11368
Langmuir 2005, 21, 11368-11372
Nonionic Surfactant Sorption onto the Bacterial Cell Surface: A Multi-Interaction Isotherm Derick G. Brown* and Khaula S. Al Nuaimi Department of Civil & Environmental Engineering, Lehigh University, Bethlehem, Pennsylvania 18015 Received May 25, 2005. In Final Form: August 21, 2005 The adsorption of linear polyoxyethylene (POE) alcohol surfactants of the form CxEy onto the surface of a Sphingomonas sp. has been examined. For this study, the alkyl chain length (x) was fixed at 12 and the POE chain length (y) was varied, with y ) 4, 7, 9, 10, and 23 ethylene oxide units. Langmuirian isotherms were observed for C12E4 and C12E23, and more complex isotherms were observed for the three intermediate POE chain length surfactants, with C12E7 and C12E9 exhibiting strong S-shaped isotherms. All isotherms showed plateaus near the critical micelle concentration (CMC) with the plateau decreasing with increasing POE chain length. A simple multi-interaction isotherm is proposed that models the sorption isotherm as the sum of two interactions. The first interaction describes monolayer adsorption, whereas the second interaction describes lateral interactions between sorbed surfactant molecules and the formation of surface aggregates. Varying ratios of these two interactions as a function of POE chain length gives rise to the variety of observed isotherm shapes. Results of the isotherm analysis suggest that lateral interactions dominate for surfactants with low POE chain lengths, and the lateral interactions decrease as the POE chain length is increased.
Introduction and Background Knowledge of surfactant adsorption to the bacterial cell surface is necessary for a number of applications. These include surfactant-enhanced biodegradation of hydrophobic organic compounds, such as polycyclic aromatic hydrocarbons (PAHs); alteration of bacterial adhesion to solid surfaces, including food preparation surfaces and medical devices; effects of surfactants in domestic sewage on bacterial transport through soils in septic tank leach fields; and surfactant solubilization of bacterial membranes and membrane-bound proteins. Although an extensive body of literature exists for surfactant sorption onto solid surfaces, there is a large gap in our knowledge regarding surfactant sorption onto bacteria, whose surfaces are more gellike in nature and consist of various lipids, proteins, and surface polymers. Despite the large number of applications involving bacteria and surfactants, there have only been a few published studies providing surfactant sorption isotherms onto bacteria. In one study, it was observed that sorption isotherms of two linear polyoxyethylene alcohol surfactants onto a Sphingomonas sp. appeared to be Langmuirian.1 The isotherms in this study were performed below the surfactants’ critical micelle concentration (CMC) so the full isotherm parameters could not be determined. Another study examined Triton X-100 sorption onto a Chlorella sp., but unfortunately all data points except for one were present above the CMC, and as such the data is only useful for identifying the sorption plateau.2 Finally, S-shaped isotherms were observed in a study of Triton X-45 and Triton X-100 sorption onto Staphylocuccus aureus.3 However, there were only five and six data points, * Corresponding Author. Address: Department of Civil & Environmental Engineering, Lehigh University, 13 East Packer Avenue, Bethlehem, PA 18015. Phone: 610-758-3543. Fax: 610758-6405. E-mail:
[email protected]. (1) Brown, D. G.; Jaffe´, P. R. Environ. Sci. Technol. 2001, 35, 20222025. (2) Komor, E.; Weber, H.; Tanner, W. Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 1814-1818.
respectively, for each of the isotherms, and no attempt was made to model the experimental data. The goal of this study was to develop detailed isotherms for nonionic surfactant sorption onto the bacterial cell surface and to determine how the surfactant structure affects the isotherm. Linear polyoxyethylene alcohol surfactants (Figure 1) were chosen for this work. These surfactants have been used in many studies with bacteria, including surfactant-enhanced biodegradation,4-7 bacterial adhesion,8,9 bacterial transport through porous media,10,11 and interactions with bacterial cell walls.12-15 For this study, the alkyl chain length (x) was fixed at 12 carbons and the POE chain length (y) was varied from 4 to 23 (Table 1), allowing determination of the effects of POE chain length on the sorption isotherm. Finally, an additional goal of this study was to use the experimental data to develop a working isotherm formulation that could readily be used to model nonionic surfactant sorption onto the bacterial cell surface. (3) Lamikanra, A.; Allwood, M. C. Microbios Lett. 1976, 1, 97-101. (4) Brown, D. G.; Guha, S.; Jaffe´, P. R. Bioremediation J. 1999, 3, 269-283. (5) Guha, S.; Jaffe´, P. R. Environ. Sci. Technol. 1996, 30, 13821391. (6) Guha, S.; Jaffe´, P. R. Environ. Sci. Technol. 1996, 30, 605-611. (7) Guha, S.; Jaffe´, P. R.; Peters, C. A. Environ. Sci. Technol. 1998, 32, 2317-2324. (8) Humphries, M.; Jaworzyn, J. F.; Cantwell, J. B. FEMS Microbiol. Ecol. 1986, 38, 299-380. (9) Humphries, M.; Jaworzyn, J. F.; Cantwell, J. B. FEMS Microbiol. Lett. 1987, 42, 91-101. (10) Brown, D. G.; Jaffe´, P. R. Environ. Sci. Technol. 2001, 35, 38773883. (11) Li, Q.; Logan, B. E. Water Res. 1999, 33, 1090-1100. (12) Brown, D. G.; Jaffe´, P. R. Biotechnol. Bioeng. 2001, 74, 476482. (13) Florence, A. T.; Tucker, I. G.; Walters, K. A., Interactions of nonionic polyoxyethylene alkyl and aryl ethers with membranes and other biological systems. In Structure/performance relationships in surfactants; Rosen, M. J., Ed.; American Chemical Society: Washington, DC, 1984; Vol. 253. (14) Tanford, C.; Reynolds, J. A. Biochim. Biophys. Acta 1976, 457, 133-170. (15) Umbreit, J. N.; Strominger, J. L. Proc. Natl. Acad. Sci. 1973, 70, 2997-3001.
10.1021/la051388k CCC: $30.25 © 2005 American Chemical Society Published on Web 10/14/2005
Bacterial Cell Surface Nonionic Surfactant Sorption
Figure 1. Nonionic surfactants used in this study, expressed as CxEy, where x is the number of carbons in the alkyl chain and y is the number of ethylene oxide units in the polyoxyethylene chain. The alkyl chain was fixed at x ) 12 and the POE chain was varied, with y ) 4, 7, 9, 10, and 23. Table 1. Properties of Surfactants Used in This Study surfactant
MW (g/moles)
CMC (mg/l)
CMC (M)
C12E4 C12E7 C12E9 C12E10 C12E23
362 494 581 626 1198
14.3 32.4 39.8 25.2 46.7
3.95 × 10-5 6.56 × 10-5 6.85 × 10-5 4.03 × 10-5 3.90 × 10-5
Materials and Methods Bacteria. A Sphingomonas sp., which was isolated from PAHcontaminated soils, was the focus of this study. This bacterium, designated culture DGB01o, has been previously used in pure and mixed-culture surfactant-enhanced biodegradation studies,4,6,7 bacterial transport studies with and without surfactants,10,16 and a surfactant-cell membrane interaction study.12 Prior studies have indicated that this bacterium is not able to biodegrade the class of surfactants used in this study.17 This bacterium is a Gram-negative rod, with approximate dimensions of 2 µm by 0.5 µm. Stock bacterial cultures were maintained on slants of R2A agar (Diffco) at 4 °C. For each experiment, the bacteria were plated on Petri dishes of R2A agar and allowed to grow for 5 days at 20 °C. They were then harvested and suspended in a baseline solution of 10-3 M CaCl2. The bacterial suspension was placed on a magnetic stirrer and allowed to mix at room temperature for a minimum of 1 h. The suspension was then centrifuged for 30 min at 1500xg, decanted, resuspended in identical baseline solution, and placed on a magnetic stirrer at 20 °C to mix overnight. The next morning the bacteria were centrifuged and resuspended as before. Bacterial concentrations were determined by measuring the absorbance of the bacterial suspension at 220 nm and using predetermined calibration curves from plate counts using the pour-plate method.18 The unit of measure from this technique is the colony forming unit (CFU), which is considered equivalent to a single bacterium. Final bacterial concentrations for the isotherm experiments were approximately 2.5 × 108 CFU/ mL. Surfactants. Five surfactantssC12E4 (Brij 30), C12E7, C12E9, C12E10, and C12E23 (Brij 35)swere obtained from Sigma and used without any further processing. The critical micelle concentrations (CMC) of the surfactants were determined from equilibrium surface tension measurements using a DuNouy surface tensiometer (Fisher Scientific). The surfactant properties are provided in Table 1. The aqueous surfactant concentrations were determined using an iodine-iodide (I-I) assay developed by Baleux19 and modified by Brown and Jaffe´.1 With this assay, a colored complex of iodineiodide and nonionic surfactant is quantified spectrophotometrically. The I-I reagent was prepared by mixing 1 g of iodine with 2 g of potassium iodide in 100 mL of deionized water. For the I-I assay, 0.25 mL of I-I reagent is added to 10 mL of the sample surfactant solution and allowed to equilibrate for 30 min. The difference in adsorption at a wavelength of 500 nm between the baseline solution with I-I reagent and the surfactant solution (16) Brown, D. G.; Stencel, J. R.; Jaffe´, P. R. Water Res. 2002, 36, 105-114. (17) Guha, S. Surfactant enhanced bioremediation: Bioavailability of hydrophobic organic compounds partitioned into the micellar phase of nonionic surfactants. Ph. D., Princeton University, Princeton, NJ, 1996. (18) American Public Health Association; American Water Works Association; Water Environment Federation, Standard Methods for the Examination of Water and Wastewater, 20th ed.; American Public Health Association: Washington, DC, 1998. (19) Baleux, B. C. R. Acad. Sci., Ser. C 1972, 274, 1617-1620.
Langmuir, Vol. 21, No. 24, 2005 11369 with I-I reagent is used to determine the surfactant concentration from predetermined calibration curves. It should be noted that, although Sigma does not provide purity or polydispersivity data for the surfactants used in this study, a highly linear response with very low accompanying error was observed with the I-I assay as a function of the POE chain length1 (data not shown), suggesting that polydispersivity in the surfactants as provided by Sigma is relatively small. Sorption Isotherm Experiment. Sorption isotherm experiments were performed by placing 25 mL of bacterial suspension into 50 mL glass centrifuge tubes. Surfactant was added to the tubes at specified concentrations. The samples were tumbled end over end on a rotary shaker for 1 h, which preliminary experiments indicated was sufficient to achieve equilibrium, followed by centrifugation at 1500xg for 1 h. The aqueous surfactant concentration was then determined with the I-I assay using a UV/visible spectrophotometer (Helios-β, Thermo Scientific). Quality control checks were performed on the I-I assay during each experiment to ensure full recovery of surfactant and that no surfactant contamination was present in the samples. The sorbed surfactant concentration was determined as the difference between the total surfactant added to each tube and the aqueous concentration determined from the I-I assay. Multi-Interaction Isotherm Model. A simple multiinteraction isotherm is proposed that accounts for monolayer adsorption via a Langmuir isotherm followed by lateral interactions and formation of surface aggregates via a modified Langmuir isotherm. First, the CMC multiple, x, is defined as the ratio between the aqueous surfactant concentration, C (mole/L) and the surfactant’s critical micelle concentration, CCMC (mole/ L)
x)
C CCMC
(1)
Then, the multi-interaction isotherm is written as the sum of the two adsorption processes
x xp S ) Γmax,1 + Γmax,2 K1 + x K2 + xp
(2)
where Γmax,1 and Γmax,2 are the maximum sorption concentrations for the two interactions (107 molecules/CFU); K1 and K2 are halfsaturation constants for each interaction (unitless); and p is an exponent (unitless). The first term in eq 2 is a Langmuir isotherm describing monolayer adsorption on the bacterial cell surface, and the second term is a modified Langmuir isotherm accounting for lateral interactions between the adsorbed surfactants and formation of surface aggregates. The form of the modified Langmuir interaction is empirical in nature, where the exponent serves to make the rise of the Langmuir isotherm steeper, inline with experimental observations. The plateau of the total isotherm, Γmax, is then given as
Γmax ) Γmax,1 + Γmax,2
(3)
Results and Discussion The sorption isotherms are shown in Figure 2 with the aqueous surfactant concentration presented as the CMC multiple and the sorbed surfactant concentration presented as 107 molecules/CFU. Three main observations can be made from these isotherms: (1) there is a maximum sorption plateau in all isotherms that occurs at or just beyond the CMC; (2) this maximum plateau increases with decreasing POE chain length; and (3) there are multiple interactions occurring with the isotherms, which are especially evident for the intermediate POE chain lengths of 7 and 9. As discussed above, there is very limited literature on nonionic surfactant sorption onto bacteria to which the current results can be related. However, comparisons can be made to studies with solid surfaces, where all three observations have been reported in the literature.
11370
Langmuir, Vol. 21, No. 24, 2005
Brown and Al Nuaimi
Figure 2. Adsorption isotherms for C12E4, C12E7, C12E9, C12E10, and C12E23 onto the Sphingomonas sp. Lines are best-fit isotherm parameters and the sums of the squared weighted residuals are presented for the two isotherms. Table 2. Best-Fit Isotherm Parameters for the Multi-Interaction Isotherm (eq 2) surfactant C12E4 C12E7 C12E9 C12E10 C12E23
(107 molecules/CFU) Γmax,1 Γmax,2 Γmax 5.52 3.72 1.56 2.19 4.44
31.5 18.4 18.2 7.93 0.543
37.0 22.1 19.8 10.1 4.98
K1 (xCMC)
K2 (xCMC)P
p
0.116 0.0134 0.0516 0.125 1.55
2.15 0.250 2.35 1.33 1.17
2.38 5.35 8.11 3.04 2.41
With regards to the first two observations, during a nonionic surfactant sorption study with silica as the solid surface, it was shown that the adsorption isotherms reached a plateau around CMC and that the plateau decreased with increasing POE chain length.20-22 This has also been observed in a previous study of nonionic surfactant sorption onto powdered activated carbon, where the plateau was shown to decrease as a power function of the POE chain length.23 For the current bacterial study, the adsorption plateaus (Table 2) are plotted as a function of the POE chain length in Figure 3. In agreement with these prior studies, the plateau for nonionic surfactant sorption onto the Sphingomonas sp. can also be represented by a power function of the POE chain length (determination of the adsorption plateau values is dis(20) Levitz, P.; Van Damme, H. J. Phys. Chem. 1986, 90, 13021310. (21) Levitz, P.; Van Damme, H.; Keravis, D. J. Phys. Chem. 1984, 88, 2228-2235. (22) Levitz, P. E. C. R. Geosci. 2002, 334, 665-673. (23) Narkis, N.; Ben-David, B. Water Res. 1985, 19, 815-824.
Figure 3. Adsorption plateau (Γmax) as a function of POE chain length.
cussed below), indicating that the adsorption plateau as a function of POE chain length is not unique to solid surfaces but also occurs with the heterogeneous cell surface. The third observation has also been reported in prior studies with nonionic surfactants. S-shaped isotherms have been observed for adsorption of Triton X-100 and Triton X-45 onto Staphylocuccus aureus.3 S-shaped isotherms have also been reported for Triton X-100 sorption
Bacterial Cell Surface Nonionic Surfactant Sorption
onto solid surfaces, including quartz, glass and silica,21,24 sand,25 soils,26,27 clays,27,28 polyethylene,29 and polystyrene.30 Sorption of CxEy surfactants exhibited S-shaped isotherms on hydrophobic polystyrene30,31 and sorption of a number of different nonionic surfactants onto polar surfaces also showed similar S-type isotherms.20,22,28,32-35 Additionally, a rhamnolipid biosurfactant from Pseudomonas aeruginosa showed similar isotherms for sorption onto two different sandy soils.36 For ionic surfactants, where S-shaped adsorption isotherms are often observed, multiple sorption regions have been proposed in the literature that relate components of the sorption isotherm to electrostatic interactions between the surfactants and solid surface and lateral interactions between the sorbed surfactant monomers.37-39 For nonionic surfactants, there are no electrostatic interactions, and it is believed that the S-curve results from strong lateral interactions between sorbed surfactant molecules.39 Specifically, it has been suggested that the sharp rise in the isotherm near the CMC is due to lateral interactions of sorbed surfactant molecules, resulting in the formation of surface aggregates such as hemimicelles, admicelles or bilayers.22,23,25,29,35,36,40 Given the similarity of the isotherms in Figure 2 to those observed in these prior studies with solid surfaces, it is likely that aggregates of surfactant molecules are also forming on the bacterial cell surface. A number of approaches have been taken to model this complex isotherm for nonionic surfactants. One study examining Triton X-100 sorption onto fine sand attempted to model the regions as two linear isotherms separated by an intermediate region with a Freundlich isotherm.25 Although this formulation adequately modeled the experimental data in that study, it is not capable of modeling the complex isotherms shown in Figure 2. Another study attempted to use the BET isotherm, but the authors concluded that the BET isotherm does not adequately reflect the S-type isotherm and suggest that this is due to micellar, rather than multilayer, adsorption.23 The ability of the multi-interaction isotherm to model the complex isotherms in Figure 2 was assessed by fitting (24) Louvisse, A. M. T.; Gonzalez, G. Adsorption of nonionic surfactants on quartz in the presence of ethanol, HCL, or CaCl2. In Surfactant-based mobility control: Progress in miscible-flood enhanced oil recovery; Smith, D. H., Ed.; American Chemical Society: Washington, DC, 1988; Vol. 373. (25) Edwards, D. A.; Adeel, Z.; Luthy, R. G. Environ. Sci. Technol. 1994, 28, 1550-1560. (26) Sun, S.; Inskeep, W. P.; Boyd, S. A. Environ. Sci. Technol. 1995, 29, 903-913. (27) Zhu, L.; Yang, K.; Lou, B.; Yuan, B. Water Res. 2003, 37, 47924800. (28) Baroumi, M.; Beurroies, I.; Denoyel, R.; Said, H.; Hanna, K. Colloids Surf. A 2003, 223, 63-72. (29) Bogdanova, Y. G.; Dolzhikova, V. D.; Summ, B. D. Colloid J. 1994, 56, 544-549. (30) Zhao, J.; Brown, W. J. Phys. Chem. 1996, 100, 3775-3782. (31) Geffroy, C.; Stuart, M. A. C.; Wong, K.; Cabane, B.; Bergon, V. Langmuir 2000, 16, 6422-6430. (32) Kiraly, Z.; Borner, R. H. K.; Findenegg, G. H. Langmuir 1997, 13, 3308-3315. (33) Kibbey, T. C. G.; Hayes, K. F. Environ. Sci. Technol. 1997, 31, 1171-1177. (34) Tronel-Peyroz, E.; Schuhmann, D.; Raous, H.; Bertrand, C. J. Colloid Interface Sci. 1984, 97, 541-551. (35) Tiberg, F.; Jonsson, B.; Tang, J.-a.; Lindeman, B. Langmuir 1994, 10, 2294-2300. (36) Noordman, W. H.; Brusseau, m. L.; Janssen, D. B. Environ. Sci. Technol. 2000, 34, 832-838. (37) Somasundaran, P.; Fuerstenau, D. W. J. Phys. Chem. 1966, 70. (38) Somasundaran, P.; Krishnakumar, S. Colloids Surf. A 1994, 93, 79-95. (39) Somasundaran, P.; Krishnakumar, S. Colloids Surf. A 1997, 123-124, 491-513. (40) Rutland, M. W.; Sender, T. J. Langmuir 1993, 9, 412-418.
Langmuir, Vol. 21, No. 24, 2005 11371 Table 3. Best-Fit Langmuir Isotherm Parameters surfactant
Γmax,L (107 molecules/CFU)
KL (xCMC)
C12E4 C12E7 C12E9 C12E10 C12E23
32.2 23.4 16.0 13.0 5.68
1.63 0.992 2.92 1.65 1.96
eq 2 to the experimental data using the nonlinear parameter estimate code PEST (version 6.1, Watermark Numerical Computing). The best-fit isotherm parameters determined by PEST are provided in Table 2, and the corresponding isotherms along with the least-squares weighted residuals from the optimization process are shown in Figure 2. For comparison, the Langmuir isotherm was also fit to the data
x S ) Γmax,L KL + x
(4)
where Γmax,L is the maximum sorption level (107 molecules/ CFU) and KL is the Langmuir half-saturation constant (unitless). The best-fit Langmuir isotherm parameters determined by PEST are provided in Table 3, and the corresponding isotherms are shown as the dashed line in Figure 2. As seen in Figure 2, the multi-interaction isotherm is able to model all of the isotherms. With the exception of C12E23, which will be discussed below, the Langmuir isotherm is not capable of accurately simulating the experimental data and this is reflected in the residuals for the isotherms shown in Figure 2. In a review of the adsorption of nonionic surfactants onto polar solid surfaces, Levitz noted that there is a steep isotherm for short POE chain lengths, a more Langmuirian isotherm for longer POE chain lengths, and sigmoidal isotherms with intermediate POE chain lengths.22 These three POE chain length regions are present in Figure 2. The short chain surfactant C12E4 shows a steeper rise than can be simulated by the Langmuir isotherm. The longer chain surfactant C12E23 can be adequately modeled with the Langmuir isotherm. And, the intermediate chain lengths of C12E7, C12E9, and C12E10 show more complex interactions, with the S-shaped isotherm very evident with C12E7 and C12E9. The observations of different regions with short, intermediate, and long POE chain lengths can be put into context by considering the contributions of the monolayer and lateral interaction components of eq 2. The contributions from these two components to the total isotherm are plotted in Figure 4 for C12E4, C12E7, and C12E23. For the short-chained C12E4, formation of surface aggregates through lateral interactions rapidly dominates as the aqueous surfactant concentration increases, resulting in a steep isotherm. As the POE chain length increases, the lateral interactions begin to decrease, resulting in the S-shaped isotherm, such as that shown for C12E7. Finally, for the longer-chained C12E23, the lateral interactions play a minor role, and the overall isotherm is dominated by the Langmuir monolayer sorption. The relationship between the monolayer sorption and lateral interactions is highlighted in Figure 5, where the ratio of Γmax,2/Γmax,1 is shown to be a linear function of the POE chain length for the range of surfactants examined in this study. The one exception is C12E9, which exhibited a stronger lateral interaction component in the isotherm than the other surfactants. One possible explanation is a specific interaction of C12E9 with the bacterial cell wall or
11372
Langmuir, Vol. 21, No. 24, 2005
Brown and Al Nuaimi
Figure 4. Interactions of two isotherm components for C12E4, C12E7, and C12E23.
Figure 5. Ratio of lateral-interaction sorption (Γmax,2 in 107 molecules/CFU) to monolayer sorption (Γmax,1 in 107 molecules/ CFU) is a linear function of the POE chain length for the range of surfactants examined in this study. The one exception is C12E9 (hollow symbol) which exhibited a stronger lateral interaction.
cell surface macromolecules. In a prior study with culture DGB01o, minimal interactions of C12E4, C12E7, C12E10, and C12E23 with the cell wall were observed for the experimental time duration used in the current study.12 However, C12E9 was not part of that prior study, so the potential for bacterial cell wall interactions to result in the higher Γmax,2/Γmax,1 ratio with C12E9 currently remains unknown.
C12E9 not withstanding, the results in Figure 5 suggest that lateral interactions and the formation of surface aggregates dominate for short POE chain lengths, whereas the monolayer sorption dominates at higher POE chain lengths. Overall, the simple isotherm model developed to represent this multi-interaction sorption provides insight into the monolayer and lateral interactions that occur during nonionic surfactant sorption as a function of both aqueous surfactant concentration and surfactant POE chain length. Ultimately, enhanced understanding of these sorption regimes as a function of concentration and POE chain length may lead to insights into the many processes involving surfactant sorption onto bacteria, including surfactant-enhanced biodegradation of hydrophobic substrates and the effects of surfactants on bacterial adhesion to surfaces and transport through porous media. Finally, the similarity of the isotherms observed in this study with bacteria and those observed in prior studies of surfactant sorption onto solid surfaces indicates that the heterogeneous nature of the bacterial cell surface did not significantly affect the isotherm shape and that the simple isotherm model developed in this study may be applicable to nonionic surfactant sorption onto surfaces other than bacteria. Acknowledgment. The authors gratefully acknowledge their valuable support during this work. This project was funded by the National Science Foundation through CAREER Grant BES-0134362 and in part by United Arab Emirates University. LA051388K
