Characterization of the Cell Surface and Cell Wall Chemistry of

Feb 27, 2008 - Frederick Mappin Building, The UniVersity of Sheffield, Mappin Street, Sheffield S1 3JD, United Kingdom. ReceiVed July 27, 2007. In Fin...
4 downloads 0 Views 233KB Size
Download PDF
4032

Langmuir 2008, 24, 4032-4040

Characterization of the Cell Surface and Cell Wall Chemistry of Drinking Water Bacteria by Combining XPS, FTIR Spectroscopy, Modeling, and Potentiometric Titrations Jesu´s J. Ojeda,† Marı´a E. Romero-Gonza´lez,*,† Robert T. Bachmann,‡ Robert G. J. Edyvean,‡ and Steven A. Banwart† Cell-Mineral Interface Research Programme, Kroto Research Institute, The UniVersity of Sheffield, Broad Lane, Sheffield S3 7HQ, United Kingdom, and Department of Chemical and Process Engineering, Sir Frederick Mappin Building, The UniVersity of Sheffield, Mappin Street, Sheffield S1 3JD, United Kingdom ReceiVed July 27, 2007. In Final Form: NoVember 20, 2007 Aquabacterium commune, a predominant member of European drinking water biofilms, was chosen as a model bacterium to study the role of functional groups on the cell surface that control the changes in the chemical cell surface properties in aqueous electrolyte solutions at different pH values. Cell surface properties of A. commune were examined by potentiometric titrations, modeling, X-ray photoelectron spectroscopy (XPS), and Fourier transform infrared (FTIR) spectroscopy. By combining FTIR data at different pH values and potentiometric titration data with thermodynamic model optimization, the presence, concentration, and changes of organic functional groups on the cell surface (e.g., carboxyl, phosphoryl, and amine groups) were inferred. The pH of zero proton charge, pHzpc ) 3.7, found from titrations of A. commune at different electrolyte concentrations and resulting from equilibrium speciation calculations suggests that the net surface charge is negative at drinking water pH in the absence of other charge determining ions. In situ FTIR was used to describe and monitor chemical interactions between bacteria and liquid solutions at different pH in real time. XPS analysis was performed to quantify the elemental surface composition, to assess the local chemical environment of carbon and oxygen at the cell wall, and to calculate the overall concentrations of polysaccharides, peptides, and hydrocarbon compounds of the cell surface. Thermodynamic parameters for proton adsorption are compared with parameters for other gram-negative bacteria. This work shows how the combination of potentiometric titrations, modeling, XPS, and FTIR spectroscopy allows a more comprehensive characterization of bacterial cell surfaces and cell wall reactivity as the initial step to understand the fundamental mechanisms involved in bacterial adhesion to solid surfaces and transport in aqueous systems.

1. Introduction The microbial population of drinking water is extensively monitored worldwide; however, the majority of measurements only focus on indicator organisms without taking into account the autochthonous population of this nutrient-deprived ecosystem.1 One of the reasons for this, apart from legislative requirements, is that the true composition of microbial communities in the drinking water systems is far from being completely understood. Up to 99.9% of the bacterial cells were unlikely to be cultured on standard media and remained undetected until new approaches were made to examine the population structure.2 Now, the application of molecular tools such as 16S rRNA gene sequencing and the development and application of highly specific oligonucleotide probes have been successfully applied to the drinking water systems in Berlin, Hamburg, Mainz (Germany), and Stockholm (Sweden).3 Analysis of drinking water biofilms showed that three of the isolated strains dominated the autochthonous biofilm population of the Berlin drinking water distribution system: Aquabacterium citratiphilum, Aquabacterium parVum, and Aquabacterium commune.4 All members of the * To whom correspondence should be addressed. E-mail: [email protected]. † Cell-Mineral Interface Research Programme, Kroto Research Institute. ‡ Department of Chemical and Process Engineering. (1) Bachmann, R. T.; Edyvean, R. G. J. Biofilms 2005, 2, 197-227. (2) Wimpenny, J.; Manz, W.; Szewzyk, U. FEMS Microbiol. ReV. 2000, 24, 661-671. (3) Kalmbach, S.; Manz, W.; Bendinger, B.; Szewzyk, U. Water Res. 2000, 34, 575-581. (4) Kalmbach, S.; Manz, W.; Wecke, J.; Szewzyk U. Int. J. Syst. Bacteriol. 1999, 49, 769-777.

genus are rod-shaped Gram-negative bacteria, are motile by means of single polar monotrichous flagella, and contain polyalkalonate and polyphosphate inclusion bodies.4 Although identified, the interactions between these bacteria and the substratum (such as pipe walls) and the bulk water phase, as well as the ability to form biofilms and colonize metal surfaces are far from clear. This microbial attachment and growth on solid surfaces (i.e., stainless steel or copper water pipes) may cause a deterioration of the water taste and water discoloration,5-7 the harboring and protection of pathogenic bacteria,8,9 and the immobilization of particulate matter and heavy metals10-12 and influence the corrosion processes of metal pipes.13-15 To reduce (5) Bays, L. R.; Burman, N. P.; Lewis, W. M. Water Treat. Exam. 1970, 19, 136-160. (6) HMSO. The Microbiology of Water 1994 Part I - Drinking Water; Report on Public Health and Medical Subjects No. 71; HMSO Books: London, 1994. (7) Ratsack, U. Neue DELIWA-Z. 1997, 5, 195-198. (8) LeChevallier, M. W.; Babcock, T. M.; Lee, R. G. Appl. EnViron. Microbiol. 1987, 53, 2714-2724. (9) Camper, A. K.; Jones, W. L.; Hayes, J. T. Appl. EnViron. Microbiol. 1996, 62, 4014-4018. (10) Percival, S.; Knapp, J. S.; Edyvean, R. G. J.; Wales, D. S. Water Res. 1998, 32, 2187-2201. (11) Gauthier, V.; Ge´rard, B.; Portal, J.-M.; Block, J.-C.; Gatel, D. Wat. Res. 1999, 33, 1014-1026. (12) Zacheus, O. M.; Lehtola, M. J.; Korhonen. L. K.; Martikainen, P. J. Wat. Res. 2001, 35, 1757-1765. (13) Beech, I. B.; Edyvean, R. G. J.; Cheung, C. W. S.; Turner, A. In Microbial Corrosion, Proceedings of the 3rd International EFC Workshop, Portugal, 1994; Tiller, A. K., Sequeira, C. A. C., Eds.; Institute of Materials: London, U.K., 1995; pp 328-337. (14) Kobrin, G.; Lamb, S.; Tuthill, A. H.; Avery, R. E.; Selby, K. A. Microbiologically influenced corrosion of stainless steels by water used for cooling and hydrostatic testing, 1997 International Water Conference, Pittsburgh, PA, NiDI Technical Series No. 10 085. (15) Keevil, C. W. Water Sci. Tech. 2004, 49, 91-98.

10.1021/la702284b CCC: $40.75 © 2008 American Chemical Society Published on Web 02/27/2008

Cell Surface and Cell Wall Chemistry of A. commune

the undesired attachment of microorganisms to solid surfaces in aqueous environments, the initial stage of bacterial adhesion must be minimized. Therefore, the fundamental mechanisms involved in the initial steps of bacterial adhesion to solid surfaces and transport in aqueous systems need to be established. Several theoretical approaches have been applied to describe bacterial adhesion to solid surfaces, such as the classical Derjaguin-Landau-Verwey-Overbeek (DLVO) theory,16-18 the extended DLVO theory,19,20 and the thermodynamic approach (surface Gibbs energy).21,22 However, some important biological factors, such as bacterial surface heterogeneity, have been largely ignored in the theoretical models mentioned above. For example, several studies by Walker et. at.23-27 have found that the heterogeneity of active sites from cell surface macromolecules, such as proteins and lipopolysaccharide-associated functional groups, controls the bacterial adhesive characteristics. The development of more accurate cell adhesion models by the inclusion of new parameters (such as co-adhesion, time dependent changes of proteins, localized hydrophobic and hydrophilic interactions, or bacterial surface changes in relation to nutrients or the environment) is a challenge for further studies, and improved conceptualization and parametrization depends strongly on a thorough understanding of the bacterial cell surface properties. Moreover, the improvement of quantitative measurement of bacterial transport and attachment efficiency, particularly noninvasive and nondestructive approaches in aquatic systems (see, e.g., ref 28), relies to a large extent on a thorough understanding of the bacterial cell wall surface and reactivity. A growing number of macroscopic acid-base potentiometric studies to characterize protonation behavior have been made in recent years (see, e.g., refs 29-33); however, acid-base titrations and values for proton dissociation constants cannot be used alone to establish the identity of surface ligands. More unequivocal identification of the functional groups responsible for the acidbase buffering capacity by means of spectroscopic techniques is often suggested along with the potentiometric measurements but is less often included in such studies. Infrared spectroscopy is a well-established technique to identify the presence or absence of functional groups, their protonation, or changes in coordination (16) van Loosdrecht, M. C.; Lyklema, J.; Norde, W.; Zehnder, A. J. B. Microb. Ecol. 1988, 17, 1-15. (17) Marshall, K. C.; Stout, R.; Mitchell, R. J. Gen. Microbiol. 1971, 68, 337-348. (18) Rutter, P. R.; Vincent, B. In Microbial adhesion and aggregation; Marshall, K. C., Ed.; Springer-Verlag; Berlin, 1984. (19) Meinders, H.; van der Mei, H. C.; Busscher, H. J. J. Colloid Interface Sci. 1995, 176, 329-341. (20) van Oss, C. J. Cell Biophys. 1989, 14, 1-16. (21) Absolom, D. R.; Lamberti, F. V.; Policova, Z.; Zingg, W.; van Oss, C. J.; Neumann, A. W. Appl. EnViron. Microbiol. 1983, 46, 90-97. (22) Busscher, H.; Weerkamp, A.; van der Mei, H.; van Pelt, A.; De Jong, H.; Arends, J. Appl. EnViron. Microbiol. 1984, 48, 980-983. (23) Redman, J. A.; Walker, S. L.; Elimelech, M. EnViron. Sci. Technol. 2004, 38, 1777-1785. (24) Walker, S. L.; Redman, J. A.; Elimelech, M. Langmuir 2004, 20, 77367746. (25) Walker, S. L.; Redman, J. A.; Elimelech, M. EnViron. Sci. Technol. 2005, 39, 6405-6411. (26) Walker, S. L.; Hill, J. E.; Redman, J. A.; Elimelech, M. Appl. EnViron. Microbiol. 2005, 71, 3093-3099. (27) Walker, S. L. Colloids Surf., B 2005, 45, 181-188. (28) Bridge, J. W.; Banwart, S. A.; Heathwaite, A. L. EnViron. Sci. Technol. 2006, 40, 5930-5936. (29) Ngwenya, B.; Sutherland, I.; Kennedy, L. Appl. Geochem. 2003, 18, 527-538. (30) Haas, J. Chem. Geol. 2004, 209, 67-81. (31) Plette, A. C. C.; van Riemsdijk, W. H.; Benedetti, M. F.; Van de Wal, A. J. Colloid Interface Sci. 1995, 173, 354-363. (32) Fein, J. B.; Boily, J.; Yee, N.; Gorman-Lewis, D.; Turner, B. F. Geochim. Cosmochim. Acta 2005, 69, 1123-1132. (33) Borrok, D.; Fein, J. B.; Kulpa, C. F. Geochim. Cosmochim. Acta 2004, 68, 3231-3238.

Langmuir, Vol. 24, No. 8, 2008 4033

states.34-36 X-ray photoelectron spectroscopy (XPS) provides a direct chemical characterization of the surface layer (2-5 nm depth), and it has been applied to investigate numerous microorganisms, proteins, and extracellular polymeric substances (EPS) (see, e.g., refs 37-46). The binding energy measured by XPS is influenced by the electron density resulting from the chemical bonds between the element of interest and the neighbor atoms. Based on the chemical shifts in a component peak, the chemical state of the element can be inferred (i.e., chemical bonding environment or multivalent states of individual elements can be obtained). XPS analysis has also been employed to estimate approximate concentrations of the basic constituents of the cell surface: polysaccharides, peptides, and hydrocarbon compounds.37,40 However, XPS and infrared spectroscopy cannot be used to study the thermodynamic stability of surface complexes as precisely as potentiometric measurements, which have high precision for proton and hydroxide ion concentrations to low detection limits. In this work, the cell surface properties of the recently isolated gram-negative Aquabacterium commune, such as surface charge, acid-base behavior, chemical composition, and changes of the organic functional groups of the cell wall at different pH values, were characterized. The combination of potentiometric titrations, modeling, XPS, and Fourier transform infrared (FTIR) spectroscopy allowed a more comprehensive characterization of the bacterial cell surface and cell wall reactivity, which is paramount in the understanding of the interactions between bacteria and their environment and the elucidation of the mechanisms involved in bacterial transport, aggregation, and biofilm formation.47 2. Materials and Methods 2.1. Preparation and Purification of A. commune. A. commune (DSMZ-11901) was grown in modified R2A medium4,48 containing (per liter of water) the following: 0.5 g of yeast extract, 0.5 g of proteose peptone no. 3, 0.5 g of casamino acid, 0.3 g of sodium pyruvate, 1.2 mL of Tween 80, 0.3 g of K2HPO4, and 0.1 g of MgSO4‚7H2O. The modified R2A medium was sterilized by autoclaving at 121 °C for 20 min. For agar cultures, 15 g of agar per liter was added before autoclaving. All chemicals were purchased from Sigma and Fisher, U.K. A. commune cultures were grown at room temperature in 5 L of autoclaved modified R2A medium under stirring. Bacterial growth was monitored by measuring the optical density by light absorption in a WPA Lightwave II UV/vis spectrophotometer at a wavelength of 600 nm. Cells were harvested at the late-log growth phase by centrifugation at 4200 rpm (3000g) for 30 min at 10 °C in a Beckman (34) Skoog, D. A.; Leary, J. J. Principles of Instrumental Analysis; Saunders College: Philadelphia, 1992. (35) Conley, R. T. Infrared Spectroscopy; Allyn and Bacon: Boston, 1972. (36) Wade, L. G. Organic chemistry; Prentice Hall: Upper Saddle River, NJ, 1995. (37) van der Mei, H. C.; de Vries, J.; Busscher, H. J. Surf. Sci. Rep. 2000, 39, 1-24. (38) Rouxhet, P. G.; Mozes, N.; Dengis, P. B.; Dufreˆne, Y. F.; Gerin, P. A.; Genet, M. J. Colloids Surf., B 1994, 2, 347-369. (39) Dufreˆne, Y. F.; Rouxhet, P. G. Colloids Surf., B 1996, 7, 271-279. (40) Dufreˆne, Y. F.; Van der Wal, A.; Norde, W.; Rouxhet, P. G. J. Bacteriol. 1997, 179, 1023-1028. (41) Yan, L.; Marzolin, C.; Terfort, A.; Whitesides, G. M. Langmuir 1997, 13, 6704-6712. (42) Beech, I. B.; Zinkevich, V.; Tapper, R.; Gubner, R.; Avci, R. J. Microbiol. Methods 1999, 36, 3-10. (43) Bruinsma, G. M.; van der Mei, H. C.; Busscher, H. J. Biomaterials 2001, 22, 3217-3224. (44) Omoike, A.; Chorover, J. Biomacromolecules 2004, 5, 1219-1230. (45) Pradier, C. M.; Rubio, C.; Poleunis, C.; Bertrand, P.; Marcus, P.; Compe`re, C. J. Phys. Chem. B 2005, 109, 9540-9549. (46) Deng, S.; Ting, Y. P. Langmuir 2005, 21, 5940-5948. (47) Eboigbodin, K.; Ojeda J. J.; Biggs, C. Langmuir 2007, 23, 6691-6697. (48) Bachmann, R. T.; Edyvean, R. G. J. Int. Biodeterior. Biodegrad. 2006, 112-118.

4034 Langmuir, Vol. 24, No. 8, 2008

Ojeda et al.

centrifuge. The biomass was washed by resuspending it in 18 MΩ water, vortexing it, and then centrifuging it in a Hermle Z231M microcentrifuge at 8000 rpm (5000g) for 10 min at 10 °C three times. A Bradford assay49 was also carried out to check for the presence of proteins resulting from the leakage of cellular contents due to lysed cells. After the washing procedure, the intact bacterial cells were imaged by atomic force microscopy (AFM) using a Dimension 3100 (Digital Instrument) instrument in the tapping mode and using a silicon tip. The images were processed using the Dimension 3100 software “Nanoscope” (see Figure S1 in the Supporting Information for an AFM image of Aquabacterium commune). 2.2. Potentiometric Titration. Potentiometric experiments were carried out according to procedures previously published in recent years.29-33 The titrations of biomass suspensions were carried out at 25 °C using NaClO4 as the background electrolyte. The NaClO4 solution was prepared by dissolving a known amount of NaClO4 (Riedel-de Hae¨n, Germany) in degassed ultrahigh quality (UHQ) water. The 0.1 M NaOH and 0.1 M HCl solutions were prepared from NaOH (Fisher, U.K.) and HCl (BDH, U.K.) using UHQ water, and the exact concentration was determined prior to the titration against primary standards. The harvested cells were resuspended in 25 mL of NaClO4 electrolyte (0.01, 0.1, or 1.0 M), and the suspension was purged with N2 (>99.99%) for 2 h to exsolve CO2 before initiating titration, yielding approximate pH values of 4-5. Following the degassing procedure, a positive pressure of N2 was maintained by allowing a gentle flow of N2 into the headspace during the titration. The suspensions were acidified to pH ≈ 3.5 using 0.1 M HCl and then titrated to pH ≈ 10 using 0.1 M NaOH. A blank without biomass was also titrated. A Bradford assay49 was carried out before and after the titration procedure (pH ) 3.5 and pH ) 10). To assess reversibility and protonation behavior, a reverse acidimetric titration was carried out following the base addition. All experiments were conducted in triplicate. All titrations were performed in a glass vessel with a lid as part of a Metrohm 718 STAT-Titrino instrument at 25 °C. The titrator was set to add successive acid or base when the absolute value of the potential drift was equal or less than 5 mV/min (0.1 mV/sec). The electrode was standardized on a proton concentration scale, [H+], and the slope deviation from the theoretical Nernst value was always within 1%. The concentration of deprotonated sites per mass of bacteria, [H+]cons/rel (mol/g), was calculated according to Fein et al.32 as follows: [H+]cons/rel ) (Ca - Cb - [H+] + [OH-])/mb

(1)

where Ca and Cb are the molar concentrations of acid and base, respectively, added at each step of the titration, brackets represent molar species concentrations, and mb is the bacterial dry weight suspension concentration (g/L). 2.3. Infrared Spectroscopy. Bacterial suspensions of A. commune in 0.1 M NaClO4 at different pH values were monitored using a liquid flow cell (Sigma-Aldrich) with CaF2 windows on a PerkinElmer Spectrum-One FTIR spectrophotometer. A total of 100 scans with a resolution of 4 cm-1 and wavenumber range from 4000 to 950 cm-1 were collected for each pH value. At the end of the flow cell study, the biomass was transferred (wet) onto a germanium crystal to obtain an attenuated total reflectance (ATR)-FTIR spectrum. ATR-FTIR measurements were performed on a Specac Silver Gate Essential Single Reflection ATR system attached to the FTIR spectrophotometer, consisting of a germanium crystal at a fixed angle of 45°. A total 100 scans with a resolution of 4 cm-1 and wavenumber range from 4000 to 950 cm-1 were collected for the sample. 2.4. X-ray Photoelectron Spectroscopy. For XPS analysis, the bacterial samples were frozen at -20 °C and freeze-dried following procedures previously described.37-40,42 The obtained powder was mounted on standard sample studs using double sided adhesive tape. (49) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.

Figure 1. (a) Typical potentiometric titration data for 3.2-4.1 g/L (dry weight) suspensions of Aquabacterium commune compared with the 0.1 M NaClO4 electrolyte. Closed symbols correspond to the forward titration data, and open symbols correspond to the back titration. (b) Fitting of three-site constant capacitance model (CCM) and nonelectrostatic model (NEM) to raw titration data. XPS measurements were made on a KRATOS AXIS 165 Ultra Photoelectron spectrometer at 10 kV and 20 mA using the Al KR X-ray source (1486.6 eV). The takeoff angle was fixed at 90°. On each sample, the data were collected from three randomly selected locations, and the area corresponding to each acquisition was 400 µm in diameter. Each analysis consisted of a wide survey scan (pass energy 160 eV, 1.0 eV step size) and high-resolution scan (pass energy 20 eV, 0.1 eV step size) for component speciation. All experiments were conducted in triplicate. The binding energies of the peaks were determined using the C1s peak at 284.5 eV. The software CasaXPS 2.3.1250 was used to fit the XPS spectra peaks. No constraint was applied to the initial binding energy values, and the full width at half-maximum (fwhm) was maintained constant for the carbon contributions in a particular spectrum.

3. Results and Discussion 3.1. Potentiometric Titrations. The bacterial suspensions exhibited proton adsorption/desorption over the pH range studied. The potentiometric titration data of A. commune in 0.1 M NaClO4 and the electrolyte without bacteria (blank) are shown in Figure 1a. During the whole titration, no evidence of saturation was observed with respect to proton adsorption. It is evident from the figure that the bacteria provided significant buffering capacity to the solution over the selected pH range when compared with the blank. This buffering capacity is due to functional groups on the bacterial surface consuming the added base by donating protons or the added acid by accepting protons. The shape of the titration curves obtained suggested the presence of functional groups with close acid-base pKa values. Similar results could be observed in all titration curves for each set of bacteria (Figure 1a), showing that although some small variability could be perceived in each set of the same bacterial (50) Fairley, N. CASA-XPS, 2.3.12 ed., Casa Software Ltd., 2006.

Cell Surface and Cell Wall Chemistry of A. commune

strain, essentially reproducible results were obtained (the variation between the titration curves was below 6% of [H+]exchanged between pH 3.5 and 10). Although a small hysteresis could be observed between acid and base titrations at the same ionic strength, results from reverse titrations did not vary strongly and suggested a reversible proton adsorption/desorption reaction. To check for the presence of proteins resulting from the leakage of cellular contents due to lysed cells, a Bradford test49 was performed on the supernatant solutions at the initial and final steps of the titration (pH ) 3.5 and 10). No evidence of proteins was found, suggesting that the titration experiments did not damage the cell wall. Titration experiments were not performed at pH below 3.5 or above 10 due the possibility of cell lysis; however, because of the limited pH coverage of these experiments, it is possible that some additional active functional groups remained untitrated, and, if titrations at lower or higher pH values could be conducted, additional groups might be found and the model of the titration curve could be affected. 3.1.1. Modeling Approaches. To calculate the acidity constants (pKa), data from each titration were fit using PROTOFIT 2.1.51-53 The constant capacitance model (CCM) and nonelectrostatic model (NEM) were used to fit the experimental data. The CCM was used to facilitate comparisons with previous studies29,56,61,67 which have used the same model. The CCM assumes that surface charge is homogeneously distributed over the bacterial surface, giving rise to relatively simple electrostatic field behavior. Although the effects of heterogeneities in the bacterial surface electrostatic field are neglected, a few studies of the distribution of active surface sites on Bacillus subtilis revealed that there is a higher concentration of surface sites at the poles of the bacteria.54,55 Table 1 summarizes the results obtained after PROTOFIT optimizations based on two, three, and four sites. The goodness of the fit is reflected in the weighted sum of squares value SS* (smaller values reflect better fits). PROTOFIT adjusts the values of the model parameters until SS* reaches a minimum.52,53 Although the two-, three-, and four-site models yielded acceptable fits to the experimental data, the two-site model provided the worst fit. The three-site model had a slightly better fit than the four-site model. Additionally, pK2 and pK3 values for the foursite model were not statistically different, indicating that these (51) Turner, B. F.; Fein, J. B. Geochim. Cosmochim. Acta 2005, 69 (suppl. 1), A49. (52) Turner, B. F.; Fein, J. B. Comput. Geosci. 2006, 32, 1344-1356. (53) Turner, F. M. PROTOFIT Version 2.0. A program for determining surface speciation constants from titration data. User’s manual; Department of Civil Engineering and Geological Sciences, University of Notre Dame: Notre Dame, IN, 2005. (54) Sonnenfeld, E. M.; Beveridge, T. J.; Doyle, R. J. Can. J. Microbiol. 1985, 31, 875-877. (55) Sonnenfeld, E. M.; Beveridge, T. J.; Koch, A. L.; Doyle, R. J. J. Bacteriol. 1985, 163, 1167-1171. (56) Fein, J. B.; Daughney, C. J.; Yee, N.; Davis, T. A. Geochim. Cosmochim. Acta 1997, 61, 3319-3328. (57) Yee, N.; Fein, J. Geochim. Cosmochim. Acta 2001, 65, 2037-2042. (58) Martinez, R. E.; Smith, D. S.; Kulczycki, E.; Ferris, F. G. J. Colloid Interface Sci. 2002, 253, 130-139. (59) Daughney, C. J.; Fein, J. B. J. Colloid Interface Sci. 1998, 198, 53-77. (60) Cox, J. S.; Smith, D. S.; Warren, L. A.; Ferris, F. G. EnViron. Sci. Technol. 1999, 33, 4514-4521. (61) Guine´, V.; Spadini, L.; Sarret, G.; Muris, M.; Delolme, C.; Gaudet, J.-P.; Martins, M. F. EnViron. Sci. Technol. 2006, 40, 1806-1813. (62) Dittrich, M.; Sibler, S. J. Colloid Interface Sci. 2005, 286, 487-495. (63) Yee, N.; Benning, L. G.; Phoenix, V. R.; Ferris, F. G. EnViron. Sci. Technol. 2004, 38, 775-782. (64) Borrok, D.; Turner, B. F.; Fein, J. B. Am. J. Sci. 2005, 305, 826-853. (65) Guine´, V.; Martins, J. M. F.; Causse, B.; Durand, A.; Gaudet, J.-P.; Spadini, L. Chem. Geol. 2007, 236, 266-280. (66) Claessens, J.; Van Lith, Y.; Laverman, A. M.; Van Cappellen, P. Geochim. Cosmochim. Acta 2006, 70, 267-276. (67) Haas, J. R.; Dichristina, T. J.; Wade, R. Chem. Geol. 2001, 180, 33-54.

Langmuir, Vol. 24, No. 8, 2008 4035 Table 1. Deprotonation Constants and pHzpc as Calculated by PROTOFIT51-53 Using the Constant Capacitance Model (CCM) no. of sites

pK1

pK2

pK3

2 3 4

3.5 3.6 3.6

6.6 5.6 5.6

8.7 5.6

pK4

pHzpc

SS*a

8.6

3.50 3.71 3.71

3.95 × 10-1 6.31 × 10-2 7.51 × 10-2

a

Weighted sum of squares. The goodness of fit is reflected in the SS* value (see text).

two pKa values may reflect the same functional group. It can be seen from Figure 1b that a three-site CCM and three-site NEM provided a reasonable fit to the raw titration data. Although the NEM describes the data reasonably, it is apparent from Figure 1b that the CCM provided a superior fit to the surface protonation behavior. However, when calculating the pK values and the surface site concentrations, the results obtained from the CCM and NEM were not statistically different, as can be seen in Table 2. For the CCM calculations, the assumed surface layer capacitance value, C, was varied to determine the best fit for each data set. It was found that, regardless of the surface layer capacitance value used, the two-site CCM provided the worst fit to the experimental data when compared to a three- or four-site model. A capacitance value of 1.2 F/m2 provided the best fit for the two-site model, whereas for the four-site model the surface layer capacitance value that best fit the data was 8.0 F/m2. The latter value implies an extremely small electrostatic field effect and thus almost negligible electrostatic interaction contribution to proton binding. Such high capacitance values were also used in other studies.32,56 Titrations with different ionic strength conditions were also tested (see Figure S2 in the Supporting Information). Although a possible ionic strength effect was observed over the range studied (the variation between the titration curves was below 20% of [H+]exchanged between pH 3.5 and 10), these effects were weak compared to the relatively large experimental uncertainties associated with biomass titrations.32,56,57-60 The lack of a significant ionic strength effect was also consistent with that found previously by other authors (see, e.g., refs 32 and 59). The curves at different ionic strength showed an intersection point around pH 3.6-3.7, where there is no effect of the salt concentration. This value could be set as the experimental pH of zero proton charge (pHzpc).31 This value was similar to the pH of zero proton charge (pHzpc) calculated by PROTOFIT using the surface complexation model (shown in Table 1). 3.1.2. Modeling Results. Table 2 summarizes the results obtained from the titrations of A. commune, compared to previously published results for other bacterial strains. Using the CCM, the mean pKa values for the three sites were 3.6 ( 0.5, 5.7 ( 0.3, and 8.6 ( 0.6, with the reported errors being 1 time the standard deviation. Similarly, when using the NEM, the pKa values were 3.5 ( 0.8, 5.7 ( 0.9, and 9.1 ( 1.6. Sites with pK ) 3.5-3.6 could be tentatively assigned to carboxyl groups (pKa of carboxyl groups varies from 2 to 629,56,61-63,67). Sites with pK ) 5.7 could be tentatively assigned to phosphate (phosphomonoester) groups (pKa of phosphate groups varies from 5.6 to 7.229,56,61,62,67). Sites with pK ) 8.6-9.1 could be tentatively assigned to hydroxyl and amine groups; the pKa of phenolic (hydroxyl) groups varies from 8 to 12, and the pKa of amine groups varies from 8.6 to 9.0,29,56,60,62,67 and hence, both amine and hydroxyl groups may be equally attributed to this pK value.29,62 The pK values and site concentrations for A. commune were similar and comparable to data previously reported in the literature

4036 Langmuir, Vol. 24, No. 8, 2008

Ojeda et al.

Table 2. Comparison of Deprotonation Constants and Surface Site Concentrations between Aquabacterium commune and Other Strains from Different Studies species

pK1

pK2

pK3

Aquabacterium 3.6 ( 0.5 5.7 ( 0.3 8.6 ( 0.6 commune, Gram-negative Aquabacterium 3.5 ( 0.8 5.7 ( 0.9 9.1 ( 1.6 commune, Gram-negative Enterobacteriaceae, 4.3 ( 0.2 6.9 ( 0.5 8.9 ( 0.5 Gram-negative Synechococcus Green, 4.85 ( 0.31 6.56 ( 0.2 8.76 ( 0.06 Gram-negative Synechococcus Red, 4.98 ( 0.16 6.69 ( 0.39 8.66 ( 0.21 Gram-negative Calothrix sp., 4.7 ( 0.4 6.6 ( 0.2 9.1 ( 0.3 Gram-negative S. putrefaciens, 5.16 ( 0.04 7.22 ( 0.15 10.04 ( 0.67 Gram-negative B. subtilis, 4.8 ( 0.14 6.9 ( 0.5 9.4 ( 0.6 Gram-positive a

C1 C2 C3 Ctot (×10-4 mol/g) (×10-4 mol/g) (×10-4 mol/g) (×10-4 mol/g) modela

ref

15.4

5.7 ( 0.9

1.7 ( 0.6

8.0 ( 3.5

CCM

this study

14.2

4.6 ( 1.5

1.9 ( 0.6

7.3 ( 3.1

NEM

this study

12.7

5.0 ( 0.7

2.2 ( 0.6

5.5 ( 2.2

CCM

29

7.0

2.6 ( 0.4

1.9 ( 0.5

2.5 ( 0.4

CCM

62

16.6

7.4 ( 1.6

4.4 ( 0.8

4.8 ( 0.8

CCM

62

14.6

3.28 ( 0.27

4.14 ( 0.31

7.16 ( 0.97

NEM

63

0.32 ( 0.02

0.09 ( 0.01

0.38 ( 0.01

CCM

67

12 ( 1

4.4 ( 0.2

6.2 ( 0.2

CCM

56

0.78 22.6

CCM ) constant capacitance model; NEM ) nonelectrostatic model.

for Gram-negative bacteria (Table 2). Moreover, the acid-base behavior of A. commune is consistent with the bacterial “universal” proton binding behavior proposed by Borrok et al.64 and suggests that, despite the model used to determine the pK values, the results are comparable to other bacteria species and the uncertainties associated with factors such as species type, ionic strength, temperature, and growth conditions are relatively small. The surface site concentrations obtained using PROTOFIT (normalized to the dry mass of bacteria) were 5.7 ( 0.9 × 10-4 mol/g for acidic sites (such as carboxyl groups), 1.7 ( 0.6 × 10-4 mol/g for neutral sites (such as phosphomonoester groups), and 8.0 ( 3.5 × 10-4 mol/g for basic sites (such as hydroxyl/ amine groups) when using the CCM and 4.6 ( 1.5 × 10-4 mol/g for acidic sites, 1.9 ( 0.6 × 10-4 mol/g for neutral sites, and 7.3 ( 3.1 × 10-4 mol/g for basic sites when using the NEM. The larger error associated with the basic sites (hydroxyl/amine groups) suggested that other effects, such as the logarithmic evolution of the error function at high pH values, may contribute to generate this uncertainty;65 however, it has also been suggested that it could be an indirect evidence of the presence of more than one functional group.29 The existence of pHzpc (Table 1 and Figure S2 in Supporting Information) indicated that the bacteria developed a positive net charge at low pH values, indicating the presence of at least one positively ionizing, plausibly amino, group. Models which only include negatively ionizing groups such as carboxyl, phosphoryl, and hydroxyl groups could not develop a net positive charge at low pH.66 The pH dependence of the cell wall charge for A. commune was similar to the ones reported for Shewanella putrefaciens (a Gram-negative bacteria) or Bacillus subtilis (a Gram-positive bacteria),66 despite the large differences in cell wall structures. A. commune and other Gram-negative bacteria possess an outermost layer composed principally of lipopolysaccarides (LPSs). This LPS layer is thought to play an important role in metal binding and cell adhesion.62,67,68 However, Gram-positive bacteria do not have LPSs but develop a surface charge similar to that of Gram-negative bacteria, indicating, as suggested by Claessens,66 that the presence of the LPS layer does not contribute significantly to the cell wall charge. A recent study by de Kerchove and Elimelech69 found that the presence of flagella is likely to (68) Langley, S.; Beveridge, T. J. Appl. EnViron. Microbiol. 1999, 65, 489498. (69) de Kerchove, A. J.; Elimelech, M. Appl. EnViron. Microbiol. 2007, 73, 5227-5234.

change the surface properties of bacteria, with the motile strains being more hydrophilic and having a higher surface acidity than nonmotile strains. These findings suggest that the single polar flagellum in A. commune strains may contribute to the titrated surface charge and affect the deposition of these bacteria onto surfaces in aquatic systems. The pHzpc ) 3.7 also indicated that A. commune was negatively charged at pH ) 7 (which is a typical pH in water distribution pipes) and electrostatic attraction with positive-charged surfaces, such as iron oxide corrosion products or calcium carbonate precipitates, may be favorable. 3.2. Infrared Spectroscopy. The infrared spectra of A. commune in 0.1 M NaClO4 electrolyte solution are shown in Figure 2. These spectra contain information about the functional groups present at different pH values during titrations. It has to be taken into account that the infrared spectra of the bacterial cells obtained using a liquid flow cell in the transmission mode could not distinguish between the cell wall and the inner cell. However, a measurement at the end of the flow cell experiment, using an ATR-FTIR system on a germanium crystal at a fixed angle of 45°, gave essentially identical spectra. The similarities between the ATR spectrum and the flow cell spectrum could be explained because the cell wall constitutes 60-70% of the total bacterial weight in a hydrated state.37 Additionally, a study conducted by Jiang et al.,70 comparing the spectra of isolated cell wall fragments of both Gram-positive and Gram-negative bacteria with the intact cells, showed that the ATR-FTIR spectra of the intact bacterial cells mostly reflected the properties of the isolated cell walls. 3.2.1. Absorption Bands of Bacterial Functional Groups. The observed infrared bands for A. commune corresponded to the presence of functional groups from macromolecules such as proteins, carbohydrates, lipids, polyphosphate groups, and other polysaccharides. The functional groups assigned for the infrared bands and the corresponding frequencies for the A. commune spectra are based on the vibration patterns previously reported for bacteria35,36,62,63,70,71 and are summarized in Table 3. At 1739 cm-1, a signal corresponding to the vibrational CdO stretching (νCdO) of carboxylic groups from membrane lipids and fatty acids was observed. This band at 1739 cm-1 showed (70) Jiang, W.; Saxena, A.; Song, B.; Ward, B. B.; Beveridge, T. J.; Myneni, S. C. B. Langmuir 2004, 20, 11433-11442. (71) Schmitt, J.; Flemming, H.-C. Int. Biodeterior. Biodegrad. 1998, 44, 1-11.

Cell Surface and Cell Wall Chemistry of A. commune

Figure 2. Infrared spectra of Aquabacterium commune in 0.1 M NaClO4 solution during titration at different pH values. Approximate band positions are indicated in the figure.

a small decrease in the intensity when decreasing the acidity, which is more apparent if the height of the peak at 1739 cm-1 is compared with the peak at 1548 cm-1. Although the band corresponding to this signal has been commonly attributed to the stretching CdO from esters and carboxylic acids,35,61,70 the deprotonation of other groups different from carboxylic acids could be causing a decrease in the intensity at higher pH values. In fact, the insert in Figure 2 shows that the band at 1739 cm-1 was actually a combination of two peaks: one at 1739 cm-1 and another peak at 1725 cm-1. The presence of a band at 1402 cm-1 has been attributed to the stretching CsO of deprotonated carboxylic groups, indicating the formation of a carboxylate anion.35,70 The amide I and II bands appeared at 1647 and 1548 cm-1, respectively, with the first one due to the stretching CdO (νCdO) of amides associated with proteins and the latter due to a combination of the bending NsH (δNsH) of amides and also contributions from the stretching CdN (νCdN) groups. These two peaks can also overlap contributions from the bending NsH and NsH2 due to amines.35

Langmuir, Vol. 24, No. 8, 2008 4037

The absorption at 1262 cm-1 could be attributed to the double bond stretching of PdO of general phosphoryl groups and phosphodiester of nucleic acids.70 Another peak corresponding to the stretching of PdO groups of phosphorylated proteins, polyphosphate products, and nucleic acid phosphodiester was found around 1085 cm-1. However, the region between 1200 and 950 cm-1 was dominated by the complex superposition of vibrations corresponding to the C-O-C and C-O-P stretching of diverse polysaccharides groups, where specific band assignments are very difficult. Additionally, there was an increase in the intensity of the band appearing at 976 cm-1, attributed to the symmetric stretching vibration of phosphoryl groups. The increase in the band was accompanied by the appearance of a band at 1225 cm-1 corresponding to the stretching of PdO in phosphates. The presence of two bands may be attributed to deprotonated phosphoryl groups PO3 and PO2 associated with the phospholipids or phosphorylated proteins. It is well-known that the phosphoryl groups can form resonance hybrids and stabilize excess charge around the double-bonded oxygen atoms and the phosphorus atom.36,72 The observed FTIR signals indicate the presence of carbon, phosphor, and nitrogen atoms in the bacterial samples. Specific FTIR signals were found to vary with pH. Changes in these signals involve contributions from acid-base reactive carboxyl, phosphoryl, and amine functional groups. This indicates that such functional groups exist on bacterial surfaces and suggests that these groups contribute to the acid-base exchange reactivity observed in the titration experiment. 3.3. X-ray Photoelectron Spectroscopy. The XPS wide scan collected for A. commune can be seen in Figure 3a. It can be observed from the spectra that the outermost cell surface layer (2-5 nm penetration depth) was mainly constituted of C, O, N, and P. The Si peak is most likely derived from the glassware used in the cultivation and extraction protocol. 3.3.1. Elements and Functional Groups. The elemental composition of the bacterial surface, resulting from integrating the C1s, O1s, N1s, and P1s peaks from the wide scan spectrum can be seen in Table 4. Nitrogen appeared at a binding energy of 398.59 eV, attributable to amine or amide groups of proteins.37,39,40,44,45 Phosphorus was found at a binding energy of 132.15 eV and can be attributed to phosphate groups.37,39,40 The presence of amine groups from proteins and phosphate groups based on the binding energies of N1s and P1s also attests to the results from potentiometric titrations (pKa ) 5.7 and 8.6) and the FTIR spectra (adsorption bands at 1647, 1548, and 976 cm-1). XPS peaks corresponding to C and O were analyzed at high resolution and deconvoluted to assess the contributions from each component (see Figure 3b and c). The carbon peak (C1s) was fit into four components: carbon bound only to carbon and hydrogen, Cs(C,H), at 282.7 eV; carbon singly bound to oxygen or nitrogen from ethers, alcohols, amines, and/or amides, Cs (O,N), at 284.1 eV; carbon doubly bonded to oxygen or singly bonded to two oxygen atoms from amides, carbonyls, carboxylates, esters, acetals, and/or hemiacetals, CdO and OsCsO, at 285.4 eV; and carbon attributable to carboxylic functions, COOR, at 286.7 eV. The oxygen peak (O1s) was best fit with two contributions: oxygen double bonded with carbon or phosphorus from carboxylic acids, carboxylates, esters, carbonyls, amides, or phosphoryl groups, CdO and PdO, at 529.7 eV; and oxygen attributable to hydroxide, phosphate, acetal, or hemiacetal, Cs OH, CsOsC, and PsOH, at 530.5 eV. (72) Walker, B. J. Organophosphorus chemistry; Penguin: Harmondsworth, U.K., 1972.

4038 Langmuir, Vol. 24, No. 8, 2008

Ojeda et al.

Table 3. Absorption Bands of the Aquabacterium commune Functional Groups wavenumber (cm-1) ≈1739-1725 ≈1647 ≈1548 ≈1402 ≈1453 ≈1384 ≈1305 ≈1300-1250 ≈1262 ≈1225 ≈1200-950 ≈1085 ≈976

functional group assignment

ref

stretching CdO of ester functional groups from membrane lipids and fatty acids; stretching CdO of carboxylic acids stretching CdO in amides (amide I band); bending -NH and -NH2 of amines N-H bending and C-N stretching in amides (amide II band); bending -NH and -NH2 of amines symmetric stretching for deprotonated COO- group bending CH2/CH3 (scissoring) symmetric stretching of COO-; bending CH2/CH3 vibration C-N from amides vibrations of C-O from esters or carboxylic acids vibrations of -COOH and C-O-H; double bond stretching of >PdO of general phosphoryl groups and phosphodiester of nucleic acids stretching of PdO in phosphates asymmetric and symmetric stretching of PO2- and P(OH)2 in phosphates; vibrations of C-OH, C-O-C, and C-C of polysaccharides. stretching PdO of phosphodiester, phosphorylated proteins, or polyphosphate products symmetric stretching vibration of phosphoryl groups

62, 70, 35

The fraction of carbon bound to oxygen or nitrogen obtained from the deconvolution of the carbon peak can be related to the sum of the atomic concentration ratios of oxygen and nitrogen with respect to carbon (O/C + N/C). A good correlation between these two independently obtained values indicates that nitrogen or oxygen is mainly bound to carbon in a 1:1 ratio.37,40 The agreement obtained from the sum O/C + N/C compared to (C(O,N))/C found in A. commune suggested that the functional groups on the cell surface correspond mainly to amides, alcohols, amines, esters, or groups in which oxygen or nitrogen is bound to carbon in a 1:1 ratio. For amide functions, when comparing the atomic concentration ratio of nitrogen with respect to carbon (N/C) and the oxygen doubly bonded to carbon (OdC)/C and carbon doubly bonded with oxygen (CdO)/C, there was a large excess of CdO with respect to N/C, indicating the presence of other CdO groups different from amides. Similarly, the oxygen double bonded with carbon (OdC)/C was in excess with respect to what could be attributable to amide, indicating the presence of other functional groups. This result, combined with the FTIR spectra (bands at 1739, 1725, and 1402 cm-1 for carboxylic acids, esters, and carboxylate anions) and the acidic groups, such as carboxylic acids, observed from the potentiometric titrations (pKa ) 3.6; 5.7 ( 0.9 × 10-4 mol/g) suggests that A. commune cell walls possess carbonyl groups not only from amides, but also from other carbonyl moieties (such as carboxylic acids and esters). 3.3.2. Molecular Composition. XPS analysis has also been used to estimate the concentration of the basic main constituents of the cell walls by assuming three classes of constituents for the bacterial surface: polysaccharides, peptides, and main features of lipidic compounds (further referred to as hydrocarbon-like products). A simple approach to compute the molecular composition of the bacterial surface was made by Rouxhet et al.38 and also applied by Dufreˆne et al.,40 consisting of comparing the measured concentration ratios N/C, O/C, and P/C with the carbon concentration and the atomic concentration ratio of nitrogen and oxygen with respect to carbon of model compounds such as Glucan (C6H10O5)n for polysaccharides (giving O/C ) 0.833, N/C ) 0.000, and [C] ) 37.0 mmol/g) and (CH2)n for hydrocarbon-like products (giving C-(C,H)/C ) 1.000 and [C] ) 71.4 mmol/g) and computing the data from the amino acid composition of the major outer membrane protein of Gram-

70, 35 70, 35 63, 70 62, 70 62, 70 71 35, 36 62, 70 70 62, 70 62 70

negative Pseudomonas fluorescens OE 28.373 for peptides (giving O/C ) 0.325, N/C ) 0.279, and [C] ) 43.5 mmol/g). For a more extensive discussion about this methodology, the reader can refer to van der Mei et al.,37 Rouxhet et al.,38 and Dufrene et al.40 By considering the above compositions, a set of three equations can be provided:

O/C ) 0.325 (CPEP/C) + 0.833 (CPS/C)

(2)

N/C ) 0.279 (CPEP/C)

(3)

1 ) (CPEP/C) + (CPS/C) + (CHC/C)

(4)

where O/C and N/C are the observed atomic concentration ratios of oxygen and nitrogen, respectively, with respect to carbon in the analyzed sample and CPEP, CPS, and CHC are the atomic concentrations of carbon present in peptides, polysaccharides, and hydrocarbon-like products, respectively. By solving the system of equations, the estimated proportions of carbon associated with each molecular constituent for A. commune are CPEP/C ) 0.175, CPS/C ) 0.249, and CHC/C ) 0.576. By using the carbon concentration of each constituent, the latter proportions can be converted into weight fractions, showing that the predominant constituents of the cell surface are hydrocarbon-like compounds (42.9 ( 2.5%), followed by polysaccharides (35.7 ( 1.8%) and peptides (21.4 ( 3.7%). Although the major predicted constituents of the cell wall are hydrocarbn-like compounds, a distinction between phospholipids and other features of lipidic compounds was not possible with this approach due to the simplifications involved in the calculations. It should also be noted that these estimated values might differ from the real composition because of the simplifications inherent in the approach used and that the XPS data were obtained using freeze-dried cells. However, cell disruption and migration of intracellular components to the surface are prevented by using well-established techniques for freeze-drying that avoid these problems.37-40 A comparison of XPS analysis and biochemical analysis of five bacterial strains published by Dufreˆne et al.40 supports the validity of using XPS to estimate the overall composition of the cell surface. (73) De Mot, R.; Proost, P.; Van Damme, J.; Vanderleyden, J. Mol. Gen. Genet. 1992, 231, 489-493.

Cell Surface and Cell Wall Chemistry of A. commune

Langmuir, Vol. 24, No. 8, 2008 4039 Table 4. Binding Energies (eV), Assignments, and Quantitation of XPS Spectral Bands of Freeze-Dried Aquabacterium communea element

peak (eV)

concn (% ( sd)b

total C total O total N total P

283.81 ( 0.14 531.03 ( 0.09 398.59 ( 0.22 132.15 ( 0.23

75.79 ( 0.90 20.00 ( 0.47 3.70 ( 0.65 0.51 ( 0.04

C1s C1s C1s C1s O1s O1s

282.72 ( 0.10 284.17 ( 0.11 285.48 ( 0.12 286.73 ( 0.11 529.50 ( 0.21 530.56 ( 0.12

61.80 ( 0.66 23.93 ( 0.69 8.07 ( 0.42 6.19 ( 0.45 22.91 ( 0.99 77.09 ( 0.99

O/Cc N/Cc P/Cc

assignment

Cs(C,H) Cs(O,N) CdO, OsCsO COOR CdO, PdO CsOH, CsOsC, PsOH

26.39 ( 0.87 4.89 ( 0.90 0.68 ( 0.06

a Atomic fractions for elemental composition were obtained from the low-resolution survey scan. Quantification of the functional group composition was obtained from the high-resolution spectra. b Mean values ( standard deviation for triplicate samples are shown. c Atomic concentration ratios with respect to total carbon, multiplied by 100.

Figure 3. (a) XPS survey spectrum collected from Aquabacterium commune, measured after freeze-drying of the cells. The peaks pertaining to C1s (b) and O1s (c) were scanned at high resolution (0.1 eV step-size) and deconvoluted to assess the local chemical environment of these elements.

The amount of polysaccharides and peptides when combined is relatively higher when compared to nonpolar hydrocarbonlike compounds, suggesting that the bacterial surface of A. commune should be hydrophilic. The atomic concentration ratio of nitrogen with respect to carbon (N/C) can also help to infer the cell surface hydrophilicity. It has been found that the intrinsic cell surface hydrophobicity increases when increasing the N/C elemental surface composition.37 van der Mei and Busscher74 studied a collection of differently cultured Streptococcus mitis and compared the results from XPS analysis and water contact angle measurements, showing that low N/C ratios (around 0.05) exhibited a water contact angle of ∼20° (hydrophilic) whereas high N/C values (around 0.15) showed contact angles of ∼100° (hydrophobic). A. commune exhibited a N/C ratio of 0.05, (74) van der Mei, H. C.; Busscher, H. J. Eur. J. Oral Sci. 1996, 104, 48-55.

suggesting a hydrophilic cell surface. Although Marshall et al.75 have warned of the danger of relying on relationships between XPS data and cell surface properties, especially for Gram-negative bacteria, the above results are also in agreement with the potentiometric data and FTIR spectroscopy data. Moreover, the relatively high amount of polysaccharides exposed on the outer membrane of A. commune is a fundamental factor controlling bacterial adhesion and transport according to Walker et al.24 and Korenevsky et al.76 Although the FTIR spectra of A. commune displayed bands composed of a combination of peaks from proteins, lipids, and carbohydrates, this information corresponded only to the identity and average protonation state of the functional groups present in the cells on a molecular level. Therefore, the elemental surface composition, the relationship between functional groups in the macromolecules, and the estimation of the weight fraction of each basic cell component by XPS was also necessary to obtain a more complete characterization of the cell wall. On the other hand, the XPS data only constrained the chemistry of the outermost 2-5 nm of the cell surface, and the chemistry of the bulk cell wall could be different than what is suggested by the XPS data alone. FTIR spectroscopy and XPS helped describe the macromolecular structure and composition of the bacterial cell wall. The study of the thermodynamic properties and densities of proton binding sites on each of those surface molecules at low detection limits relied on the potentiometric measurements, which have much higher precision for proton and hydroxide concentrations to low detection limits, compared to the spectroscopic techniques.

4. Concluding Remarks The cell surface properties of Aquabacterium commune were examined by potentiometric titrations, modeling, and FTIR spectroscopy. This bacterium was chosen as a model microorganism to show how the combination of these methods allows a more comprehensive characterization of bacterial cell surfaces and cell wall reactivity. Parameters such as surface charge, acidbase behavior, chemical composition, and changes of the organic functional groups of the cell wall at different pH values could be characterized in this bacteria for the first time. (75) Marshall, K. C.; Pembrey, R.; Schneider, R. P. Colloids Surf., B 1994, 2, 371-376. (76) Korenevsky, A.; Beveridge, T. J. Microbiol. 2007, 153, 1872-1883.

4040 Langmuir, Vol. 24, No. 8, 2008

The potentiometric titrations showed that the protonationdeprotonation process between pH ) 3.5 and 10 was reversible for this specific micro-organism. Potentiometric titrations, modeling, and FTIR spectroscopy provided evidence of the presence of three active sites in the bacterial surface: carboxylic groups, with a concentration of 5.7 ( 0.9 × 10-4 mol/g for the carbonyl groups, 1.7 ( 0.6 × 10-4 mol/g for the phosphoryl groups, and 8.0 ( 3.5 × 10-4 mol/g for the hydroxyl/amine groups. The total concentration of active sites was 15.4 × 10-4 mol/g, which is a comparable value similar to other Gram-negative bacteria. The pHzpc obtained when titrating the same mass of bacteria at three different electrolyte concentrations was around 3.7, which was also predicted when modeling the potentiometric titrations using PROTOFIT. This value indicates that A. commune is negatively charged in its typical neutral pH environment. This is an important result due to the role of electrostatic interactions in initial cell attachment and adhesion to solid surfaces such as pipe walls and metal corrosion films. FTIR spectroscopy showed that carboxylic and phosphoryl groups are sensitive to changes in solution pH and the deprotonation of these two groups could be the cause of the negative charge developed at pH 7. The calculated molecular composition of the surface of A. commune as analyzed by XPS was as follows: hydrocarbon-like compounds (43%), polysaccharides (36%), and peptides (21%). The results from XPS, potentiometric titrations, modeling, and FTIR spectroscopy presented here showed that the combination of these techniques effectively helps to characterize the bacterial surface and to describe the changes of functional groups

Ojeda et al.

in the cell wall as a necessary step to understand the interactions taking place between bacteria and their environment. These findings are crucial to allow further investigation of the molecular interaction taking place between cell walls and solid surfaces, and to describe the physical-chemical parameters influencing cell adhesion. The improved description of the cell wall chemistry and reactivity, specifically the understanding of contributions to adhesion forces and interaction energies in terms of key groups of macromolecules, that may influence hydrophobicity of the cell wall, can support theoretical adhesion models by improving their conceptualization and parametrization. Acknowledgment. This work has been funded by the Engineering and Physical Sciences Research Council (EPSRC) and the Biotechnology and Biological Sciences Research Council (BBSRC) as part of the Cell-Mineral Interface Research Program (GR/S72467/01(P)). J.J.O. acknowledges funding from the EPSRC Platform Grant strategic funding (GR/S87416/01). The authors also thank Zhenyu Zhang, Department of Physics and Astronomy, University of Sheffield, for providing AFM images of A. commune, Gautam Mishra and Dr. Sally McArthur, Department of Engineering Materials, for support given with the XPS analysis, and Jonathan Bridge, Department of Civil and Structural Engineering, for useful discussions about particle transport, adhesion, and DLVO theory. Supporting Information Available: Atomic force microscopy images to calculate bacterial dimensions (Figure S1) and titrations at different ionic strength (Figure S2). This material is available free of charge via the Internet at http://pubs.acs.org. LA702284B