Cockeysville, Maryland, USA) or TSB containing 10% human serum or 50% human urine at 37 °C.

The cells were harvested by centrifugation at 37 °C for 5 min at 1000 g (Sorvall TC6, Dulont, Newtown, Pennsylvania, USA) and washed three times with phosphate buffer saline (PBS, 0.15 mol l−1) for flow experiments and with water (Millipore, Mosheim, France) for contact angle measurements.

Serum and urine collection

Serum and urine were collected from 5 healthy men, between the ages of 20 and 45.

They were pooled and sterilised by filtration employing 0.45 μm pore-size filters (Millipore, Mosheim, Francia) and volumes of approximately 10 ml and 50 ml of sterile serum and urine, respectively were stored at −20 °C in sterile flasks.

Prior to use, the flasks were warmed at room temperature.

Adhesion experiments—the flow chamber

Glass (Menzel-Glaser, Braunschweig, Germany) was employed as the substratum for the adhesion experiments in the parallel plate flow chamber.

It was cleaned by sonication for 5 min in 2% v/v commercially available surfactant solution (Vilenet, Vileda Iberica, Barcelona, Spain), rinsed thoroughly with water, sonicated in water for 5 min, and finally inserted in the flow chamber.

The parallel plate flow chamber (dimensions: l × w × h = 7.6 × 3.8 × 0.06 cm) has been described in detail before.15

For enumeration of the adhered bacteria, the flow chamber was placed on the stage of a microscope (Leitz Diaplan, Leica, Wetzlar, Germany) equipped with a 40x ultra-long working distance objective (Leica, Wetzlar, Germany) with a numerical aperture of 0.5.

The images obtained from the microscope focussed on microorganisms adhered to the bottom plate of the flow chamber were registered by a CCD camera (Hitachi, Tokyo, Japan).

Images were recorded in sets of three images, captured at intervals of a tenth of a second.

Adhered bacteria were identified from non-adhered ones by comparing the three images and counting the fixed bacteria.

An image covered a surface area of 0.025 mm2.

Experimentally, E. faecalis ATCC29212 bacterium grown in TSB or in TSB with serum or urine, were suspended in PBS to a concentration of 3 × 108 cells ml−1 and the suspension flowed through the system at 37 °C for 4 h.

A pulse-free flow (0.034 ml s−1) was created by the hydrostatic pressure, and the suspension was recirculated by a Miniplus 2 peristaltic pump (Gilson, Villers le Bel, France).

During the first 30 min of adhesion, a large number of images were taken to determine the initial deposition rate, after that images were taken every 10 min approximately up to 4 h.

Experiments were repeated at least three times with separately cultured bacteria.

The numbers of adhered bacteria were compared using an unpaired Students t-test.

Changes were statistically significant if P < 0.05.

P gives the probability of observing differences between samples means when the confidence interval selected is of 95%.

Contact angle measurements. Surface energy calculations. Interaction free energy

Water (Milli-Q Plus), formamide (puriss > 99.0%, Fluka, Switzerland) and diiodomethane (purum > 98%, Fluka, Switzerland) contact angles on lawns of partially dried bacteria at 37 °C without the residual water were determined using the sessile drop technique.16

The values of the Lifshitz–van der Waals (LW) surface tension component (γLW), electron-donor (γ) and electron-acceptor (γ+) parameters of the acid–base (AB) surface tension component (γAB) of these three liquids at 37 °C are the following: for water γLW = 20.3 mJ m−2, γ = 25.0 mJ m−2, γ+ = 25.0 mJ m−2, γAB = 50.0 mJ m−2; for formamide γLW = 36.7 mJ m−2, γ = 38.3 mJ m−2, γ+ = 2.2 mJ m−2, γAB = 18.4 mJ m−2 for diiodomethane γLW = 48.1 mJ m−2, γ = 0.0 mJ m−2, γ+ = 0.7 mJ m−2, γAB = 0.0 mJ m−2 being calculated in terms of the constant relation for the γ+ and γ parameters for water.17

Briefly, bacteria suspended in demineralised water were layered onto 0.45 μm pore size filters (Millipore, Molsheim, France) using a negative pressure.

Filters were left to air dry at 37 °C for 45 min.

The time of drying was the time at the “plateau contact angles” (θ) could be measured,17 and has been checked previously.

After that, the filters were introduced in an environmental chamber G211 (krüss, Hamburg, Germany), connected to a thermostat to maintain the temperature at 37 °C.

Before measuring the contact angle with a probe liquid, the chamber was allowed to saturate with vapour of the liquid employed.

Subsequently, the contact angles were obtained analysing the images captured with a computer.

The surface free energy components, γLW, , were calculated from the three-equation system by applying the Young–Dupré equation to each probe liquid (L): where γL = γLWL + γABL is the surface tension of the probe liquid and the subscript B denotes bacteria.

The total interaction free energy, ΔGT, between microorganisms and substrata through water (W) as a function of the separation distance (d) is calculated by the sum of the Lifshitz–van der Waals (LW), the acid–base (AB) and the electrical interaction free energies (ΔGLW, ΔGAB and ΔGEL, respectively) as proposed by the extended DLVO theory (X-DLVO).13,14

These three components are calculated as follows.

Expression of ΔGLW: A is the Hamaker constant, A = −12πd20ΔGLWadh, a is the microorganism radius, d0 the distance of the closest approach between two surfaces and ΔGLWadh is obtained through the Lifshitz–van der Waals surface free energy as follows, subscripts B and S being bacterium and substratum, respectively, and W indicates that surface free energy of the suspending liquid is nearly equal to water.

Expression of ΔGAB: Expression of ΔGEL: ε0 is the dielectric constant of the vacuum and εr the relative dielectric constant of the suspending liquid.

ζB and ζS are the zeta potentials of bacteria and substrata and the subscript S denotes substratum.

Interaction free energies predict favourable adhesion when they are negative and no adhesion when they are positive.

Results and discussion

The addition of serum and urine to the growth medium of E. faecalis ATCC29212 involves a series of changes in its physico-chemical surface properties and adhesion behaviour as the following tables and figures indicate.

Physico-chemical surface properties

The first columns of Table 1 contain the water, formamide and diiodomethane contact angles measured at 37 °C on lawns of bacteria grown in TSB (control), TSB with 10% serum and TSB plus 50% urine.

Serum increases the water contact angle of microorganisms which, in terms of hydrophobicity, means that under such a condition E. faecalis becomes more hydrophobic.16,18

On the other hand, urine has no effect on the surface hydrophobicity of the studied bacteria, which remains invariable compared to that of the control, within the experimental error.

The formamide contact angle does not depend on the constituents of the growth medium, while the diiodomethane contact angle only changes for urine-grown cells.

These contact angles of the probe liquids provide, through eqn. (1), the components and parameters of the surface free energy of the cells, which also appear in Table 1.

Serum-grown cells have the highest Lifshitz–van der Waals surface free energy component, but the lowest acid–base surface free energy component, making the total surface free energy in this case very similar to that of the control and urine-grown cells.

The asymmetry found between the electron-donor and electron-acceptor parameters of γAB is very similar in the cases of the control and urine-grown cells and very much higher than that for serum-grown cells.

According to Van Oss et al.,19 this result emphasises the higher hydrophobicity of serum-grown cells compared to the other two samples studied, taking into account that the degree of asymmetry between γ+ and γ is an indicator of the level of hydrophilicity.

The change that serum introduces in the cellular surface hydrophobicity of cells is in agreement with the results obtained by microbial adhesion to hydrocarbons (MATH method) and recently published by this group.20

In relation to the electrical characterization of the cell surface, the last column of Table 1 presents the values of the zeta potential of bacteria when they are suspended in PBS, a liquid also employed in the adhesion experiments.

The data show that, within the experimental error, the addition of urine to the culture medium produces a very small increase in the effective negative charge on the bacterial surface, while the addition of serum makes the net negative charge higher.

It is interesting to point out that when we say that ζ is a measure of the effective charge, we mean that it measures the electrical potential which one particle manifests when it approaches another particle.

In this context, it is the quantity which characterises the electrical force between the approaching particles, as can be verified below.

Additional information on the surface charge of the microorganisms is provided by the isoelectric point (i.e.p.), which measures the value of the pH when the net surface charge is equal to zero.

Thereby, Fig. 1(a) shows the zeta potential versus pH for the three cases studied, when cells are suspended in KPi—buffer of low ionic strength, with physiological properties and widely used in determining the i.e.p. of biological samples.21,22

When urine is added to the culture medium, the ζ–pH curve is in most cases similar to that of the control, with the exception of those points obtained around pH 3.3.

Nevertheless, serum diminishes the net surface charge of E. faecalis over the entire pH range.

These graphs indicate that the i.e.p.s of control and urine-grown cells are nearly equal (1.9 and 2.0, respectively), while the i.e.p. of serum-grown cells is higher and equal to 2.9.

We also determined the i.e.p. when cells were suspended in the same liquid as that employed in the adhesion experiments, i.e. PBS, a liquid with a higher ionic strength.

The results are presented in Fig. 1(b).

Firstly, it has to be mentioned that the amount of acid or base required to change the pH of PBS suspensions was very much higher than in the case of KPi, due to the higher stability of the buffer PBS.

Despite the fact that PBS makes the curves less pronounced than those obtained with KPi, because of the reduction in the net surface charge of microorganisms (due to the compression of the electrical double layer), the effects of serum and urine on the zeta potential are similar to those observed in the case of KPi.

This means that urine does not introduce important changes in ζ over the entire pH range selected, while serum makes the microbial surface less negatively charged, especially at low pHs.

Based on the curves (Fig. 1(b)), it can be seen that neither the control nor the urine-grown cells have their i.e.p.s in the pH range studied, while the i.e.p. of serum-grown cells is approximately 1.8, a unit less than with KPi, in agreement with the ionic characteristics of both suspending liquids.

The increase in the i.e.p. of serum-grown cells with respect to the control and the invariability of the i.e.p. in the case of the urine-grown cells is positively correlated with the higher and equal hydrophobicity of serum-grown cells and urine-grown cells, respectively, compared to that of the control.

In addition to the previous analysis of the i.e.p. as a magnitude to characterise the electrical behaviour of the cell surface, the i.e.p. value has been related by several authors to chemical components of the bacterial surface such as proteins.11,23,24

It is widely accepted that a direct relation between i.e.p.s and the ratio nitrogen–carbon (N/C, determined by X-ray electron spectroscopy) exists.11,23,24

The explanation is based on the excess of amino groups, NH4+, positively charged, against phosphate groups, PO43−, negatively charged on the bacterial surface.

This picture causes a reduction in the net value of ζ over the pHs, with the consequent increase in the i.e.p..

Some authors11,24 have related the increase in NH4+ groups (and also the i.e.p. increase) with a rise in the number of proteins exposed to the bacterial surface which, due to their hydrophobic character, contributes to an increase in the cellular surface hydrophobicity.

This result is in total agreement with the hydrophobic behaviour of E. faecalis as shown in the first column of Table 1, where serum-grown cells are presented as the most hydrophobic microorganisms.

The effect of serum on the hydrophobicity of the cell surface of enterococci is in agreement with the results obtained by Ljungh and Wadstrom25 when they determined the hydrophobicity of different microorganisms (among them E. faecalis) grown with serum.

Nevertheless, these authors could not attribute the phenomenon to a specific serum compound.

The direct relation between i.e.p. and hydrophobicity found in this work, in addition to the correlations between i.e.p., hydrophobicity and NH4+ obtained by our group and other researchers in previous studies, may indicate that the presence of proteins, contained in serum, on the bacterial surface is responsible for the increase in cellular surface hydrophobicity of serum-grown bacteria.

Theoretical predictions and experimental results of adhesion

The main objective when analysing the physico-chemical surface parameters of E. faecalis grown in TSB and TSB supplemented with serum or urine is to check whether they are able to predict the initial adhesion of the microorganisms.

On this basis, interaction free energy, obtained through the above physico-chemical surface parameters, is related to the number of adhered cells discussed in this section.

When a bacterium approaches a surface, the value of the interaction free energy, ΔGT, is a function of the separation distance between both phases, as described by the extended DLVO theory (X-DLVO).

When representing ΔGTversus the separation distance (d) many systems present two minima, one called the secondary minimum (in most cases no higher than 100 kT, k and T being Boltman's constant and absolute temperature, respectively) and the other called the primary minimum, located at shorter distances, much deeper and separated from the secondary minimum by a potential barrier (in the order of some tens of kT).

The state of stable equilibrium corresponds to the primary minimum and the ‘jump’ from the secondary to the primary minimum is relatively easy for cells, which use in their metabolism an amount of energy in the range of 106–108kT.26

When calculating the ΔGT in biological systems to predict non-specific bacterial adhesion, the minimal separation distance (1.57 Å) between cells and substrata is estimated in order to form a new interface.27

This implies that bacteria falls into the primary minimum, an assumption that cannot be completely true if the primary minimum does not exist.

This reasoning led us in this work to take into account not only the value of ΔGT at 1.57 Å but also to draw the ΔGT curves as a function of the separation distances between the interacting surfaces.

The graphics are presented in Figs. 2–4, each relating to one of the samples studied.

At first glance, one can observe the similarity between Figs. 2 and 4, corresponding to control and urine-grown cells, while Fig. 3 appears very different from the others, not only at high separation distances but also when in close approximation (compare Figs. 2(a), 3(a) and 4(a) and Figs. 2(b), 3(b) and 4(b)).

This result is well correlated with all the data belonging to the previously mentioned physico-chemical characterisation of the cells, in which control and urine-grown cells behave in a similar way and differently from serum-grown cells.

On analysing each type of force as a function of the separation distance, Lifshitz–van der Waals interaction free energy governs the interaction between E. faecalis and glass at high separation distances (this is clearly verified from 50 Å for all the cases studied).

At intermediate distances (from approximately 10 Å to 50 Å), the Lifshitz–van der Waals and mainly the electrical force control the sign and magnitude of the total interaction free energy, while at short distances (below 10 Å, see especially Figs denoted with (b)) the acid–base force is the dominant interaction.

With respect to the shape of ΔGT, serum-grown cells exhibit a deep primary minimum (Fig. 2(b)), while control and urine-grown cells lack this potential well.

This fact implies, from a physico-chemical point of view, that serum-grown cells would have a far higher initial adhesion to glass than in the other cases.

The high potential barrier that control and urine-grown cells have at the minimal separation distance means that the bacteria are not able to reach this state and so, from a theoretical point of view, the adhesion to glass of E. faecalis grown in TSB and in TSB supplemented with urine is not possible.

The cells in the last two cases would remain in the secondary minimum, which for the control is about −6 kT and for urine-grown cells −0.5 kT, at distances of 35 Å and 59 Å, respectively, far from the minimal interaction distance required to form a new interface.

In order to validate the theory, the interaction free energy of the adhesion process of E. faecalis to glass should be checked with the experimental adhesion data.

In this sense, Fig. 5 represents the number of adhered cells per square centimetre throughout the adhesion period in a parallel plate flow chamber.

It can be clearly seen that the adhesion to glass of E. faecalis grown with serum is higher than that of the control and urine-grown cells (P < 0.05), the last two cases being mostly similar (P > 0.05).

These adhesion data are in good agreement with the theoretical predictions of adhesion dictated by the X-DLVO theory, in the sense that ΔG values anticipate favourable adhesion to glass of serum-grown cells against the control or urine-grown bacteria.

In these last two cases, adhesion is expected to be similar between them and lower than serum-grown cells.


The present study brings to light that components of serum modify the physico-chemical surface of E. faecalis by enhancing its hydrophobicity, isoelectric point and by changing the components and parameters of the surface free energy.

Such changes predict a favourable adhesion to glass taking into account the X-DLVO theory, which is in agreement with the experimental adhesion data.

On the other hand, the supplementation of the culture medium with urine does not produce any effect on the hydrophobicity and surface charge of microorganisms, as shown by the similar adhesion curves obtained for this sample and that of the control.

This work can help to deepen in the knowledge of the infections by E. faecalis, which in the last decades has become an important nosocomial pathogen.

Its ability to modify its physico-chemical properties in the presence of serum, favouring its adhesion to inert surfaces, could, for example, explain the importance of enterococcus in endocarditis.

Moreover, enterococcal infections of the urinary tract could be related to other kinds of forces, different from physico-chemical interactions.

This last point remains unclear and warrants further investigation.