ARTICLE pubs.acs.org/Langmuir
Non-Invasive Vibrational SFG Spectroscopy Reveals That Bacterial Adhesion Can Alter the Conformation of Grafted “Brush” Chains on SAM Emilie Bulard,*,† Ziang Guo,† Wanquan Zheng,† Henri Dubost,† Marie-Pierre Fontaine-Aupart,† Marie-No€elle Bellon-Fontaine,‡ Jean-Marie Herry,‡ Romain Briandet,‡ and Bernard Bourguignon† † ‡
Institut des Sciences Moleculaires d’Orsay, ISMOCNRS, Universite Paris Sud, B^at. 350 91405 Orsay cedex, France INRA AgroParisTech, UMR 1319 Micalis, 91300 Massy, France ABSTRACT: Understanding bacterial adhesion on a surface is a crucial step to design new materials with improved properties or to control biofilm formation and eradication. Sum Frequency Generation (SFG) vibrational spectroscopy has been employed to study in situ the conformational response of a self-assembled monolayer (SAM) of octadecanethiol (ODT) on a gold film to the adhesion of hydrophilic and hydrophobic ovococcoid model bacteria. The present work highlights vibrational SFG spectroscopy as a powerful and unique non-invasive biophysical technique to probe and control bacteria interaction with ordered surfaces. Indeed, the SFG vibrational spectral changes reveal different ODT SAM conformations in air and upon exposure to aqueous solution or bacterial adhesion. Furthermore, this effect depends on the bacterial cell surface properties. The SFG spectral modeling demonstrates that hydrophobic bacteria flatten the ODT SAM alkyl chain terminal part, whereas the hydrophilic ones raise this ODT SAM terminal part. Microorganism-induced alteration of grafted chains can thus affect the desired interfacial functionality, a result that should be considered for the design of new reactive materials.
1. INTRODUCTION All solidliquid interfaces are potentially subject to bacterial adhesion and biofilm formation. When they concern pathogens, these microbial consortia are involved in nosocomial and food born infections.1,2 Thus, the control of bacterial adhesion to abiotic surfaces remains a current challenge for the design of new materials with improved properties such as functional or brush coatings to prevent and control surface biocontamination3 but also to control biofilms eradication. Bacterial adhesion on an inert surface is largely governed, as for any colloidal particles, by non-covalent molecular interactions (coulombian, van der Waals, Lewis acidbase). These interactions depend on the nature of the bacteria cell envelopes (peptidoglycan, exoproteins, exopolysaccharides, pili, flagela, and so forth) and on the physicochemical properties of the surfaces.4,5 Their macroscopic and microscopic analyses have been largely developed (topography and roughness, hydrophobic/hydrophilic character, acid/base properties, surface potential, and so forth) in order to propose predictive models for bacteria adhesion.68 However, divergences between these models and experimental results have been reported particularly in the case of surfaces functionalized with physisorbed or grafted chains.911 So the question arises whether bacterial adhesion modifies the conformation and the physicochemical properties of these grafted chains or not. To answer this question, it is necessary to use a biophysical technique able to provide information on interactions at the molecular level with selectivity for the solid-bacteria interface under in situ conditions (e.g., in aqueous environment in order to preserve cell viability). r 2011 American Chemical Society
SFG vibrational spectroscopy has recently been applied to biomolecular systems.1214 Its high surface specificity and sensitivity has allowed us to understand interactions at the molecular level between an organized molecular system and a substrate. SFG is also emerging as a tool more adapted than infrared and Raman spectroscopy to study the molecular conformation of SAMs.15 The method was used here for the first time to investigate the reactivity of a hydrophobic model surface (ODT SAM) to the adhesion of a single layer of ovococcoid hydrophilic and hydrophobic model bacterial strains suspended in aqueous solution. SAMs are well suited for such SFG experiments due not only to their high molecular order16 but also to their stability in air and in contact with aqueous surfaces. They are also considered in view of potential application as coatings able to resist bacterial contamination. This work reveals that both water and bacteria interact with the ODT layer and modify its conformation. The hydrophobic cells tend to flatten the SAM terminal groups while the hydrophilic ones have a brush effect stronger than that of an aqueous solution.
2. EXPERIMENTAL SECTION 2.1. Materials. Substrates consisted of borosilicate glass coated with a polycrystalline gold film of 250-nm thickness and annealed in an oven at 600 °C over 30 s. A self-assembled monolayer of octadecanethiol was Received: January 17, 2011 Revised: February 15, 2011 Published: March 16, 2011 4928
dx.doi.org/10.1021/la200205e | Langmuir 2011, 27, 4928–4935
Langmuir
ARTICLE
Figure 1. AFM topographic images (10 10 μm2): Left. gold surface, Right. ODT SAM/gold surface. created on the gold-coated substrate employing the following procedure: the coated surface was dipped in a 1 mM ODT solution in absolute ethanol during 3 h, rinsed in absolute ethanol, and dried under nitrogen flow. As described previously,16 we obtained an ODT SAM film thickness of ∼2 nm under these conditions. Surface topography of these substrates was characterized by Atomic Force Microscopy (AFM) in contact mode (PicoSPM, Molecular Imaging, ScienTec, Palaiseau, France) operating under air at 22 °C. For these experiments, we used a cantilever (silicon nitrides gold-coated oxide-sharpened, ScienTec, Palaiseau, France) with a spring constant of ∼0.38 N 3 m1. Topographic images were acquired at a scanning rate of 1 line/s and for 512 lines/image. The gold surface topography (Figure 1A) presents plantens of 12 μm2 delimitated by grooves of 2030 nm depth. This is typical of Au deposits on glass substrates. It could be improved only using Au single crystals, but this would imply time-consuming processing and very high cost. After dipping into the ODT solution, the ODT SAM/gold support after heating presents the same characteristics as the gold surface alone: the chemical treatment did not alter the surface topography (Figure 1B).
Figure 2. L. lactis after 1 h 30 min adhesion on an ODT SAM/gold surface: (A) SEM image (3 2.5 μm2) in a surface region with a high density of grooves; (B) SEM image showing bacteria on a large gold planten: bacteria do not necessarily adhere on grooves; (C) epifluorescence image (50 50 μm2) of bacterial coverage in aqueous solution; and (D) similar fluorescence image showing that the homogeneity of the bacterial film was spoiled by exposure to air.
2.2. Bacterial Growth Conditions and Characterizations. Bacteria used for the adhesion on ODT SAMs were L. lactis ssp. cremoris strains MG1363 and its mutant PRTPþ expressing the PrtP protease.17,18 Bacteria were stored at 20 °C in M17 broth (Difco) containing 0.5% (w, vol) glucose and 50% (v/v) of glycerol. They were subcultured twice in M17-glucose broth at 30 °C, until a stationary phase was reached. Finally, they were cultivated overnight (working cultures) at 30 °C. About 90 mL of bacteria from final working cultures were harvested by centrifugation (10 min, 7000 g, 4 °C), washed twice, and resuspended in 25 mL of distilled water or potassium nitrate solution (101 M, Sigma-Aldrich) at a final cell density of approximately 109 cells/mL. The main physicochemical difference between the two strains is their cell surface properties: the Lactoccocus lactis MG1363 has a polysaccharide pellicle on its cell wall,17 and the Lactococcus lactis MG1363 PRTPþ has an external envelope with anchored proteinase PRTP.18 The Microbial Adhesion To Solvents (MATS) method19 was employed for the evaluation of the hydrophobic/hydrophilic character of the cell surface of the two L. lactis strains. Experimentally, bacteria in distilled water or in nitrate potassium solution (2.4 mL) were mixed for 120 s at maximum intensity on a vortex-type agitator with 0.4 mL of different apolar solvents (decane and hexadecane). The mixture was allowed to stand for 15 min, to ensure complete separation of the two phases. Then a sample (1 mL) was carefully removed from the aqueous phase and its optical density was measured at 400 nm. The microbial adhesion in each solvent was calculated using the formula: %adhesion ¼ ð1 OD=OD0 Þ100 where OD0 was the optical density of the bacterial suspension before mixing with the solvent and OD the absorbance after mixing and phase separation.
Figure 3. Experimental GaAs reference spectra in air (black dots) and in water (red dots). The liquid water absorption spectrum is also shown (blue line). The green line represents the fit of the spectrum in water obtained from the experimental GaAs spectrum in air corrected by water absorption by a layer of thickness 5 μm. For both solvents, the natural L. lactis strain MG1363 presented a weak affinity (7 ( 5%), whereas the L. lactis strain MG1363 PRTPþ presented an affinity of (90 ( 5) % revealing a stronger hydrophobic character. 2.3. Bacterial Adhesion. To investigate the ODT SAM response to bacterial adhesion, ∼200 μL of the bacteria suspension at 109 cells/ mL in distilled water or in potassium nitrate buffer (101 M) was deposited over the ODT SAM. The solution was allowed to incubate for 90 min and then washed with the suspending fluid to remove the nonadherent bacteria. Care was taken to prevent bacteria from being exposed to air, because this was found to damage the homogeneity of the bacterial deposit (Figure 2). For SFG measurements, a CaF2 plate was added to ensure a uniform thin water layer above the surface sample. Its thickness was ∼5 μm (Figure 3), which is sufficiently low to allow infrared radiation to reach the sample. Morphology of adherent bacteria on ODT SAM was analyzed by Scanning Electron Microscopy (SEM) as previously described.20 As can 4929
dx.doi.org/10.1021/la200205e |Langmuir 2011, 27, 4928–4935
Langmuir
ARTICLE
Figure 4. Experimental (blue dots) and fitted (red lines) SFG spectra and deconvoluted vibrational bands (green lines) of an ODT SAM: (A) in air; (B) in the presence of distilled water; (C) in the presence of a potassium nitrate solution; (D) in the presence of hydrophilic L. lactis bacteria in distilled water after 90 min of adhesion; and (E) in the presence of hydrophobic L. lactis bacteria PRTPþ in distilled water after 90 min of adhesion. The SFG reference spectrum (purple dashed line) is also presented. The experimental SFG scheme is added below. be observed in Figure 2A, Lactococcus lactis cells maintain their characteristic ovococcoid morphology after adhesion on the functionalized gold surface. There is no evidence in our collection of MEB images that bacteria are preferentially localized on the grooves of the film (Figure 2B). This is probably related to the small thickness of the grooves (with respect to the size of bacteria), and also to the coccoid shape of the bacteria (without extracellular appendages such as flagella, etc...) which presumably does not help them to exploit such small inhomogeneities of the surface. It is also important to mention that the SFG spectra can be safely assigned to the flat regions: in SFG, the signal scales as the square of the number of molecules; the grooves represent no more than a few % of the surface, implying that they contribute to less than 0.1% to the SFG spectra. To control the bacterial surface coverage on the ODT SAM in situ, epifluorescence microscopy measurements were performed. Adherent bacteria were stained with the nucleic acid dye acridin orange (0.01% in water) for 15 min in the dark. The dye solution was washed and replaced by pure water before mounting the sample under a Leica DM2 microscope equipped with an Olympus Camedia C5060WZ digital camera. The pictures show that our procedure ensures a homogeneous bacterial deposit (Figure 2C), by contrast to the case where bacteria were exposed to air (Figure 2D). The density of bacteria on the ODT SAM is (5 ( 2) 105 bacteria/mm2 as estimated by numerations of the cells released by 5 min of sonication in NaCl solution (150 mM). Assuming an average bacterium section of 1 μm2, this corresponds to a (50 ( 20) % of the bacterial surface coverage, showing that with the incubation time used, we remain in the first bacterial colonization step. This corresponds to 2200 ODT molecules across the diameter of each bacterium.
2.4. Sum Frequency Generation Vibrational Spectroscopy. Details about our broad band SFG setup can be found in our previously published work.16 Tunable IR pulses (4 μJ, 145 fs and 150 cm1 bandwidth) and “visible” pulses (800 nm, 2 μJ, adjustable duration and bandwidth 16 ps 152.5 cm1) were superimposed on the sample in a collinear copropagating configuration at the incident angle of ∼66° in p polarization (Figure 4). In this experimental geometry, the laser spot size on the sample surface was ∼100 μm. The generated SFG signals were collected during 100 to 300 s to obtain an acceptable signal-to-noise ratio and analyzed by a highresolution detection system (a spectrometer of resolution 0.4 cm1 at 650 nm, equipped with a cooled CCD camera). Vibrational spectra were obtained in the region covered by the IR laser (in this case, the region of CH stretch) with a spectral resolution determined by the width of the visible laser. The vibrational bands are superimposed to a so-called nonresonant background showing the spectral profile of the IR laser and arising in fact from the broad resonant response of the gold surface electronic states. In order to deconvolute the vibrational bands from this nonresonant background, experimental spectra are fitted to the standard formula:21,22 IðωSFG Þ µ gðωIR Þ χNR 3 eij þ
∑v
2 Av ωIR ωv þ iΓ
where g(ωIR) is the IR laser spectral profile recorded on a GaAs reference sample which provides only a non-resonant SFG signal. The term χNR.eij is the Au constant nonresonant response with phase j and Aν, ων, and Γ are the Lorentzian amplitude, frequency, and half width of mode ν, respectively. 4930
dx.doi.org/10.1021/la200205e |Langmuir 2011, 27, 4928–4935
Langmuir
ARTICLE
Table 1. Results of SFG Spectrum Deconvolution (Intensity Ratio R between Symmetric and Asymmetric CH3 Bands, See Text) and Modeling (Φ Angles and CH3 Tilt Angles of A and B Molecules in the Adsorption Model of Ref 15) a parameter R ((0.5)
Φ angle A/B (deg) ((3°)
CH3 tilt angle A/B (deg) ((1.0°)
ODT in air
3.7
242/332
55.1/15.8
ODT in distilled water ODT in KNO3 solution (101 M)
5.8 5.4
250/340 249/339
52.5/11.9 52.5/11.9
ODT with hydrophilic bacteria in distilled water
8.8
260/350
48.9/7.5
ODT with hydrophilic bacteria in KNO3 solution
8.2
260/350
48.9/7.5
ODT with hydrophobic bacteria in distilled water
3.8
238/328
55.8/16.8
ODT with hydrophobic bacteria in KNO3 solution
3.5
237/327
56.6/18.3
support
a
Uncertainties characterize the observed dispersion of the results over ∼10 samples.
Γ is related to the decay of vibrational energy to the surface and it is supposed to be the same for all CH modes. The reference GaAs spectrum is measured in air. Therefore, the IR laser profile g(ωIR) must be corrected from absorption by the water layer that surrounds the bacteria (Figure 3). Taking into account the liquid water absorption spectrum, the fitting procedure provides a typical water layer thickness of ∼5 μm. The phase j is characteristic of the surface electronic structure and may depend on the chemical nature of adsorbates or the medium surrounding the surface. The vibrational modes are expected to have the same relative phases when vibronic coupling is absent as is the case with alkyl chains. χNR and Aν are effective second order non-linear susceptibilities, which include geometric and Fresnel factors and several components of the second order susceptibility tensor, according to the surface symmetry. In our fitting procedure, the well-known CH stretch frequencies are fixed and the width of each stretching mode is the same for all bands. The value of the phase in the different media is also fixed and corresponds to an average obtained over ∼50 experiments. Such a fitting ensures that we cannot obtain multiple solutions by letting too many parameters vary.23 We only admit a phase change from air to water, because the first water layers may electrostatically affect the Au substrate.
3. RESULTS 3.1. ODT SAM in Air and Solvent Environment. Before the ODT SAM was dipped into the bacterial suspension, SFG vibrational spectra in the wavenumber range 28003050 cm1 were collected from the substrate in air, in water and in a 101 potassium nitrate solution using the same beam geometry (Figure 4AC, respectively). Experimental spectra obtained in water and KNO3 solution are similar but significantly different from that measured in air, which was analyzed below by our fitting procedure. It should also be noted that the overall spectrum intensity decreases by ∼50% for aqueous solution with respect to air, which can be attributed to a collection of factors: reflection losses on the CaF2 window (∼22%), absorption and scattering by water. The SFG spectrum of the ODT SAM in air has been previously reported.15 Its deconvolution reveals that the spectrum mainly consists of three strong vibrational bands assigned to symmetric and asymmetric stretching vibrations of the CH3 group (2875 and 2964 cm1, respectively) and the Fermi resonance of the CH3 symmetric mode with two quantas of the CH3 bending mode (2936 cm1) (Figure 4A). The bands belonging to CH2 groups at 2851 and 2918 cm1 (corresponding to the symmetric and asymmetric stretching vibrations) have a very weak SFG intensity. This is usually considered as a proof of the all-trans conformation of the ODT alkyl chain: local centers of symmetry in the middle of CC bonds result in the cancellation of the SFG signals of adjacent CH2 groups. Recent calculations show that the situation is more complex. The weak CH2 band intensities are due to a combination
of local centrosymmetry and orientation: unpaired CH2 dipole moments lie close to the surface plane which results in very small SFG intensity. In addition to these five bands, we have taken into account two weak bands at 2905 and 2973 cm1 in the deconvolution procedure (Figure 4). They were recently assigned to the SCH2 group,24 based on the observation of a systematic deviation between experimental and fitted spectra around 2905 cm1 and on ab initio calculations of the electronic and vibrational structure of alkanethiol molecules. This band was already observed but not assigned.25,26 The relative CH3 band intensities are directly related to the orientation of the CH3 group symmetry axis with respect to the surface normal and are maximal when the transition moment is normal to the metallic surface. For a CH3 axis nearly perpendicular to the surface, the symmetric stretch intensity Isym = (Asym/Γ)2 is maximal because its transition moment is parallel to the symmetry axis, while the doubly degenerate asymmetric stretch intensity, Iasym, is minimal because its transition moments are perpendicular to the symmetry axis. In order to characterize the CH3 intensities, we define a phenomenological parameter, R, equal to the intensity ratio between the symmetric (including the Fermi resonance) and the asymmetric vibrations. R ¼ ðIsym þ IFermi Þ=I asym According to the relative change in the vibrational intensities of the CH3 groups observed in air and in aqueous solution (Figure 4A,B), the corresponding R ratio increases from air (R = 3.7) to water (R = 5.8) while similar R values were obtained for water and KNO3 solution (Table 1). This reveals that in an aqueous environment, the CH3 symmetry axis is closer to the surface normal than in air. It was also found that water has an impact on the phase of the nonresonant response which decreases from 290° to 260°. These variations of R can be related to molecular conformation changes of the ODT chains. For this purpose, SFG spectrum modeling has been performed according to the adsorption model proposed by Bourguignon et al.,15,16 which allows us to calculate the SFG spectrum of adsorbed alkanethiol molecules for any conformation of the alkyl chain. The molecular hyperpolarizability is calculated as the sum of the hyperpolarizabilities of all CH2 and CH3 groups that contribute differently according to their orientation. The hyperpolarizability tensors of the modes of individual CH2 and CH3 groups were calculated ab initio. The simulation takes into account geometrical and optical parameters, such as incidence angles, polarization of laser beams, Fresnel coefficients, and optical indexes. These parameters depend on whether the ODT SAM/gold surface is in air or in contact with an aqueous environment. When the ODT chains are self-assembled, their conformation is constrained in a 4931
dx.doi.org/10.1021/la200205e |Langmuir 2011, 27, 4928–4935
Langmuir
Figure 5. Angles that characterize the ODT molecules. Molecules are all-trans except near the S and two first C atoms, implying that most of the C backbone is planar. Black filled circles represent carbon atoms. Hydrogen atoms are omitted for clarity. The gray filled circle represents the sulfur atom. Θ and Φ are the molecular tilt and twist angles. The directions of the dipole moments of CH3 (symmetric μs3, asymmetric in plane μas3ip and out of plane μas3op) and CH2 nearest to the surface (symmetric μs2 and asymmetric μas2) are also indicated. All unexplicited angles correspond to trans, tetrahedral conformation, except around the S and the two C atoms nearest to the surface. The angles around these three atoms are determined by Θ, Φ, and Ψ1. The methyl tilt angle with respect to the surface normal depends only on Θ and Φ.
nearly all-trans conformation, except near the substrate. The conformation is then characterized by three independent angles: Φ, the “twist” angle characterizing the rotation of the molecular C backbone around the average chain axis and is also related to the orientation of the methyl terminal part of the ODT SAM; Θ, the tilt angle of the all-trans part of the alkyl chain with respect to the surface normal; and Ψ1, the dihedral angle about the C(0)C(1) bond nearest the surface that allows adaptation of the chain geometry to the available adsorption sites (Figure 5). All of the remaining angles that describe the chain geometry are determined by tetrahedral geometry and all-trans conformation of bonds. The angles about the S atom can be calculated from Θ, Φ, and Ψ1. The CH3 tilt angle, i.e., the tilt of the CH3 symmetry axis with respect to the surface normal, depends only on Θ and Φ. Furthermore, it was established that the ODT SAM structure is characterized by the presence of two types of molecular conformation, subscripted A and B.15 These two conformations of the ODT chain grafted on the SAM exposed to air are characterized by ΘA= ΘB = 30°, ψ1A = 90°, ψ1B = 100° and ΦA= 242 ( 3°, ΦB = 332 ( 3°, respectively. The C backbones of the two conformations are perpendicular (ΦB = ΦA þ 90°), which optimizes van der Waals interactions.15 We have first calculated the ODT SAM SFG spectrum in water assuming that only the optical parameters change from air to water (i.e., the ODT conformation is not changed). This simulation fails to reproduce the experimental spectrum in aqueous environment and strongly suggests that the ODT chain conformation is modified.
ARTICLE
Thus, the values of the three angles Φ, ψ1, and Θ must be reconsidered. The interaction energy between the ODT SAM and water is presumably much too weak to disorganize the self-assembly which involves more than 1 eV per ODT molecule due to many intermolecular interactions all along the chains. Therefore, we expect that Θ does not change since it would imply to compress the chains if it decreases, or break the assembly if it increases. This idea is confirmed by the fact that the intensities of the CH2 vibrational bands remain very small in water (Figure 4B) as in air (Figure 4A), suggesting that the C backbone is not affected by interaction with water. Thus, in what follows the changes of the CH3 band intensities and the resulting ratio R are interpreted as a consequence of variations of Φ only. It must be noted that changing Φ implies that Ψ1 must also change in order to maintain the bonding to the appropriate adsorption site on Au, but this does not affect the CH3 SFG response. To summarize our spectrum analysis, we use the standard SFG formula of Section 2.4 to extract the parameters of the Lorentzians that contribute to the spectrum. To avoid unphysical results, we limit the number of fitted parameters: frequencies are fixed at the values of the literature, the phase is the same for all vibrational bands, all spectra recorded in air and in water. Then we compare the fitted relative intensities of the three methyl bands (characterized by the phenomenological parameter R for convenience) with our adsorption model, which calculates SFG spectra as a function of molecular conformation and orientation. To account for the weakness of interactions with bacteria and water, we restrict changes of the molecular conformation to the rotation Φ of the C backbones. Therefore, we extract from the comparison with experiment the value of Φ and the corresponding value of the CH3 axis tilt angle. The best modeling for the SFG spectra measured in water was obtained for the twist angle values ΦA = 250 ( 3° and ΦB = 340 ( 3° (Table 1). This corresponds to a weak decrease of the CH3 tilt angle in water by comparison to air (Table 1): the aqueous environment has a weak but measurable “brush effect” on the ODT SAM (Figure 6.B). Such a slight adjustment of the ODT SAM conformation is consistent with weak interactions between the grafted chains and the solvent and thus confirms the relevance of our adsorption model.15,16 3.2. ODT SAM in the Presence of Lactococcus lactis Adherent Cells. We have investigated the effect of bacteria adhesion on ODT SAM for the hydrophilic and hydrophobic strains, both in water and in potassium nitrate solution (0.1 M). This solution with a high ionic strength allows the neutralization of all charges that could be present on the bacterial surface or on the ODT interface and ensures that only van der Waals interactions occur between bacteria and the ODT CH covalent bonds. For both bacteria, the R ratios obtained in distilled water and in nitrate potassium solution are closely the same (Table 1). This result indicates that there is no repulsive charge effect on the ODT SAMbacteria interaction. Significant differences are observed in vibrational CH3 band intensities between the SFG spectra measured in water and in the presence of hydrophilic Lactococcus lactis bacteria (Figure 4D) and hydrophobic Lactococcus lactis bacteria (Figure 4E). It should also be noted that the overall SFG intensity in the presence of bacteria is slightly smaller than in pure water presumably due to light scattering by bacteria. The observed spectral changes in the SFG spectra could result from a modification of the ODT SAM conformation in the 4932
dx.doi.org/10.1021/la200205e |Langmuir 2011, 27, 4928–4935
Langmuir
ARTICLE
Figure 6. Schematic ODT SAM conformations: (A) in air and (B) in distilled water. Parts C and D show the two extreme possibilities of methyl orientation adjustment allowed by the rotation of the C backbone planes about the average molecular axis. Also, parts C and D correspond ideally to hydrophilic and hydrophobic interactions, respectively, but the amplitude of rotation is in fact restricted by the limited possibility of adjustment of the AuSC bonds. The experiments and modeling show that the amplitude of rotation is only 23° (for the C backbone plane) and 11° (for the methyl tilt angle) from hydrophobic to hydrophilic bacteria.
presence of bacteria, and/or from an additional signal coming from the bacteria themselves. To test this latter hypothesis, we have recorded SFG spectra of bacteria on a bare gold surface. They show only a non-resonant response coming from gold (data not shown), a result which ascertains that bacteria do not contribute to the SFG spectra. The spectral SFG analysis must take into account the fact that bacteria could not cover all the ODT SAM surface. In this case, the deconvolution sums up the contribution of “free” ODT in water (Figure 4B) and that of ODT in contact with bacteria. The experimental spectra are fitted to our model as a function of the bacterial coverage ranging from 30% to 100%. The best fits were obtained for an ODT SAM coverage of ∼60% for each strain of bacteria (Figures 4D,E), in good agreement with the experimental coverage values estimated from epifluorescence and SEM images. In this condition, the deconvolution of the SFG spectra reveals that hydrophilic bacterial adhesion results in an increase of the R ratio by comparison to the ODT interface exposed to distilled water (Table 1). The corresponding twist angles determined from the SFG spectra modeling are ΦA = 260 ( 5° and ΦB = 350 ( 5°. Consequently, the methyl tilt angle decreased by 3.6° and 4.4° for A and B molecules, respectively (Table 1) in comparison to the ODT conformations in water. This analysis demonstrates that hydrophilic bacteria raise up the ODT SAM methyl groups and decrease the exposure of the CH2 groups (near the methyl ones) to bacteria (Figure 6C). By contrast, hydrophobic bacterial adhesion results in a decrease of the R ratio, of the twist angles (Table 1) and in an increase in the methyl tilt angle (3.3° and 4.9° for A and B molecules, respectively) with respect to water. The effect of hydrophobic bacteria is to flatten the ODT SAM methyl groups and to increase the exposure of the neighboring CH2 groups to the environment (Figure 6D).
4. DISCUSSION AND CONCLUSIONS The present work demonstrates that water and bacteria interact weakly with the chemical groups on the top of the chains (the methyl and to a small extent the next CH2 group). These interactions result in a slight adjustment of the twist angles of the grafted chains as illustrated in Figure 6. The SAM organization remains largely controlled by stronger intermolecular interactions along the
ODT chains and by chemisorption to the gold surface: the basic organization, including the tilt angle is unchanged in the presence of water or bacteria. Furthermore, the behavior of the ODT molecules depends on the hydrophobic or hydrophilic character of the molecules covering the ODT SAM. The fact that methyl groups tend to point up when exposed to water or hydrophilic bacteria and down when exposed to hydrophobic bacteria is difficult to interpret quantitatively because the SAM-bacteria interface involves fluctuating weak interactions between numerous molecular species: ODT, complex biological molecules attached to bacterial envelopes and water. However, it can be rationalized from the general ideas that have emerged from the understanding of the solvation processes of nonpolar (hydrophobic) and polar (hydrophilic) molecules.27 According to the fact that phase separation is more favorable than mixing, the interface area between hydrophobic and hydrophilic phases is minimized and the number of hydrogen bonds between polar molecules in the hydrophilic phase is maximized. This induces a local order near the interface due to a hydrogen bonding network. The interaction between non-polar molecules is maximized by increasing the contact area between them (as occurs within the SAM for example). These ideas can be translated to the case of a non-polar SAM upon exposure to distilled water or to a KNO3 solution. The formation of an ordered layer of water molecules on the top of the ODT SAM results in an entropy decrease, i.e., an increase in the Gibbs free energy of the system ODT SAMsolution.27 This entropy penalty is minimized by straightening the methyl terminal part of the alkyl chain which limits the number of water molecules in contact with the hydrophobic ODT SAM. It is also compensated by the increase in the number of hydrogen bonds between them: this decreases the internal energy and so the system free energy. The brush effect of aqueous environment onto the ODT SAM achieves a new equilibrium of the system. In the case of bacteria, it is more difficult to describe the interaction because the involved molecules are much more complex and are attached to the bacterial envelope. The experimental results show that hydrophilic bacteria, like water, raise the methyl terminal part of the grafted chains. Thus, the minimization of the interface area remains the driving force. An unexpected result is that the raising of the ODT SAM effect observed in the presence of bacteria is stronger than that in presence of water. This precludes the 4933
dx.doi.org/10.1021/la200205e |Langmuir 2011, 27, 4928–4935
Langmuir presence of many water molecules between the SAM and the bacteria: otherwise, the SFG response to the hydrophilic bacteria would be the same as that of water. Contrary to the hydrophilic strain, hydrophobic bacteria are observed to flatten the methyl terminal part of the ODT SAM. This facilitates interactions of the bacteria with the methylene group next to the methyl one. Flattening the top of the ODT chains thus maximizes attractive interactions between bacteria and the SAM. As such, the present work demonstrates the high sensitivity of vibrational SFG spectroscopy to probe in situ the effect of bacterial adhesion on a functionalized self-assembled monolayer. By monitoring and modeling the spectral intensity changes of the CH stretching modes, we are able to relate the SFG spectral changes to a flattening or a rising of the terminal bond of the alkyl chain when it interacts with hydrophobic or hydrophilic bacteria, respectively. The main result of this work is the direct evidence at the molecular level that even in a case of small interactions, bacteria adhesion onto an ordered molecular film results in measurable changes of the film conformation. Such unanticipated effects on the grafted chain conformation in the presence of bacteria suggest that a general understanding of the bacteriamolecular film interaction requires one to probe the film at the molecular level. According to these fundamental results, the inefficiency of grafted surfaces to prevent bacterial adhesion911 and/or to improve bacterial decontamination could be explained by a modification of the grafted surface by bacteria and so the alteration of the desired functionality of molecular films at the first stage of bacterial adhesion. We must keep in mind that the presence of proteins in the surrounding medium and at a later stage of biofilm formation, proteins emitted by the bacteria themselves will further modify the properties of the substrate.28,29 This is now under investigation and should be considered for the design of new materials.
’ AUTHOR INFORMATION Corresponding Author
*E-mail:
[email protected].
’ ACKNOWLEDGMENT The authors wish to thank the MIMA2 platform (INRA, Massy, France) for allowing us to use the scanning electron microscope and the CPBM platform of LUMAT (Orsay, France) for the use of their biological infrastructures. ’ REFERENCES (1) Lindsay, D.; von Holy, A. Bacterial biofilms within the clinical setting, what healthcare professionals should know. J. Hosp. Infect. 2006, 64, 313–325. (2) Wong, A. C. L. Biofilms in food processing environments. J. Dairy Sci. 1998, 81, 2765–2770. (3) Nejadnik, M, R.; van der Mei, H.; Norde, W.; Busscher, H. J. Bacterial adhesion and growth on a polymer brush-coating. Biomaterials 2008, 29, 4117–4121. (4) Giaouris, E.; Chapot-Chartier, M.-P.; Briandet, R. Surface physicochemical analysis of natural Lactococcus lactis strains reveals the existence of hydrophobic and low charged strains with altered adhesive properties. Int. J. Food Microbiol. 2009, 131, 2–9. (5) Speranza, G.; Pederzolli, C.; Lunelli, L.; Canteri, R.; Pasquardini, L.; Carli, E.; Lui, A.; Maniglio, D.; Brugnara, M.; Anderle, M. Role of
ARTICLE
chemical interactions in bacterial adhesion to polymer surfaces. Biomaterials 2004, 25, 2029–2037. (6) Wiencek, K. M.; Fletcher, M. Bacteria adhesion to hydroxyl- and methyl- terminated alkanethiol self-assembled monolayers. J. Bacteriol. 1995, 177, 1959–1966. (7) Boonaert, C. J. P.; Dufr^ene, Y. F.; Derclaye, S. R.; Rouxhet, P. G. Adhesion of Lactococcus lactis to model substrata, direct study of the interface. Colloids Surf., B 2001, 22, 171–182. (8) Vacheethasanee, K.; Marchant, R. E. Surfactant polymers designed to suppress bacterial (Staphylococcus epidermidis) adhesion on biomaterials. J. Biomed. Mater. Res. 2000, 50, 302–312. (9) Cheng, G.; Zhang, Z.; Chen, S.; Bryers, J. D.; Jiang, S. Inhibition of bacterial adhesion and biofilm formation on zwitterionic surfaces. Biomaterials 2007, 28, 4192–4199. (10) Zdyrko, B.; Klep, V.; Li, X.; Kang, Q.; Minko, S.; Wen, X.; Luzinov, I.. Polymer brushes as active nanolayers for tunable bacteria adhesion. Mater. Sci. Eng., C 2009, 29, 680–684. (11) Sivakumar, P. M.; Iyer, G.; Natesan, L.; Doble, M. 30 -Hydroxy4-methoxychalcone as a potential antibacterial coating on polymeric biomaterials. Appl. Surf. Sci. 2010, 256, 6018–6024. (12) Chen, X.; Tang, H.; Even, M. A.; Wang, J.; Tew, G. N.; Chen, Z. Observing a molecular knife at work. J. Am. Chem. Soc. 2006, 128, 2711–2714. (13) Koffas, T. S.; Kim, J.; Lawrence, C. C.; Somorjai, G. A. Detection of immobilized protein on latex microspheres by IR-Visible sum frequency generation and scanning force microscopy. Langmuir 2003, 19, 3563–3566. (14) Wang, J.; Buck, S. M.; Chen, Z. J. Sum frequency generation vibrational spectroscopy studies on protein adsorption. J. Phys. Chem. B 2002, 106, 11666–11672. (15) Bourguignon, B.; Zheng, W.; Carrez, S.; Ouvrard, A.; Fournier, F.; Dubost, H. Deriving the complete molecular conformation of selfassembled alkanethiol molecules from sum-frequency generation vibrational spectra. Phys. Rev. B 2009, 79, 125433. (16) Guo, Z.; Zheng, W.; Hamoudi, H.; Dablemont, C.; Esaulov, V. A.; Bourguignon, B. On the chain length dependence of CH3 vibrational mode relative intensities in sum frequency generation spectra of self assembled alkanethiols. Surf. Sci. 2008, 602, 3551–3559. (17) Chapot-Chartier, M. P.; Vinogradov, E.; Sadovskaya, I.; Andre, G.; Mistou, M. Y.; Trieu-Cuot, P.; Furlan, S.; Bidnenko, E.; Courtin, P.; Pechoux, C.; Hols, P.; Dufr^ene, Y. F.; Kulakauskas, S. Cell surface of Lactococcus lactis is covered by a protective polysaccharide pellicle. J. Biol. Chem. 2010, 285 (14), 10464–10471. (18) Habimana, O.; Le Goff, C.; Juillard, V.; Bellon-Fontaine, M.-N.; Buist, G.; Kulakauskas, S.; Briandet, R. Positive role of cell wall anchored proteinase PrtP in adhesion of Lactococci. BMC Microbiol. 2007, 7, 36. (19) Bellon-Fontaine, M.-N.; Rault, J.; van Oss, C. J. Microbial adhesion to solvents, a novel method to determine the electrondonor/electron-acceptor or Lewis acid-base properties of microbial cells. Colloids Surf., B 1996, 7, 47–53. (20) Kamgang, J. O.; Briandet, R.; Herry, J.-M.; Brisset, J. L.; Naïtali, M. Destruction of planktonic, adherent and biofilm cells of Staphylococcus epidermidis using a gliding discharge in humid air. J. Appl. Microbiol. 2007, 103, 621–628. (21) Yeganeh, M. S.; Dougal, S. M.; Polizzotti, R. S.; Rabinowitz, P. Interfacial atomic structure of a self-assembled alkyl thiol monolayer/Au (111), a sum frequency generation study. Phys. Rev. Lett. 1995, 74, 1811–1814. (22) Superfine, R.; Huang, J. Y.; Shen, Y. R. Nonlinear optical studies of the pure liquid/vapor interface, vibrational spectra and polar ordering. Phys. Rev. Lett. 1991, 66, 1066–1069. (23) Busson, B.; Tadjeddine, A. Non-uniqueness of parameters extracted from resonant second-order nonlinear optical spectroscopies. J. Phys. Chem. C 2009, 113, 21895–21902. (24) Bulard, E.; Zheng, W.; Bourguignon, B. Surf. Sci., in press. (25) Ron, H.; Cohen, H.; Matlis, S.; Rappaport, M.; Rubinstein, I. Self assembled monolayers on oxidized metals. 4. Superior n-alkanethiol monolayers on copper. J. Phys. Chem. B 1998, 102, 9861–9869. 4934
dx.doi.org/10.1021/la200205e |Langmuir 2011, 27, 4928–4935
Langmuir
ARTICLE
(26) Ruan, C. M.; Bayer, T.; Meth, S.; Sukenik, C. N. Creation and characterization of n-alkylthiol and n-alkylamine self-assembled monolayers on 316L stainless steel. Thin Solid Films 2002, 419, 95–104. (27) Chottard, J.-C., Depezay, J. C.; Leroux, J. P.Chimie Fondamentudes Biologiques Et Medicales, II Structure Moleculaire; Hermann: tale, E Paris, 1995, pp 214218. (28) Keselowsky, B. G.; Collard, D. M.; Garcia, A. J. Surface chemistry modulates fibronectin conformation and directs integrin binding and specificity to control cell adhesion. J. Biomed. Mater. Res. 2003, 66A, 247–259. (29) Keselowsky, B. G.; Collard, D. M.; Garcia, A. J. Surface chemistry modulates focal adhesion composition and signaling through changes in integrin binding. Biomaterials 2004, 25, 5947–5954.
4935
dx.doi.org/10.1021/la200205e |Langmuir 2011, 27, 4928–4935