Quartz tuning fork studies on the surface properties of Pseudomonas aeruginosa during early stages of biofilm formation

Colloids and Surfaces B: Biointerfaces 102 (2013) 117–123

Contents lists available at SciVerse ScienceDirect

Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb

Quartz tuning fork studies on the surface properties of Pseudomonas aeruginosa during early stages of biofilm formation ˜ b , Laura González a , Eduard Torrents b , Antonio Juárez b,c , Manel Puig-Vidal a,∗ Jorge Otero a , Rosa Banos a b c

SIC-BIO, Bioelectronics and Nanobioengineering Group, Departament d’Electrònica, Universitat de Barcelona, Marti i Franques 1, 08028 Barcelona, Spain Institute for Bioengineering of Catalonia (IBEC), Baldiri Reixac 15-21, 08028 Barcelona, Spain Departament de Microbiologia, Facultat de Biologia, Universitat de Barcelona, Avda. Diagonal 643, 08028 Barcelona, Spain

a r t i c l e

i n f o

Article history: Received 23 February 2012 Received in revised form 19 July 2012 Accepted 7 August 2012 Available online 19 August 2012 Keywords: Atomic force microscopy Biofilms Elasticity Pseudomonas aeruginosa Tuning fork

a b s t r a c t Scanning probe microscopy techniques are powerful tools for studying the nanoscale surface properties of biofilms, such as their morphology and mechanical behavior. Typically, these studies are conducted using atomic force microscopy probes, which are force nanosensors based on microfabricated cantilevers. In recent years, quartz tuning fork (QTF) probes have been used in morphological studies due to their better performance in certain experiments with respect to standard AFM probes. In the present work QTF probes were used to measure not only the morphology but also the nanomechanical properties of Pseudomonas aeruginosa during early stages of biofilm formation. Changes in bacterium size and the membrane spring constant were determined in biofilms grown for 20, 24 and 28 h on gold with and without glucose in the culture media. The results obtained using the standard AFM and QTF probes were compared. Both probes showed that the bacteria forming the biofilm increased in size over time, but that there was no dependence on the presence of glucose in the culture media. On the other hand, the spring constant increased over time and there was a clear difference between biofilms grown with and without glucose. This is the first time that QTF probes have been used to measure the nanomechanical properties of microbial cell surfaces and the results obtained highlight their potential for studying biological samples beyond topographic measurements. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Cystic fibrosis (CF) is the most common life-shortening recessive genetic disease. There is no cure and, although medical research has improved the outlook for patients, more than half of them die before they are 18 years old [1]. There are several symptoms of CF, but the most important is the lung disease associated with chronic infections. In the initial stages of the illness, the lungs are mainly colonized by Staphylococcus aureus and Haemophilus influenzae but, during the entire life of the patient, Pseudomonas aeruginosa infection is the most prevalent [2]. The reason why CF patients are predisposed to these infections remains a very active area of research and a definitive answer has yet to be found [3]. Hence, preventing infection is relatively difficult and another important area of research is antibiotic therapy that can be given once the patient is infected (the average survival of a CF patient without P. aeruginosa infection is more than 10 years longer than that of infected patients [4]). P. aeruginosa usually forms biofilms in the lungs. Biofilms are communities of bacteria that adhere to the lung

∗ Corresponding author. Tel.: +34 934039161; fax: +34 934039161. E-mail address: [email protected] (M. Puig-Vidal). 0927-7765/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfb.2012.08.013

surface [5]. In the early stages of biofilm formation, bacteria adhere to the surface and then initiate the production of an extracellular polymeric matrix (EPS – extra polymeric substance) [6,7]. The EPS is mainly formed by polysaccharides, proteins and nucleic acids. These form a matrix around the cells. Bacteria inside the EPS matrix show increased resistance to extreme survival conditions such as the lack of nutrients or the presence of antibiotics in the media [8]. The mechanism of biofilm formation and why the bacteria show increased resistance to conventional antibiotic treatments are not well understood. The heterogeneity of the EPS-bacterial structure and the difficulty that antibiotics encounter in passing through the polysaccharide matrix are the main issues under current investigation [9]. In recent years, the evolution of nanotechnologies has opened the door to the study of several fields of biomedicine with an unprecedented level of detail [10,11]. Of the multiple applications of nanotechnology in microbiology, a topic that presents a particular challenge is that of biofilm formation behavior [12]. Of the multiple nanocharacterization techniques, atomic force microscopy (AFM) is of special interest because it is capable of imaging soft, non-coated samples under physiological conditions [13,14]. AFM can also be used to measure the mechanical, electrical and chemical properties of biological samples [15–17]. In recent

118

J. Otero et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 117–123

years, several AFM studies on biofilm structure and formation have been published. The main objective of all of them was to better understand the behavior of these biological structures: their morphology and cohesion and why they show increased resistance to treatment. Ref. [18] studied the morphology of P. aeruginosa biofilms over different materials, along with the effect of hot water washing; the objective of the study was to engineer better materials for washing machines. The elasticity of the bacteria forming the biofilm was reported in [19,20]: these studies tried to understand why the biofilm bacteria have a harder outer membrane in comparison with non-biofilm forming bacteria, and the influence of the buffer in the measurements. In a similar way, Ref. [21] attempted to relate the adhesion between the tip and the bacteria with their pathogenicity. The effect of some antibiotics on morphology and elasticity was reported in [22]. Some research groups have gone one step further by trying to implement novel AFM modes to indirectly measure the interaction and adhesion between the biofilm and the surface [23,24]. Nevertheless, such measurements are indirect (for example, involving looking at the energy dissipation in the friction between the tip and the biofilm while scanning), and the results are difficult to interpret. The AFM technique is based on the measurement of the interaction force between a nanometric radius tip and the sample surface. The tip is usually at the end of a microfabricated cantilever that bends when a force occurs. This deflection is measured by focusing a laser at the end of the cantilever and measuring its reflection in a position-sensitive photodiode. An alternative method that is increasing in popularity involves the use of a quart tuning fork (QTF) probe. Originally introduced by [25] to maintain a constant distance between the optical fiber and the sample in near-field scanning optical microscopy (NSOM), QTF probes have shown their ability to image the sample surface with nanometric accuracy [26,27]. A tip is mounted in the QTF resonator and oscillated parallel to the sample surface. The force interactions produce changes in the amplitude of oscillation and the resonant frequency of the probe. Due to its specific mechanical structure and the piezoelectric properties of quartz, QTF probes present higher quality factors than microfabricated cantilevers: this is of special importance when working in liquid media [28]. Furthermore, the elimination of the laser–photodiode system allows easier integration with optical microscopy and the cooperation of multiple probes over the same sample surface [29]. Only a few studies have been carried out using QTF probes for the measurement of biological samples [30–32], and the use of these probes to measure nanomechanical properties and for molecular recognition is even less common [33,34]. The aim of the present work was to investigate the feasibility of using QTF probes for morphological and nanomechanical studies in cells able to form biofilms, such as P. aeruginosa. AFM and QTF probes were used to measure properties of the samples, and the results obtained with the two techniques were compared.

thin gold film over freshly cleaved mica using the method described in [36]. Experiments of the P. aeruginosa biofilms grown on flat gold substrates were conducted at room temperature in a nitrogen atmosphere. Samples were cultured for the specified times and in the selected conditions and then washed with distilled water. Bacteria were not fixed to the substrate using a biochemical agent because the adhesive properties of the biofilm were sufficient to image them if the correct probe was used (soft enough probe). Samples were dried immediately before performing the measurements: they were placed inside the microscope sample holder and then the instrument was protected by a glass bell, which was immediately filled with nitrogen. After that, all measurements were performed within the first 24 h following sample preparation. Rather than measuring the turgor pressure, which is measured when samples are immersed in liquid and depends on the pH and osmotic pressure [37], samples were dried to enable measurement of the mechanical properties of the membranes.

2.2. AFM and QTF imaging All AFM and QTF experiments were performed using a Cervantes AFM microscope (Nanotec Electrónica S.L., Madrid, Spain). Two commercial OMCL-RC800PSA (Olympus Corp., Tokyo, Japan) cantilevers were used (one for each set of samples) for AFM imaging. These cantilevers were individually calibrated prior to use via the Sader method [38], with a normal spring constant K of 0.62 N/m and 0.58 N/m, respectively. The QTF probes were home-fabricated by attaching a chemically sharpened optical fiber (using the meniscus etching described in [39], immersing the fiber end in a 48% Hydrofluoric acid solution and using Isooctane as a protection layer) to one of the tines of a decapsulated commercial AB38T (Abracon Corp., California, US) QTF with a nominal resonant frequency of 32.768 Hz. The tip radius was about 200 nm, according to measurements made using the TGG01 pyramidal calibration structure (Mikromash, Tallin, Estonia), with a nominal curvature radius below 10 nm [40]. The static spring constant K of the QTF was calculated from geometric parameters [41] as 10,080 N/m (Fig. 1). AFM images were acquired in dynamic mode using standard laser–photodiode measurement of the cantilever deflection via the method described in [42,43]. QTF images were acquired using the scheme presented in Fig. 2b: the QTF nanosensor was driven electrically to its resonant frequency. The free amplitude of oscillation A and quality factor (Q, the relation between the amplitude at the resonant frequency f0 and the bandwidth [44]) were set to 1 nm and 1000, respectively, using the specifically developed electronic driver [45,46] shown in Fig. 2b The mechanical amplitude of motion was calculated from the current flowing through the device [47] using a lock-in amplifier, according to the equation derived in [48]:

 A=

Vrms · Irms · Q K · 2f0

(1)

2. Materials and methods 2.1. Sample preparation P. aeruginosa cells were grown statically in petri dishes containing liquid LB medium at 37 ◦ C. One cm2 flat gold substrates were placed in the petri dishes just after inoculating them (1:100) from an overnight liquid LB culture. At the desired time intervals (20 h, 24 h and 28 h), these gold substrates were removed for analysis. Biofilm development by P. aeruginosa was studied under two conditions: with and without glucose in the culture medium. The presence of the sugar increases biofilm development [35]. Flat gold substrates (Zrms < 1 nm) were prepared by thermal evaporation of a

where Vrms is the amplitude of the excitation signal and Irms is the measured current. If the QTF sensor oscillates parallel to the sample surface, in what is usually named the shear mode, force interactions between the tip and the sample result in changes in the amplitude of oscillation and the resonant frequency [49] of the device. The measured phase can then be used as input in a phase-locked-loop (PLL) to maintain the QTF in resonance by maintaining a constant 0 phase. The amplitude of oscillation A is used as the main feedback signal for the Z sample movement (see [50] for a complete description of the AM mode using a PLL). The sample is then scanned in X and Y directions and the Z movements of the scanner are used to maintain the amplitude

J. Otero et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 117–123

119

Fig. 1. (a) Phase contrast optical image of the fabricated tuning fork probe. (b) Scheme of the QTF measurement setup. The QTF probe is excited electrically by setting the amplitude of oscillation and the quality factor. The current through the device is measured by a lock-in amplifier: the phase is used to maintain the QTF in resonance using a PLL while the amplitude is used as the main feedback signal to maintain a constant interaction between the tip and the sample and to reconstruct the surface topography.

calibrated prior to the measurements by obtaining a series of force curves over the substrate. Force measurement using the AFM probe is straightforward as it is a force sensor. By modeling the cantilever as an ideal spring, the normal force can be calculated from the cantilever deflection using Hooke’s law: FN = −Kn · d

(2)

where Kn is the normal spring constant and d is the measured deflection of the cantilever. Photodiode sensitivity was calibrated prior to the experiment by pushing a non-deformable surface and simultaneously recording the Z piezoelectric movement and the electrical signal response. Force measurement using the QTF probe is more complex. The electrically driven tuning fork resonates at a frequency: Fig. 2. Force curve scheme for the loading part. The tip is the sample by moving the Z piezo while the force is recorded. The curve over the substrate is the reference if no deformation is assumed. The curve over the sample, which is deformable, presents a non-linear region (a) and a linear region (b) of lower slope than the reference curve.

K M

2f0 =

(3)

where M is the QTF mass and K its static spring constant. When the tip approaches the surface, both friction and elastic forces appear: Ff = M x˙

of oscillation constant and to reconstruct the 3D topography of the sample. Images (both in the AFM and QTF experiments) were acquired at very low interaction forces to minimize the vertical deformation of the biofilms: the amplitude set-point was fixed at 95% of the free oscillation amplitude of the nanosensors (forces in the range of a few nanoNewtons, below 18 nN which is a force small enough to not damage the cells [19]). The scan speed was set to a low value (5 ␮m/s) to avoid lateral deformation of the sample and potential bacterial detachment. After a series of wide area images were acquired over the same zone to ensure that the sample and setup were stable, smaller images of zones of interest were acquired; these images were post-processed using the WSxM software [51].

where  is the damping coefficient and k is the tip-sample interaction constant, defined in [52] as k=

 2 f f0



−1 K

(5)

where f is the frequency measured during the interaction. To measure the elastic properties of the sample and compare the results with the measurements obtained with the AFM cantilever, the normal force is the tangential component of the elastic force [33,52]:



FN =

2.3. Force spectroscopy After imaging the biofilms, 20 bacterial cells were selected from each sample for mechanical characterization. The normal deflection vs. Z tip-sample distance was measured using the AFM probe. The frequency shift vs. Z tip-sample distance was measured using the QTF probe. A series of 20 measurements over the center of the top (25 nm uncertainty due to repeatability errors in the piezoelectric scanner) of each bacterium were performed. Probe sensitivity was

(4)

Fe = kx

f2 f02



−1

K ·A

(6)

where A is the amplitude of oscillation in the X direction. For small forces, if f = f0 + f and f  f0 , the expression can be rewritten: FN =

2 · K · A · f f0

(7)

Once the force was calculated from the deflection or frequency shift using Eq. (2) or (7), the force was obtained as a function of the Z displacement (force curves).

120

J. Otero et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 117–123

Fig. 3. AFM topographic images of Pseudomonas aeruginosa biofilm samples adhered to gold substrates, imaged in dynamic mode. (a) Growth without glucose for 20, 24 and 28 h, respectively. (b) Growth with glucose for 20, 24 and 28 h, respectively. The Z scale is 300 nm.

2.4. Force curve analysis The nanomechanical properties of the sample can be quantified from the force curves via comparison with the curves obtained when the tip is pushed onto the substrate (assumed to be non-deformable) [53]. Force curves over a soft sample present a non-linear region at low loading forces (Fig. 2, zone a) and a linear region at high loading forces (Fig. 2, zone b). The simplest and most common model used to fit the non-linear region is the Hertz model, which is valid for a parabolic tip and a flat sample [54]: √ 4E R F= ı3/2 (8) 3(1 − 2 ) where E is the elastic modulus of the sample, R is the tip radius,  the Poisson’s ratio of the sample, and ı the deformation (calculated as the difference in the Z actuator movement for a given force with respect to a non-deformable surface). If the sample is assumed to have a linear-elastic deformation, the linear region can be modeled as the interaction of two springs. The procedure and the validation of the model have been extensively described elsewhere [19,37], and the spring constant of the sample ksample can be calculated using the following equation: ksample = msubstrate ·

msample 1 − msample

(9)

where msubstrate is the slope of the linear fit of force curve over the substrate and msample is the linear fit of the high-force region of the force curve over the sample. Absolute elastic modulus quantification from AFM data remains controversial. There are several uncertainties in the calibration procedure, such as the shape of the tip, the angle of the indentation and the Poisson’s ratio [22]. Moreover, the Hertz model assumes a flat surface and is only valid for small indentations, which have been shown that are not only a function of the mechanical properties of the sample [55,56]. The objective of this work was to demonstrate the feasibility of using the QTF probe in nanomechanical measurements of bacterial biofilms and to compare it with the standard AFM technique. Relative measurements, which are less dependent

on calibrations and assumptions, are more suitable for this purpose and are usually adequate [57,58]. Thus the two-springs model was used to measure the mechanical properties of the biofilms and to allow comparisons between the different samples and techniques. 3. Results 3.1. Biofilm morphology Biofilm morphology was studied using the standard AFM technique and the QTF probe (Figs. 3 and 4). Analysis of the morphology was consistent with previous experiments in the field of biofilms [7,59]. As expected, both incubation time and the presence of glucose influenced biofilm formation. Examination of the samples at 20 h revealed individual bacteria exhibiting flagellae (marked by an arrow in the images). Biofilms had started to form in the 24-h samples, the surface was completely covered by the biofilm in the 28-h samples, in which EPS was also present. Analysis of the observations made using the AFM showed that there were no qualitative differences in the morphology of the bacteria as a function of the presence of glucose in the culture media. The AFM data were used to quantify the morphological data due to their higher lateral resolution. The size of 20 bacteria from each sample was measured, using an ellipse approximation (Fig. 5). The comparison was made between the diameters and not the volumes of the bacteria because variations in the mechanical properties of the bacteria could lead to underestimation of the cell volume. Thus the softer the sample was, the greater the error made in quantifying the Z value. There was an increase in the size of the bacteria as the biofilm developed in both cases, but there were no differences between the biofilms grown with or without glucose. 3.2. Nanomechanical properties In contrast to the morphological data, evaluation of nanomechanical properties by AFM revealed significant differences between cells grown with and without glucose (Fig. 6). The results revealed no significant effect of incubation time on the spring constant of bacteria cultured in LB medium without

J. Otero et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 117–123

121

Fig. 4. QTF topographic images of Pseudomonas aeruginosa biofilm samples natively adhered to gold substrates, imaged in shear mode. (a) Growth without glucose for 20, 24 and 28 h, respectively. (b) Growth with glucose for 20, 24 and 28 h, respectively. The Z scale is 300 nm.

glucose. However, there was a significant change in the stiffness of biofilms formed by the glucose-cultured cells over time. Hence, biofilms grown with glucose exhibited higher spring constants for the same culture time. Time dependence of the stiffness was more difficult to analyze for the 28 h measurements. This may be because the spring constants of the cantilevers used were 0.58 N/m and 0.62 N/m (one cantilever was used for each set of samples), and the closer these values are to that of the sample spring constant, the more sensitive the measurement. If, as shown in the experiments conducted with the QTF, bacteria cultured for 28 h had a much higher spring constant, the K value would have been underestimated. However, by looking at the frequency shift as a function of the tip-sample distance, it was possible to collect data using the QTF sensor, providing evidence that the slopes of the curves had a relation with the

Fig. 5. The evolution of bacterial size as a function of culture time and the presence of glucose in the growth media showed that the bacteria increased in size over time but were unaffected by the presence or absence of glucose. Error bars represent standard deviation.

elastic properties of the bacteria (Fig. 7a). Spring constants were calculated using the equations described above. The results are shown in Fig. 7b. The experimental results tendencies were consistent with those obtained using the AFM cantilever: there were differences in the spring constant of bacteria cultured with and without glucose; moreover, the change in mechanical properties as a function of culture time was more clearly shown in experiments performed with the QTF. 4. Discussion and conclusions Here we have studied the morphology and nanomechanical properties of biofilms by using QTF-based nanosensors. The goal was to compare these measurements with those obtained with a standard AFM cantilever. With respect to the sample preparation,

Fig. 6. Evolution of the spring constant as a function of culture time and the presence of glucose in the growth media measured using the AFM cantilever. Error bars represent standard deviation.

122

J. Otero et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 117–123

Fig. 7. (a) Example of frequency shift vs. distance for the one data set. (b) Evolution of the bacterial spring constant as a function of culture time and the presence of glucose in the growth media measured using the QTF. Error bars represent standard deviation.

we decided to dry the samples instead of measuring them in buffer solution. It has been reported that there are not artifacts in the measurements if biofilms are dried [22]. Furthermore, it has been reported that that there are no differences between washed and unwashed samples and new or old biofilms [7]. In addition, it must be considered that measurements in buffer solution are more difficult to interpret: it has been shown that pH and osmotic pressure influence and alter the measured mechanical properties [37,56]. These parameters are hard to control with the commercial equipment we used, and the presence of artifacts could not have been ruled out. Because of the above referred points, we considered that, to compare the results between the two different sensors, the best scenario was to measure the stiffness in dry samples in nitrogen, even knowing that the measured spring constants will be higher under these conditions [20] than the ones reported in the studies conducted in a buffer solution. In our case, we can compare the differences and tendencies between the measurements performed with the two different sensors. In both cases, the stiffness was higher for the biofilms cultured with glucose and there was an increase of the spring constant with the culture time. As reported in [19], large spring constants are typical of biofilms, so the spring constants we measured with the AFM cantilever fit in the higher range of values reported in the literature. Values of spring constants measured with the QTF are higher than

these values. There could be a couple of reasons for this fact: the first one is related with the sensitivity problem mentioned previously; the second one is the difficulty to calibrate the spring constant of the QTF sensors. We calculated it from geometrical parameters, but there is a high controversial in the literature on how to give a correct value for this static spring constant of the sensor, and there are works suggesting that the geometrical calculation overestimates 4–5 times the real spring constant of the sensor [60]. As values given in the literature cover one order of magnitude [60,61] we decided to not over adjust the spring constant of our sensors to match the results of the cantilevers, and to use the most accepted method to calculate it (geometrical). The present work shows the potential for using quartz tuning fork sensors to measure the morphology and nanomechanical properties of biological samples. Images acquired using the standard AFM technique and the QTF demonstrated that the evolution of bacterial biofilms can be evaluated with the latter approach. Consistency in the results of the mechanical measurements showed that the QTF could be an alternative to the standard AFM cantilevers for investigating differences in sample properties. We were also able to correlate a well-characterized physiological process with QTF-obtained mechanical properties. The presence of glucose in the culture medium influences biofilm formation and this, in turn, can be detected by the QTF probe. This example highlights the ability of QTF approaches to characterize and discriminate between biological samples with different mechanical properties. In addition to demonstrating that the elasticity results obtained using the QTF are equivalent to AFM measurements, the use of these novel sensors has some remarkable advantages. The most direct benefit of using a QTF is its high quality factor when operating in liquid. The high static spring constant allows very low amplitudes of oscillation, which together with the high Q leads to high sensitivity in the detection of the frequency shift and interaction forces. Several reports [62,63] have shown that fN forces can be measured using these sensors. When compared with a cantilever, QTFs are extremely stable with respect to temperature and exhibit very low energy dissipation. Hence force measurements are stable enough to allow work with biological samples. Given that there is no need to use a laser and a photodiode to measure the interaction, the potential applications of QTFs are many. For this reason, QTFs are the most commonly used sensors in NSOM applications [64], where a laser could interfere in the near-field optical measurements. The self-sensing properties of the sensor also open up a wide range of applications in biomedical research. The tip is only a few mm long, so sample access is much better than with standard cantilevers; using a QTF, experiments can be performed in petri dishes, well-plates and other array-like sample preparations (which require the use of classical staining or fluorescence techniques [65,66]). The long tip and laser-free detection also facilitate the integration of more than one sensor in the measurement setup. This makes it possible to conduct complex experiments with more than one probe. Ultimately, the integration of QTF sensors in microanalysis systems and lab-on-a-chip devices could solve some of the problems associated with mass detection using microfabricated cantilevers. Acknowledgments This work was supported in part by the Spanish Ministerio de Educación under project TEC2009-10114 and grant AP200601079, and also by the Regional Catalan authorities under project VALTEC09-2-0058 and 2009SGR66. References [1] F.A. Ratjen, Respir. Care 54 (2009) 595.

J. Otero et al. / Colloids and Surfaces B: Biointerfaces 102 (2013) 117–123 [2] A. Bush, E.W. Alton, J.C. Davies, U. Griesenbach, A. Jaffe (Eds.), Cystic Fibrosis in the 21st Century, Prog. Respir. Res., Karger, Basel, 2006, p. 138. [3] S. Van daele, M. Vaneechoutte, K. De Boeck, C. Knoop, A. Malfroot, P. Lebecque, J. Leclercq-Foucart, L. Van Schil, K. Desager, F. De Baets, Eur. Respir. J. 28 (2006) 740. [4] Cystic Fibrosis Foundation, Patient Registry 1995. Annual data report, Cystic Fibrosis Foundation, Bethesda, 1996. [5] S. Moreau-Marquis, B.A. Stanton, G.A. O’Toole, Pulm. Pharmacol. Ther. 21 (2008) 595. [6] I.B. Beech, L. Hanjaknsit, M. Kalaji, A. Neal, V. Zinkevich, Microbiology 145 (1999) 1491. [7] I. Auerbach, C. Sorensen, H. Hansma, P. Holden, J. Bacteriol. 182 (2000) 3809. [8] N. Høiby, T. Bjarnsholt, M. Givskov, S. Molin, O. Ciofu, Int. J. Antimicrob. Agents 35 (2010) 322. [9] R. Donlan, ASAIO J. 46 (2000) S47. [10] G.M. Whitesides, Nat. Biotechnol. 21 (2003) 1161. [11] R. Freitas, Nanomed: Nanotechnol. Biol. Med. 1 (2005) 3. [12] Y. Dufrêne, Nat. Rev. Microb. (2008), http://dx.doi.org/10.1038/nrmicro1948. [13] P. Parot, Y.F. Dufrêne, P. Hiterdorfer, C. Le Grimellec, D. Navajas, J. Pellequer, S. Scheuring, J. Mol. Recognit. 20 (2007) 418. [14] S.R. Cohen, A. Bitler, Curr. Opin. Colloid Interface Sci. 13 (2008) 316. [15] S. Iyer, R. Gaikwad, V. Subba-Rao, C. Woodworth, I. Sokolov, Nat. Nanotechnol. 4 (2009) 389. [16] L. Fumagalli, G. Ferrari, M. Sampietro, G. Gomila, Nano Lett. 9 (2009) 1604. [17] T. Fisher, A. Oberhauser, M. Carrion-Vasquez, P. Marszalek, J. Fernandez, Trends Biochem. Sci. 24 (1999) 379. [18] Y. Oh, N. Lee, W. Jo, W. Jung, J. Lim, Ultramicroscopy 109 (2009) 874. ˜ [19] C. Volle, M. Ferguson, K. Aidala, E. Spain, M. Núnez, Colloids Surf. B: Biointerfaces 67 (2008) 32. [20] F. Teran, R. Carlson, J. Monds, R. Veeh, F. Hu, P. Stewart, R. Lal, G. Ehrlich, R. Avci, J. Bacteriol. 191 (2009) 2512. [21] B. Park, T. Haines, N. Abu-Lail, Colloids Surf. B: Biointerfaces 73 (2009) 237. [22] P. Eaton, J.C. Fernandes, E. Pereira, M.E. Pintado, F.X. Malacata, Ultramicroscopy 108 (2008) 1128. [23] A. Atabek, T. Camesano, J. Bacteriol. 189 (2007) 8503. [24] V. Vadillo-Rodriguez, et al., J. Bacteriol. 190 (2007) 4225. [25] K. Karrai, R.D. Grober, Appl. Phys. Lett. 66 (1995) 1842. [26] H. Göttlich, R.W. Stark, J.D. Pedarnig, W.M. Heckl, Rev. Sci. Instrum. 71 (2000) 3104. [27] H.P. Rust, M. Heyde, H.J. Freund, Rev. Sci. Instrum. 77 (2006) 043710. [28] D. Ebeling, H. Hölscher, H. Fuchs, B. Anczykowski, U.D. Schwarz, Nanotechnology 17 (2006) S221. [29] J. Otero, L. Gonzalez, G. Cabezas, M. Puig-Vidal, IEEE Trans. Mechatron. (2012) 1. [30] G.Y. Shang, W.H. Qiao, F.H. Lei, J.F. Angiboust, M. Troyon, M. Manfait, Ultramicroscopy 105 (2005) 324. [31] M. Koopman, I. de Bakker, M.F. Garcia-Parajo, N.F. van Hulst, Appl. Phys. Lett. 83 (2003) 5083. [32] Y. Takahashi, H. Shiku, T. Murata, T. Yasukawa, T. Matsue, Anal. Chem. 81 (2009) 9674. [33] V.T. Tung, S.A. Chizhik, T.X. Hoai, J. Eng. Phys. Thermophys. 82 (2009) 140.

[34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64]

[65] [66]

123

M. Hofer, et al., Ultramicroscopy 110 (2010) 605. M.J. Chen, Z. Zhang, T.R. Bott, Colloids Surf. B: Biointerfaces 43 (2005) 61. M. Hegner, P. Wagner, G. Semenenza, Surf. Sci. 291 (1993) 39. M. Arnoldi, M. Fritz, E. Bäuerlein, M. Radmacher, E. Sackmann, A. Boulbitch, Phys. Rev. E 62 (2000) 1034. J. Sader, J. Chon, P. Mulvaney, Rev. Sci. Instrum. 70 (1999) 3967. P. Hoffmann, B. Dutoit, R. Salathé, Ultramicroscopy 61 (1995) 165. U. Hübner, W. Morgenroth, H.G. Meyer, T. Sulzbach, B. Brendel, W. Mirandé, Appl. Phys. A 76 (2003) 913. Y. Qin, R. Reifenberger, Rev. Sci. Instrum. 78 (2007) 063704. T.A. Camesano, M.J. Natan, B.E. Logan, Langmuir 16 (2000) 4563. A.V. Bolshakova, O.I. Kiselyovaa, A.S. Filonov, O.Y. Frolova, Y.L. Lyubchenko, I.V. Yaminsky, Ultramicroscopy 86 (2001) 121. J. Jahng, M. Lee, H. Noh, Appl. Phys. Lett. 91 (2007) 023103. J. Otero, M. Puig-Vidal, G. Cabezas, L. Gonzalez, ES Patent 201031135, July 23, 2010. L. Gonzalez, J. Otero, G. Cabezas, M. Puig-Vidal, Sens. Actuators A 184 (2012) 112. F. Giessibl, Appl. Phys. Lett. 73 (1998) 3956. J. Liu, A. Callegari, M. Stark, M. Chergui, Ultramicroscopy 109 (2008) 81. J. Rychen, T. Ihn, P. Studerus, A. Herrmann, K. nsslin, H. Hug, P. van Schendel, H. Güntherodt, Appl. Surf. Sci. 157 (2000) 290. H. Edwards, L. Taylor, W. Duncan, J. Appl. Phys. 82 (1997) 980. I. Horcas, R. Fernandez, J.M. Gomez-Rodriguez, J. Colchero, J. Gomez-Herrero, A.M. Baro, Rev. Sci. Instrum. 78 (2007) 013705. K. Karrai, I. Tiemann, Phys. Rev. B 62 (2000) 13174. W.F. Heinz, J.H. Hoh, Trends Biotechnol. 17 (1999) 143. K. Johnson, et al., Proc. R. Soc. Lond. A 324 (1971) 301. F. Gaboriaud, B.S. Parcha, M.L. Gee, J.A. Holden, R.A. Strugnell, Colloids Surf. B: Biointerfaces 62 (2008) 206. J. Arfsten, S. Leupold, C. Bradmöller, I. Kampen, A. Kwade, Colloids Surf. B: Biointerfaces 79 (2010) 284. E. A-Hassan, W.F. Heinz, M.D. Antonik, N.P. D’costa, S. Nageswaran, C.A. Schoenenberger, J.H. Hoh, Biophys. J. 74 (1998) 1564. A. Vinckier, G. Semenza, FEBS Lett. 430 (1998) 12. C. Díaz, P.L. Schilardi, R.C. Salvarezza, M.F. de Mele, Colloids Surf. B: Biointerfaces 82 (2011) 536. A. Castellanos-Gomez, N. Agraït, G. Rubio-Bollinger, Nanotechnology 20 (2009) 215502. M. Koopman, B.I. de Bakker, M.F. Garcia-Parajo, N.F. van Hulst, Appl. Phys. Lett. 83 (2003) 5083. M. Barbic, L. Eliason, J. Ranshaw, Sens. Actuators A 136 (1997) 564. T. Ludwig, J. Phys. A: Math. Theor. 41 (2008) 164025. K. Awsiuk, A. Bernasik, M. Kitsara, A. Budkowski, P. Petrou, S. Kakabakos, S. Prauzner-Bechcicki, J. Rysz, I. Raptis, Colloids Surf. B: Biointerfaces 90 (2012) 159. M. Sandberg, A. Määttänen, J. Peltonen, P.M. Vuorela, Int. J. Antimicrob. Agents 32 (2008) 233. M. Müsken, S. Di Fiore, U. Römling, S. Häussler, Nat. Protoc. 5 (2010) 1460.