3348
Langmuir 2008, 24, 3348-3357
Mucin-Electrolyte Interactions at the Solid-Liquid Interface Probed by QCM-D Zsombor Feldo¨to¨,† Torbjo¨rn Pettersson,*,†,‡ and Andra De˘ dinaite˘ †,§ Department of Chemistry, Surface Chemistry, Royal Institute of Technology, Drottning Kristinas Va¨g 51, SE-10044 Stockholm, Sweden, ForceIT, MossVa¨gen 14, SE-153 37 Ja¨rna, Sweden, and Institute for Surface Chemistry, Box 5607, SE-114 86 Stockholm, Sweden ReceiVed October 29, 2007. In Final Form: January 24, 2008 The interaction between mucin and ions has been investigated by employing the quartz crystal microbalance technique with measurement of energy dissipation. The study was partially aimed at understanding the adsorption of mucin on surfaces with different chemistry, and for this purpose, surfaces exposing COOH, OH, and CH3 groups were prepared. Mucin adsorbed to all three types of functionalized gold surfaces. Adsorption to the hydrophobic surface and to the charged hydrophilic surface (COOH) occured with high affinity despite the fact that in the latter case both mucin and the surface were negatively charged. On the uncharged hydrophilic surface exposing OH groups, the adsorption of mucin was very low. Another aim was to elucidate conformational changes induced by electrolytes on mucin layers adsorbed on hydrophobic surfaces from 30 mM NaNO3. To this end, we investigated the effect of three electrolytes with increasing cation valance: NaCl, CaCl2 and LaCl3. At low NaCl concentrations, the preadsorbed layer expands, whereas at higher concentrations of NaCl the layer becomes more compact. This swelling/compacting of the mucin layer is fully reversible for NaCl. When the mucin layer instead is exposed to CaCl2 or LaCl3, compaction is observed at 1 mM. For CaCl2, this process is only partially reversible, and for LaCl3, the changes are irreversible within the time frame of the experiment. Finally, mucin interaction with the DTAB cationic surfactant in an aqueous solution of different electrolytes was evaluated with turbidimetry measurements. It is concluded that the electrolytes used in this work screen the association between mucin and DTAB and that the effect increases with increasing cation valency.
Introduction The term “mucins” denotes a large group of glycoproteins consisting of a linear polypeptide backbone with alternating naked and densely glycosylated regions. The carbohydrate part of mucins typically constitutes 70-80% of the polymer mass.1,2 It consists of neutral, acidic, branched, and linear oligosaccharide side chains covalently bound to the polypeptide backbone. Typical sugars present in mucins are GlcNAc, GalNAc, Gal, Fuc, and sialic acids.2 Mucin molecules carry a net negative charge due to the presence of sialic acid groups and a few sulfate groups on some of the oligosaccharides, but cationic sites are also present.1,2 In general terms, mucins are large macromolecules consisting of subunits with a molecular weight of 0.25-5 million. In the native state, the molecular weight of mucin is in the range of 1-50 million.3 Mucins fulfill numerous biological functions, with most of them related to the transport, defense, and protection of bodily surfaces. For instance, they play an important role in controlling the permeability of oral mucosal surfaces.4 The molecular structure of mucins allows them to bind water effectively, and thus the presence of mucins on oral surfaces helps to keep them hydrated5 and lubricated.4 There is a considerable amount of data indicating that the carbohydrate moieties of salivary mucins * To whom correspondence should be addressed. E-mail: torbjorn@ forceit.eu. † Royal Institute of Technology. ‡ ForceIT. § Institute for Surface Chemistry. (1) Lee, S.; Mu¨ller, M.; Rezwan, K.; Spencer, N. D. Langmuir 2005, 21, 8344. (2) Strous, G. J.; Dekker, J. Crit. ReV. Biochem. Mol. Biol. 1992, 27, 57. (3) Carlstedt, I.; Sheehan, J. K.; Corfield, A. P.; Gallagher, J. T. Essays Biochem. 1985, 20, 40. (4) Adams, D. J. Dent. Res. 1975, 54, B19. (5) Tabak, L. A.; Levine, M. J.; Mandel, I. D.; Ellison, S. A. J. Oral Pathol. Med. 1982, 11, 1.
play a major role in the non-immune clearance of microbial flora.6 Alternatively, mucin can modulate oral microbial flora by promoting the attachment and growth of certain micro-organisms and clearing others.5 It is interesting that in work by Shi et al. it was demonstrated that by using mucin coatings it is possible to create surfaces with increased resistance to bacterial adhesion.7,8 The adsorption of mucin onto various types of surfaces is of great importance in biomaterial applications. Therefore, extensive studies aimed at understanding the behavior of mucins in contact with surfaces have been conducted.9-17 Proust and co-workers9-11 demonstrated that it is not possible to attain a real plateau value for bovine submaximillary mucin (BSM) adsorption even after 20 h or more. Durrer and co-workers12 showed that the adsorption of pig gastric mucin (PGM) and BSM on carboxylate- and aminofunctionalized polystyrene latexes over a broad range of pH values (3-7.5) is driven not only by electrostatic attraction. (6) Tabak, L. A. Crit. ReV. Oral Biol. Med. 1990, 1, 229. (7) Shi, L.; Ardehali, R.; Caldwell, K. D.; Valint, P. Colloids Surf., B 2000, 17, 229. (8) Shi, L.; Ardehali, R.; Valint, P.; Caldwell, K. D. Biotechnol. Lett. 2001, 23, 437. (9) Proust, J. E.; Baszkin, A.; Perez, E.; Boissonnade, M. M. Colloids Surf. 1984, 10, 43. (10) Perez, E.; Proust, J. E.; Baszkin, A.; Boissonnade, M. M. Colloids Surf. 1984, 9, 297. (11) Proust, J. E.; Baszkin, A.; Boissonnade, M. M. J. Colloid Interface Sci. 1983, 94, 421. (12) Durrer, C.; Irache, J. M.; Duchene, D.; Ponchel, G. J. Colloid Interface Sci. 1995, 170, 555. (13) Malmsten, M.; Blomberg, E.; Claesson, P.; Carlstedt, I.; Ljusegren, I. J. Colloid Interface Sci. 1992, 151, 579. (14) Lindh, L.; Glantz, P. O.; Carlstedt, I.; Wickstro¨m, C.; Arnebrant, T. Colloids Surf., B 2002, 25, 139. (15) Svensson, O.; Lindh, L.; Ca´rdenas, M.; Arnebrant, T. J. Colloid Interface Sci. 2006, 299, 608. (16) Pearce, E. I. F. Calcif. Tissue Int. 1981, 33, 395. (17) Dedinaite, A.; Lundin, M.; Macakova, L.; Auletta, T. Langmuir 2005, 21, 9502.
10.1021/la703366k CCC: $40.75 © 2008 American Chemical Society Published on Web 02/12/2008
Mucin-Electrolyte Interactions
Malmsten et al.13 investigated the adsorption of rat and pig gastric mucins, RGM and PGM, respectively, on hydrophobic substrates using ellipsometry and the interferometric surface force apparatus, demonstrating that both types of mucins adsorbed on hydrophobic surfaces and that the more charged RGM formed a thinner adsorbed layer than the less charged PGM. The adsorption of human salivary mucin MUC5B18 and BSM has been studied on hydrophobic and hydrophilic surfaces.14 It has also been shown that these mucins can form multilayers with chitosan using layerby-layer deposition methods.15,17 Because mucin is a polyelectrolyte, it is expected that salt type and concentration will affect its properties. Indeed, it has been shown that Ca2+ ions are important to the mechanism that controls the storage of mucin inside secretory cells.19 Calcium is supposed to play a role in cystic fibrosis because it is found at abnormally high concentrations in the mucus that lines the respiratory epithelia in individuals suffering from this disease.20 Calcium is complexed by mucin, and this induces a conformational change in the mucin molecule. Consequently, the mobility and diffusivity of respiratory mucus are reduced.21 It has further been demonstrated that Ca2+ increases the intrinsic viscosity of mucin present in saliva at the same time as the apparent pore size in the mucus gel is decreased.18 Interactions between mucins and multivalent ions, such as Al3+ and Cr3+, in the bulk have shown20,22 that these ions cause mucin aggregation. The same has also been observed at the air-water interface with Cr(III) complexes,23 and similar mucin-Cr(III) complexes have also been imaged with AFM on mica surfaces.24 Nevertheless, there exist only a limited number of reports concerned with the influence of metal ions on mucin, and the number of investigations dealing with mucin-metal ion interactions at the solid-liquid interface is even smaller. Thus, this topic remains poorly understood.20,22 The aim of the present study is to gain understanding of how preadsorbed mucin layers interact with mono-, di-, and trivalent cations. To this end, we have investigated the adsorption of mucin on gold surfaces modified with various ω-terminated thiols. The effect of electrolytes on these layers was further investigated using QCM-D, which allows us to obtain information on both the sensed mass (i.e., adsorbed amount together with associated water) and conformational changes. By using turbidimetry, we have also explored how inorganic cations affect the aggregation of mucin in the absence and presence of a cationic surfactant. Materials and Methods Materials. BSM (catalog number M3895) with a molecular weight of about25 7 × 106 g/mol was purchased from Sigma and used as received. The BSM, according to the manufacturer specifications, contains 12% sialic acid residues. In passing, we note that this commercial mucin contains some residues of other proteins (e.g., bovine serum albumin26). The following chemicals were used as received; pure ethanol 99.5% from Kemetyl, sodium nitrate (NaNO3) suprapur from Merck, sodium chloride (NaCl) pro analysi from Merck, calcium chloride (CaCl2) pro analysi from Merck, lanthanum (18) Raynal, B. D. E.; Hardingham, T. E.; Sheehan, J. K.; Thornton, D. J. J. Biol. Chem. 2003, 278, 28703. (19) Verdugo, P.; Deyrup-Olsen, I.; Aitken, M.; Villalon, M.; Johnson, D. J. Dent. Res. 1987, 66, 506. (20) Exley, C. J. Inorg. Biochem. 1998, 70, 195. (21) Varma, B. K.; Demers, A.; Jamieson, A. M.; Blackwell, J.; Jentoft, N. Biopolymers 1990, 29, 441. (22) Shrivastava, H. Y.; Nair, B. U. J. Biomol. Struct. Dyn. 2003, 20, 575. (23) Shrivastava, H. Y.; Dhathathreyan, A.; Nair, B. U. Chem. Phys. Lett. 2003, 367, 49. (24) Shrivastava, H. Y.; Dhathathreyan, A.; Nair, B. U. Chem. Phys. Lett. 2003, 369, 534. (25) Bastardo, L.; Claesson, P.; Brown, W. Langmuir 2002, 18, 3848. (26) Feiler, A. A.; Sahlholm, A.; Sandberg, T.; Caldwell, K. D. J. Colloid Interface Sci. 2007, 315, 475.
Langmuir, Vol. 24, No. 7, 2008 3349 chloride heptahydrate (LaCl3‚7H2O) reagent grade from Scharlau, 1-hexadecanethiol (mercaptohexadecan) 95% from Fluka, 11mercapto-undecanoic acid (mercaptodecanoic acid) 95% from Aldrich, 11-mercapto-1-undecanol (mercaptodecanol) 97% from Aldrich, and N-dodecyl-N-N-N-trimethylammonium bromide (DTAB) sigma ultra from Sigma. All solutions were made by using water pretreated with a Milli-RO 18 unit, followed by purification with a Q-PAK unit. The outgoing water (resistivity > 18 MΩ cm) was filtered through a 0.2 µm filter; the organic content of the outgoing water was less than 6 ppb, and the pH of the Milli-Q water was ca. 5.6. Methods. QCM-D measurements were performed with a D300 from Q-Sense (Gothenburg, Sweden) using quartz crystals coated with gold (Q-Sense, Gothenburg, Sweden) with a fundamental frequency of 4.95 MHz. All measurements were obtained in a special temperature-controlled room. The temperature of the measuring cell was further controlled to 23.0 ( 0.02 °C using Peltier elements. The primary data from the measurements are the change in the resonance frequency of the oscillator, ∆f, and the change in the dissipation value, ∆D, where the dissipation is defined as D ) Edis/(2πEst), with Edis being the energy dissipated during one period of oscillation and Est being the energy stored. The crystals were placed in contact with chromosulphuric acid (BIC) for 3 × 5 min prior to use, washed with Millipore water and then with 99.5% ethanol, and finally dried with a gentle jet of filtered N2 gas. In the next step, the gold crystals were placed in 1 mM mercaptohexadecan (in 99.5% ethanol) for 24 h and then transferred to 99.5% ethanol where the surfaces were stored until use, but for not longer than 1 week. Prior to use, the crystals were dried with N2 gas. The gold surface carrying a monolayer of mercaptohexadecan will be referred to as the CH3 surface, emphasizing the end group exposed to the solution. The same procedure was adopted with mercaptodecanoic acid and mercaptodecanol to produce surfaces denoted the COOH surface and the OH surface, respectively. The thiolation of the gold surfaces changes the wettability of the surfaces. The equilibrium contact angle of water was measured for the CH3 surfaces, OH surfaces, and COOH surfaces to be 102, 14, and 17°, respectively, and these values are consistent with those reported by Ederth et al., who concluded that the thiol layers were tightly packed and in the crystalline state.27,28 The first step of the experimental procedure is to obtain a stable baseline that characterizes the resonance frequency and dissipation prior to mucin adsorption. This baseline was obtained in 30 mM NaNO3, and all subsequent measurements are related to this baseline. After this, a solution containing 25 ppm mucin in 30 mM NaNO3 was injected, and the adsorption was followed for 60 min. At this point, mucin-free 30 mM NaNO3 was injected despite the fact that the adsorbed amount had not reached equilibrium. It has previously been noted that equilibrium values for the adsorbed mass of mucin are very difficult to achieve because of the presence of large mucin aggregates.10,16,29 A rinsing time of 10 min was sufficient to reach a new stable frequency reading. The same rinsing procedure was repeated once more. To study the effect of salt on the preadsorbed mucin layer, the background electrolyte concentration was changed, starting with 1 mM NaCl, and again an equilibration time of 10 min was found to be sufficient to reach a new stable reading. This was followed by injections of solutions with increasing salt concentrations (10, 50, and 100 mM, respectively). After the 100 mM step, 30 mM NaNO3 was injected to achieve the same final bulk composition as at the start of the experiment. The same procedure was repeated for solutions of CaCl2 and LaCl3. The measurements were repeated at least twice with good reproducibility. The reproducibility can be illustrated by the lines in Figures 5-7 that illustrate the results from three independent measurements of mucin adsorption to CH3 surfaces. When the electrolyte type and concentration are changed, it affects the bulk density, F, and viscosity, η. These changes induce changes (27) Ederth, T.; Claesson, P.; Liedberg, B. Langmuir 1998, 14, 4782. (28) Ederth, T.; Liedberg, B. Langmuir 2000, 16, 2177. (29) Dedinaite, A.; Bastardo, L. Langmuir 2002, 18, 9383.
3350 Langmuir, Vol. 24, No. 7, 2008
Feldo¨to¨ et al.
in frequency and dissipation that are not due to the adsorbed layer. The frequency shift can, according to Kanazawa and Gordon, be calculated as30-32 ∆fn ) xnf03/2
( ) Fs0ηs0 πµqFq
1/2
- xnf03/2
( ) Fsηs πµqFq
1/2
(1)
where n denotes the overtone number, f0 is the fundamental frequency of the quartz crystal in air (4.95 MHz), Fq is the specific density of quartz (2648 kg/m3), µq is the shear wave velocity of quartz (2.95 × 1010 kg/ms2), and subscripts s0 and s denote the reference electrolyte (in our case, 30 mM NaNO3) and the electrolyte under study, respectively. The change in dissipation due to changes in the bulk has been suggested by Rodahl and Kasemo to be given by33 1 Fqtq
∆Dn ) xn
x
ηsFs 1 - xn 2πfn Fqtq
x
ηs0Fs0 2πfn
(2)
where tq is the thickness of the quartz crystal (3.3 × 10-4 m). In the present study, these changes are substantial and have to be taken into account when evaluating the data. The frequency and dissipation are also affected by adsorption to the crystal surface and the mucin layer and by conformational changes of the adsorbed layer. The Sauerbrey equation34 is commonly used to obtain the sensed mass, ∆m, from the change in resonance frequency, ∆f, ∆m ) -
Fqtq∆fn nf0
(3)
where ∆fn is the change in resonance frequency for overtone n. The Sauerbrey equation is a good approximation when changes in the dissipation value, ∆D, are less than 10-6/5 Hz of ∆f.35,36 We note that the sensed mass includes the mass of the adsorbing species as well as changes in the mass of the solvent that oscillates with the crystal. It is also affected by the viscoelastic properties of the layer. The Johannsmann model37 is more accurate and informative because it uses QCM data from several overtones to create a regression line that derives the true sensed mass. We note that the true sensed mass still contains contributions from both the adsorbing species and the solvent oscillating with the crystal. The advantage of the Johannsmann model over the Sauerbrey method is that it accounts for the viscoelasticity of the adsorbed layer. Johannsmann derived an equation from the acoustic impedance and shear modulus of quartz resonators that relates the value of ∆f to the various properties of the resonator and the adsorbed layer
(
)
(2πf)3Ff2d3 1 δ ˆf ≈ -f0 2πfFd + Jˆ (f) 3 π xFqµq
(4)
where δ ˆf is the shift in the complex frequency, d is the thickness of the film, f is the resonance frequency of the crystal in contact with solution, and Jˆ (f) is the complex shear compliance. Equation 4 can be rewritten more conveniently by using the equivalent mass, m ˆ *, defined by m ˆ*)-
xFqµq δ ˆf 2f0
f
(5)
and one thus obtains (30) Kanazawa, K. K.; Gordon, J. G. Anal. Chim. Acta 1985, 175, 99. (31) Kanazawa, K. K.; Gordon, J. G. Anal. Chem. 1985, 57, 1770. (32) Reed, C. D.; Kanazawa, K. K.; Kaufman, J. H. J. Appl. Phys. 1990, 68, 1993. (33) Rodahl, M.; Kasemo, B. Sens. Actuators, B 1996, 37, 111. (34) Sauerbrey, G. Z. Phys. 1959, 155, 206. (35) Brewer, S. H.; Glomm, W. R.; Johnson, M. C.; Knag, M. K.; Franzen, S. Langmuir 2005, 21, 9303.
m ˆ * ) m0 + Jˆ (f)
Fq(2πf)2d2 3
(6)
The true sensed mass, m0, was calculated under the assumption that Jˆ (f) is independent of the frequency in the accessible frequency range. A plot of the equivalent mass against the square of the resonance frequency for different overtones gives the true sensed mass as the y intercept.37 A density and sound velocity analyzer, DSA 5000 from Anton Paar (Austria), was used for the density measurements that were carried out at 23 °C. The kinematic viscosity was measured at 23 °C with a Lauda PVS 1 (processor viscosity system) from Lauda, Dr. R. Wobser, GMBH & Co. KG, Lauda-Ko¨nigshofen, Germany) using an Ubbelohde dilution viscometer 0C from Schott Instruments GmbH (Mainz, Germany). The absolute viscosity was obtained by multiplying the kinematic viscosity by the density. The capillary constant was calculated from repeated measurements on Millipore water at 25 °C using the value of 0.89 × 10-3 mPa s for the viscosity of water.38 The turbidity of mucin-surfactant mixtures in water and in 10 and 100 mM NaCl, CaCl2, and LaCl3 solutions was measured using a ratio turbidimeter HACH model 18900 from Hach Company (Loveland, CO). In this turbidimeter, the cylindrical sample cell is placed in a chamber protected from stray light and illuminated with white light. It is known that in strongly associating systems, as is in our case, the order of addition may affect both the size and composition of the polyelectrolyte-surfactant aggregates formed.39 This is due to the presence of long-lived nonequilibrium states. Because of this fact, the procedure where surfactant is added to the polymer-containing solution was adhered to in all cases. Mixing was achieved by gently turning the test tube upside down five times. It is important to avoid bubble formation because it results in erroneous turbidity values. The measurements were made 15 min after mixing the solutions. Individual samples were prepared for each surfactant concentration.
Results Effect of Surface Properties on Mucin Adsorption. The true sensed mass and dissipation change observed as a result of mucin adsorption on various thiolated gold surfaces are shown in Figure 1a,b. Mucin was adsorbed from a 25 ppm solution in 30 mM NaNO3 at pH 5.5-6.0. The hydrophilic uncharged surface exposing OH groups adsorbs the least amount of mucin. In fact, the adsorption is remarkably small considering that the sensed mass includes both the mass of the mucin and that of water associated with the layer. The dissipation change due to mucin adsorption is also small (Figure 1b). A plot of ∆D vs ∆f, as in Figure 2, results in a straight line with a slope equal to -0.022 s, which closely passes through the origin. Thus, the data indicate that all adsorbing molecules adopt a similar surface conformation independently of if they arrive in the early or late stage. This is reasonable considering the very limited adsorption of mucin on the OH surface. The sensed mass due to mucin adsorption on the hydrophobic uncharged surface with exposed CH3 groups is significantly larger than on the OH surface (Figure 1a). The dissipation change due to mucin adsorption on the CH3 surface (Figure 1b) is large, suggesting the formation of an extended layer. The ∆D versus ∆f plot in Figure 2 is linear down to -60 Hz with a slope equal to -0.035 s. The higher slope as compared to that obtained for the OH surface further emphasizes that the layer formed on the (36) Ho¨o¨k, F.; Kasemo, B.; Nylander, T.; Fant, C.; Sott, K.; Elwing, H. Anal. Chem. 2001, 73, 5796. (37) Johannsmann, D.; Mathauer, K.; Wegner, G.; Knoll, W. Phys. ReV. B 1992, 46, 7808. (38) Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 87th ed.; Taylor and Francis: Boca Raton, FL, 2007. (39) Mezei, A.; Me´sza´ros, R.; Varga, I.; Gila´nyi, T. Langmuir 2007, 23, 4237.
Mucin-Electrolyte Interactions
Figure 1. QCM-D data for the adsorption of mucin on OH, COOH, and CH3 surfaces. Mucin is injected at time 0. (a) The true sensed mass according to the Johannsmann model and (b) the corresponding dissipation vs time for the third overtone.
Figure 2. ∆D vs ∆f plots of the mucin adsorption measurements illustrated in Figure 1 for the three surfaces. Straight lines are inserted to display the adsorption trends. For each curve, the time increases from left to right.
CH3 surface is more extended than that on the OH surface. However, because the initial adsorption is very rapid the instrument does not allow for data collection at low adsorption densities on this surface. Thus, it is not possible to compare the results obtained with the CH3 surface with those for the OH surface at the same sensed mass. Nevertheless, an extrapolation of the data for the CH3 surface shows that the trend line does not pass through the origin. Thus, we conclude that the molecules
Langmuir, Vol. 24, No. 7, 2008 3351
adsorbing initially, before we are able to capture any data, adopt a flatter conformation on the surface than those arriving later. This is further emphasized by the upturn in the ∆D versus ∆f plot that is observed at frequency shifts below -60 Hz. The adsorption of mucin on the hydrophilic, negatively charged surface with COOH groups exposed results in an even higher sensed mass than mucin adsorption on the hydrophobic CH3 surface. The dissipation is much higher than that on the OH surface but somewhat lower than that on the CH3 surface. Just like for the CH3 surface, an extrapolation of the data in Figure 2 does not go through the origin; the slope of this line is -0.033 s. As for the CH3 surface, the molecules that arrive first to the COOH surface adopt a flatter conformation than those arriving later. However, unlike for the CH3 surface, no additional upturn in the ∆D versus ∆f plot is noted at the highest adsorbed masses (lowest frequencies). We note that the dissipation values for mucin on the CH3 surface and on the COOH surface are considerably higher than for globular proteins that form rigid monolayers (e.g., BSA on COOH-modified gold) but lower than for globular proteins forming multilayers (e.g., myoglobin on COOH-modified gold).40 Effects of Salt on Preadsorbed Mucin Layers. The QCM-D raw data for the adsorption of mucin on a hexadecanthiol-treated gold surface, the CH3 surface, as monitored by the third overtone are shown in Figure 3. After stabilization of the baseline in 30 mM NaNO3, mucin (25 ppm, in 30 mM NaNO3) was injected at t ) 0 in Figure 3. The adsorption was interrupted after 60 min, even though equilibrium was not reached, by rinsing the surface with 30 mM NaNO3 twice. Next, the adsorbed mucin layer was subjected to aqueous LaCl3 solutions of different concentrations (1, 10, 50, and 100 mM), and the experiment was ended with the injection of the initial salt solution (30 mM NaNO3). When 1 mM LaCl3 was introduced, both the sensed mass and the dissipation decreased. Exposure to a more concentrated solution of LaCl3 (10 mM) resulted in increased sensed mass and dissipation. At 50 and 100 mM LaCl3, the sensed mass and dissipation increased even further. After finally rinsing with 30 mM NaNO3, the sensed mass and dissipation decreased to approximately the level observed in 1 mM LaCl3. Thus, the layer does not return to its initial state in 30 mM NaNO3. We note that the observed responses can be due to changes in bulk viscosity and density and/or changes in the adsorbed layer. Before interpreting the results presented above, we have to take a closer look at the changes expected from the changes in bulk solution properties. To determine the effect of changing electrolyte and electrolyte concentration on the QCM-D data, the density and viscosity for the different electrolytes were measured at 23 °C. As seen in Figure 4, both quantities increase linearly with the electrolyte concentrations. The slopes, reported in Table 1, were used to calculate the density and viscosity, respectively, for the different electrolyte concentrations used in the experiments. Our viscosity and density measurements are in good agreement with the data published elsewhere.41-44 The QCM-D response due to these changes was calculated using eqs 1 and 2. By subtracting these values from the measured changes in frequency and dissipation, the QCM-D response due to changes in the preadsorbed layer is obtained. In Figures 5-7, we refer to the total QCM-D response (40) Kaufman, E. D.; Belyea, J.; Johnson, M. C.; Nicholson, Z. M.; Ricks, J. L.; Shah, P. K.; Bayless, M.; Pettersson, T.; Feldo¨to¨, Z.; Blomberg, E.; Claesson, P.; Franzen, S. Langmuir 2007, 23, 6053. (41) Out, D. J. P.; Los, J. M. J. Solution Chem. 1980, 9, 19. (42) Isono, T. J. Chem. Eng. Data 1984, 29, 45. (43) Zhang, H.-L.; Han, S.-J. J. Chem. Eng. Data 1996, 41, 516. (44) Zhang, H.-L.; Chen, G.-H.; Han, S.-J. J. Chem. Eng. Data 1997, 42, 526.
3352 Langmuir, Vol. 24, No. 7, 2008
Figure 3. QCM-D data for mucin adsorption on a CH3 surface from 30 mM NaNO3, followed by rinsing with mucin-free 30 mM NaNO3. Next, the aqueous electrolyte solution was changed to LaCl3 at different increasing concentrations (1, 10, 50, and 100 mM, respectively). Finally, the system was rinsed with 30 mM NaNO3. (a) ∆f for the third overtone vs time and (b) dissipation vs time for the third overtone.
due to changes in bulk and adsorption layer properties as the “total response” whereas the response due to changes in the layer is referred to as “layer response”. We note that the layer response is calculated relative to 30 mM NaNO3 that was the electrolyte used during the initial adsorption and rinsing process. In Figures 5-7, the QCM-D data, obtained as described for Figure 3, using NaCl, CaCl2, and LaCl3, respectively, are presented. The results before and after accounting for bulk effects are shown. The true sensed mass is calculated according to eq 6 using the bulk corrected values as input in the Johannsmann model. As seen in Figures 5-7, the correction for changes in bulk viscosity and density is very important. Without considering bulk changes in viscosity and density, the sensed mass for the 1 mM solutions (also at 10 mM for NaCl and CaCl2) would have been underestimated whereas it would have been overestimated at the higher ionic strengths. The trend for the corrected dissipation change is opposite to that for the corrected sensed mass. In the following evaluation of the data reported in Figures 5-7, we will consider only the true sensed mass and the dissipation changes that can be associated with the response of the adsorbed layer. Johannsmann’s model reports a higher sensed mass for the mucin layer in 50 and 100 mM NaCl than in the reference solution, 30 mM NaNO3; see Figure 5. The corrections for changes in bulk properties and layer viscoelasticity are most important in 1 and 10 mM NaCl, whereas the corrections are small in 50 and
Feldo¨to¨ et al.
Figure 4. (a) Density and (b) viscosity as a function of electrolyte (NaCl, CaCl2, and LaCl3) solution concentration. Data for 30 mM NaNO3 are also shown. Table 1. Slopes of the Linear Relation for Density and Viscosity Measured at 23 °C
NaCl CaCl2 LaCl3
density slope (kg/mol)a
viscosity slope (N m s/mol)b
0.04125 0.09152 0.22545
8.39091 × 10-8 3.07191 × 10-7 6.10043 × 10-7
a Density ) (slope × concentration of electrolyte) + 997.54 (kg/m3). Viscosity ) (slope × concentration of electrolyte) + 0.00093247 (Pa s).
b
100 mM solutions. It is found that the true sensed mass increases with increasing NaCl concentration. However, the dissipation change decreases with increasing NaCl concentration. When comparing the measurements before and after exposing the preadsorbed mucin layer to the different NaCl concentrations, it is observed that the true sensed mass and the dissipation change are the same in both cases. Thus, the changes induced in the mucin layer by NaCl are reversible. When CaCl2 (Figure 6) is introduced, both the true sensed mass and the dissipation observed for the preadsobed mucin layer are lower than in the reference solution, 30 mM NaNO3. The general trend is that the true sensed mass increases slightly with increasing CaCl2 concentration in the range of 1-100 mM, whereas the dissipation change decreases with increasing CaCl2 concentration. When comparing the measurements before and after exposing the preadsorbed mucin layer to the different CaCl2 concentrations, the true sensed mass is found to be very similar whereas the dissipation change is slightly lower after exposure to CaCl2. This demonstrates that the changes induced in the
Mucin-Electrolyte Interactions
Langmuir, Vol. 24, No. 7, 2008 3353
Figure 5. (Top) Sensed mass and dissipation change at different NaCl concentrations. “Total response” is the sensed mass obtained from eq 3, and the dissipation change is directly measured. “Layer response” is obtained using eqs 1 and 2, and here the effects due to changes in bulk density and viscosity are subtracted. “True sensed mass” is obtained using the Johannsmann model that accounts for the layer viscoelasticity. The dashed lines correspond to the values for the mucin layer in contact with 30 mM NaNO3. (Bottom) The sensed mass and dissipation change for the mucin layer in 30 mM NaNO3 before and after exposure to the different concentrations of NaCl. Data obtained at 100 mM NaCl are also shown.
mucin layer by CaCl2 are not completely reversible on the time scale (60 min) of the measurements. The mucin layer response to the different LaCl3 concentrations is illustrated in Figure 7. In general, the true sensed mass and the dissipation are lower than in the reference solution, and both of these quantities increase with increasing LaCl3 concentration. The dissipation change is dramatically lower than in the reference solution for all concentrations of LaCl3. When comparing the measurements before and after exposing the preadsorbed mucin layer to the different LaCl3 solutions, the true sensed mass and the dissipation change are drastically lower after the exposure to LaCl3. Thus, LaCl3 induces irreversible changes, over the 60 min measurement period, within the adsorbed layer. Effect of Salt on Aqueous Mixtures of Mucin and DTAB. The turbidity of DTAB solutions in the absence of mucin was determined in 10 and 100 mM CaCl2 solutions at DTAB concentrations ranging from 0.05 to 135 mM (1 cmc DTAB in water ) 14.4 mM),45 and the data are shown in Figure 8b. At concentrations below 15 mM of DTAB, the turbidity is low and comparable to that of a pure electrolyte solution, demonstrating that no large aggregates are present under these conditions. Above this concentration, the turbidity is gradually increasing with increasing surfactant concentration owing to the growing micellar concentration. For a given DTAB concentration in this range, (45) Abuin, E. B.; Scaiano, J. C. J. Am. Chem. Soc. 1984, 106, 6274.
the turbidity is larger at the higher CaCl2 concentration, suggesting that larger aggregates are formed at higher CaCl2 concentrations. The turbidity of 25 ppm mucin solutions in water is around 0.4 NTU (nephelometric turbidity units), clearly higher than that of pure water, demonstrating the presence of some mucin aggregates. Such aggregates in solutions of the same Sigma BSM as employed in this study have been observed previously by dynamic light scattering.25 They have also been visualized on surfaces by AFM.29 The addition of NaCl to the mucin solutions (Figure 8a) containing a low concentration of surfactant results in no change in turbidity, whereas the addition of CaCl2 (Figure 8b) results in a small increase in turbidity. The increase in turbidity is due to the effective screening of the mucin-mucin repulsion by the Ca2+ ions, and this causes some aggregation. The effect is somewhat stronger in 100 mM CaCl2 than in 10 mM CaCl2 (Figure 8b). In the presence of LaCl3, the turbidity increase at low surfactant concentrations is even more pronounced (Figure 8c), which is due to the higher valency of the La3+ cation. These data demonstrate that the presence of multivalent ions increases mucin aggregation in the bulk solution. When DTAB is added to the mucin solution in the absence of multivalent ions, there is no measurable change in turbidity up to 2.5 mM DTAB (Figure 8a-c). Above this concentration but still well below the cmc of DTAB, the turbidity increases significantly and reaches a maximum value at a DTAB
3354 Langmuir, Vol. 24, No. 7, 2008
Feldo¨to¨ et al.
Figure 6. (Top) Sensed mass and dissipation change in different CaCl2 concentrations. “Total response” is the sensed mass obtained from eq 3, and the dissipation change is directly measured. “Layer response” is obtained using eqs 1 and 2, and here the effects due to changes in bulk density and viscosity are subtracted. “True sensed mass” is obtained according to the Johannsmann model that accounts for the layer viscoelasticity. The dashed lines correspond to the values for the mucin layer in contact with 30 mM NaNO3. (Bottom) The sensed mass and dissipation change for the mucin layer in 30 mM NaNO3 before and after exposure to the different concentrations of CaCl2. Data for 100 mM CaCl2 are also shown.
concentration of 15 mM. Upon further increases in DTAB concentration, the turbidity decreases (as seen in Figure 8a-c). Closer inspection of the turbidity curves in Figure 8a-c reveals that the multivalent metal ions are able to suppress the binding of the cationic DTAB to mucin. The peak in turbidity is strongly reduced in 10 mM LaCl3 as compared to that in the salt-free solution. At 100 mM CaCl2 and at 100 mM LaCl3, the turbidity peak has completely vanished. Clearly, increasing the ionic strength screens the DTAB-mucin electrostatic interaction, and a higher ion valency results in more efficient screening.
Discussion Driving Force for Mucin Adsorption. In line with other studies,13-15 we find that mucin adsorbs extensively to nonpolar surfaces (Figure 1). We note that mucin adsorption to nonpolar surfaces is counteracted by electrostatic forces because the adsorption results in the accumulation of negative charges in the surface region. Nevertheless, the adsorption is extensive, and this indicates the predominance of an adsorption mechanism other than electrostatics. We suggest that the most important driving force for mucin adsorption on the nonpolar surface is hydrophobic interactions (i.e., the removal of contacts between water and the hydrophobic surface as well as between water and the nonpolar parts of the mucin1,15). Negatively charged mucin also adsorbs significantly on negatively charged surfaces such as mica, silica, and hydroxy-
apatite.14,16,17 In line with this, we find that mucin readily adsorbs onto the COOH surface. We note that though both mucin and the surface carry net negative charges mucin also contains positive amino acid residues. Thus, it is suggested that electrostatic forces also promote adsorption in this case. Because our experiments were conducted in 30 mM NaNO3 where the Debye length is 1.8 nm, the added salt efficiently screens the interaction between the surface and charges on mucin that are localized in the outer region of the adsorbed layer. This means that positive patches in the mucin structure have the possibility to interact favorably with the surface without creating too strong of a repulsion between the negative charges on mucin and the surface. This hypothesis could be tested by performing adsorption studies at different ionic strengths and pH values. Indeed, Lindh et al.14 found an increased adsorption of mucin on oxidized silicon at higher ionic strengths, which is consistent with this hypothesis. The very low mucin adsorption on the hydrophilic uncharged surface is in line with the discussion above. In this case, electrostatic forces between the mucin molecules counteract adsorption, just as on the hydrophobic uncharged surface. However, in contrast to the CH3 surface, hydrophobic interactions are expected to be weak on the OH surface. One may speculate that the small adsorption observed is due to the formation of hydrogen bonds between mucin and the surface. We do not propose that the mucin-surface hydrogen bonds are stronger than those formed by mucin-water and surface water, but water
Mucin-Electrolyte Interactions
Langmuir, Vol. 24, No. 7, 2008 3355
Figure 7. (Top) Sensed mass and dissipation change in different LaCl3 concentrations. “Total response” is the sensed mass obtained from eq 3, and the dissipation change is directly measured. “Layer response” is obtained using eqs 1 and 2, and here the effects due to changes in bulk density and viscosity are subtracted. “True sensed mass” is obtained according to the Johannsmann model that accounts for the layer viscoelasticity. The dashed lines correspond to the values for the mucin layer in contact with 30 mM NaNO3. (Bottom) The sensed mass and dissipation change for the mucin layer in 30 mM NaNO3 before and after being exposed to the different concentrations of LaCl3. Data for 100 mM LaCl3 are also shown.
released as a result of adsorption will also result in the formation of new water-water hydrogen bonds. If the free-energy change associated with this change in hydrogen bonding is negative, then it would contribute to the adsorption driving force. Effect of Salt on Mucin-Surfactant Interactions. The turbidity of mucin solution as a function of DTAB concentration in water goes through a maximum (Figure 8). We attribute the increase in turbidity to mucin-DTAB association primarily driven by the electrostatic attraction between the positively charged head group of the surfactant and negative charges of the sialic acid residues present in the mucin. As a result of association, the charge of the mucin-DTAB aggregates decreases, and because of the incorporation of DTAB, hydrophobic domains form. Both factors favor aggregation. It is likely that the mucin-DTAB aggregates have zero net charge at the turbidity maximum. Indeed, the electrophoretic mobility measured at this point, 15 mM DTAB, was found to be close to zero. However, the mobility data are only qualitative, and the adsorption of mucin on the walls of the electrophoresis capillary cell complicated these measurements. The turbidity decrease upon further increase in DTAB concentration allows us to conclude that the association of mucin and DTAB proceeds above the charge neutralization point, now driven by hydrophobic interactions between the surfactant tails. Because DTAB brings excess positive charges to the mucinDTAB aggregates, deaggregation occurs, and the turbidity of the mixtures is reduced. However, it remains notably higher than
that at low (below 2.5 mM) DTAB concentrations. In these respects, the mucin-DTAB association follows the same trends as observed for many other polyelectrolyte-surfactant systems.46,47 We also note that no sedimentation was observed in any of the samples during the time span of our measurements (15-30 min). The turbidity peak decreases with increasing electrolyte concentration (Figure 8), indicating that salt addition counteracts mucin-surfactant association. The effect is more pronounced with La3+ and Ca2+ than with Na+ ions. In part, this effect is due to the screening of attractive electrostatic DTAB-mucin interactions as a result of the higher ionic strength. However, the effect does not simply scale with the ionic strength (100 mM NaCl has a lower inhibiting effect than 10 mM LaCl3), and this indicates that the multivalent ions also bind to the negative sites on mucin. Effect of Salt on Preadsorbed Mucin Layers. In the result section, we showed that for the correct interpretation of QCM-D data it is important to account for changes in bulk liquid density and viscosity with electrolyte concentration. Only after correcting the data for these bulk changes it is possible to draw conclusions about changes in the adsorbed layers due to variations in salt concentration and type. For instance, if the true sensed mass of (46) Naderi, A.; Claesson, P. M.; Bergstrom, M.; Dedinaite, A. Colloids Surf., A 2005, 253, 83. (47) Dedinaite, A.; Claesson, P. M. Langmuir 2000, 16, 1951.
3356 Langmuir, Vol. 24, No. 7, 2008
Figure 8. (a-c) Turbidity as a function of DTAB concentration in 25 ppm mucin solutions in water (b) (a) in 25 ppm mucin solutions in 10 and 100 mM NaCl (2 and 9, respectively), (b) in 10 and 100 mM CaCl2 solutions in water (4 and 0, respectively) and in mucin solutions containing 10 and 100 mM CaCl2 (2 and 9, respectively), and (c) in mucin solutions containing 10 and 100 mM LaCl3 (2 and 9, respectively).
the layer is increased, then this indicates that the amount of water within the layer that oscillates with the crystal has increased and/or that something else is adsorbing to the layer. To distinguish between different possible scenarios, one can use the information from the dissipation measurements. If the corrected dissipation increases, then this means that the adsorbed layer has entered a more extended viscoelastic state. In the second step of the analysis,
Feldo¨to¨ et al.
we used the Johannsmann model (eq 4) to calculate the true sensed mass by taking into account the viscoelasticity. Hence, a large difference between the sensed mass prior to performing the Johannsmann analysis and the true sensed mass means that the layer has a large viscoelasticity. The charge density of the mucin layer can be changed in two different ways. First, increased ionic strength results in increased ionization of weak acidic and basic groups because the repulsive interactions between these groups are screened.48 In our case, mainly sialic acid (pKa ≈ 2.6)49 residues will become more extensively deprotonated at higher ionic strengths. However, if metal ions bind to the charged groups, then this would result in decreased charge. We note that for a constant ionic strength and adsorbed amount of mucin an increased charge density of the layer is expected to result in a more extended layer structure. Now we turn our attention to what is happening with the mucin layer when it is exposed to various salt solutions. No direct ion binding to the CH3 surface is expected because this surface is uncharged and nonpolar.27,28 In Figure 3, we display results from a typical QCM-D experiment. When mucin is adsorbed to the surface from 30 mM NaNO3, the dissipation is high, indicating the formation of a non-rigid layer. The repulsion between the negatively charged sialic acids contributes to this layer structure. Because the preadsorbed mucin layer is exposed to a lower concentration of NaCl, the sensed mass decreases, and the dissipation increases (Figure 5). The decrease in the sensed mass can in principle be due to the desorption of mucin or the expulsion of water. On the basis of the rapid change in the QCM-D response and the evolution of this response with electrolyte concentration, we regard the desorption hypothesis as highly unlikely. In fact, the desorption of polymers is typically a very slow process,50 as also observed in QCM-D measurements,51 whereas the changes observed here are completed within a few minutes. Furthermore, the increase in sensed mass observed at higher salt concentrations would then indicate the further adsorption of mucin, an event that is impossible because there is no mucin in the bulk solution. The fact that the changes observed are fully reversible (Figure 5) indeed demonstrates that desorption does not occur. Thus, the adsorbed amount remains constant throughout the measurements. The sensed mass in 30 mM NaNO3 on the CH3 surface is around 4 mg/m2. Comparing this with the adsorbed mass of mucin on hydrophobized silica from 30 mM NaNO3 as determined by reflectometry (about 2 mg/m2)29 indicates that about 50% of the sensed mass is due to water that is hydrodynamically coupled to the layer. The fact that the sensed mass decreases and the dissipation increases when the NaCl concentration is decreased was unexpected because a decrease in the true sensed mass in most cases indicates a compaction and expulsion of water, which results in a decrease in dissipation. In contrast, we find that the dissipation increases, which indicates swelling of the layer. We propose that this contradiction can be resolved by considering AFM force curves measured at a similar condition between two layers of mucin.52 In 0.1 mM NaCl, very long-range steric forces are observed, extending well above 500 nm (250 nm/layer), which is above the typical decay length of the acoustic wave (200 nm)33 for the QCM-D at the third overtone. Thus, we conclude that the QCM-D in this case is insensitive to the state of the outer part (48) Healy, T. W.; White, L. R. AdV. Colloid Interface Sci. 1978, 9, 303. (49) Scheinthal, B. M.; Bettelheim, F. A. Carbohydr. Res. 1968, 6, 257. (50) Dijt, J. C.; Stuart, M. A. C.; Fleer, G. J. Macromolecules 1992, 25, 5416. (51) Plunkett, M. A.; Claesson, P. M.; Rutland, M. W. Langmuir 2002, 18, 1274. (52) Pettersson, T.; Feldo¨to¨, Z.; Claesson, P. M.; Dedinaite, A. Prog. Colloid Polym. Sci., in press, 2008.
Mucin-Electrolyte Interactions
of the layer, and the conclusion from the QCM-D measurements is that the true sensed mass of the inner region decreases as a result of the extensive swelling and the swelling also results in an increased viscoelasticity in this region. The reason for the extensive swelling is that the repulsion between the negative charges in the mucin layer is significantly less screened in 1 mM than in a 30 mM 1:1 electrolyte. When the NaCl concentration is increased, the mass increases and the dissipation decreases, and we have returned to a layer thickness that can be fully probed by QCM-D (but, of course, with a higher sensitivity for the inner part). The effect of CaCl2 on the mucin layer is illustrated in Figure 6. At 1 mM, the true sensed mass decreases, as does the dissipation. This indicates a compaction of the mucin layer, which results in the expulsion of water. This is due to electrostatic screening of the repulsion between sialic acid groups on mucin. When the concentration of CaCl2 is gradually increased up to 100 mM, the true sensed mass increases slightly whereas the dissipation decreases to some extent. The decreased dissipation suggests a progressive compaction of the mucin layer. When the layer is rinsed with 30 mM NaNO3 after being exposed to 100 mM CaCl2, the mass is similar, but the dissipation is somewhat lower than observed originally in 30 mM NaNO3. Thus, some of the calcium ions remain in the mucin layer, causing limited compaction. Finally, we turn our attention to Figure 7, where the effect of LaCl3 is illustrated. The mucin layer response to LaCl3 is similar to its response to CaCl2, with the difference being that the mass increases more as the concentration of LaCl3 is increased. When the layer is returned to 30 mM NaNO3 after exposure to 100 mM LaCl3, a significantly smaller true sensed mass and dissipation are observed as compared to the values found for the original layer in 30 mM NaNO3. Thus, some of the lanthanum ions also remain associated with the mucin layer after rinsing. These ions compact the layer (as shown by the low dissipation value); therefore, less water contributes to the true sensed mass. To summarize, the general differences between the effects of the three different electrolytes are that the changes induced by NaCl are completely reversible, whereas for CaCl2 there is a small difference between the final layer in 30 mM NaNO3 compared with the starting point in 30 mM NaNO3. The difference between the final and initial states of the layer after exposure to LaCl3 is much larger. This suggests that some Ca2+ ions are irreversibly bound (on the time scale of the experiment) to the mucin layer and that La3+ binds even more strongly to mucin.
Conclusions All electrolytes used in this work are able to suppress the binding of the DTAB cationic surfactant to mucin by screening electrostatic attractive forces. Furthermore, the multivalent ions
Langmuir, Vol. 24, No. 7, 2008 3357
also appear to be able to block the negative sites on mucin by directly binding to mucin. Mucin adsorbs extensively on nonpolar surfaces. The main driving force for adsorption on the nonpolar surface is hydrophobic interaction (the removal of contacts between water and the hydrophobic surface as well as between water and the nonpolar parts of the mucin). Mucin also adsorbs strongly to the negatively charged polar surface. We argue that despite the fact that mucin and the surface are both negatively charged electrostatic forces promote adsorption: although mucin carries a net negative charge, it contains positive amino acid residues. Thus, it seems plausible to suggest that these play an important role in the adsorption process. On the uncharged polar surface, the adsorption is low because neither hydrophobic nor electrostatic forces are important. We suggest that the limited adsorption observed on this surface is caused by changes in free energy due to changes in hydrogen bonding between mucin-water, surface-water, water-water, and mucin-surface, respectively. The preadsorbed mucin layers were exposed to NaCl, CaCl2, and LaCl3 solutions. In the case of sodium chloride ions at low concentrations, we conclude that only a part of the conformational change can be followed with the QCM-D because some tails and loops extend further away from the surface than the penetration depth of the QCM shear wave. When the concentration of NaCl is increased, the mucin layer contracts as a result of the screening of electrostatic repulsive forces. This swelling/compacting of the mucin layer is fully reversible for NaCl. In the case of CaCl2, the mucin layer becomes compacted at 1 mM. At higher ionic strengths, the mucin layer becomes even more compact. The mucin layer response is not fully reversible as demonstrated by the dissipation data. This suggests that some calcium ions remain bound to the mucin layer. In the case of LaCl3, the mucin layer behaves similarly as when CaCl2 is used. The main difference is that the compaction of the mucin layer due to exposure to LaCl3 solution is irreversible over the period of our measurements (60 min). Clearly, multivalent ions strongly affect surface layers of mucin, which is expected to have a strong negative impact on the biological function of mucin and mucus as a hydrating and lubricating coating. Acknowledgment. We thank Professor Per Claesson for valuable discussions during the preparation of the manuscript. T.P. and A.D. acknowledge financial support from the Swedish Research Council, VR. Sarah Lundgren is acknowledged for valuable discussions on the QCM bulk corrections. We also thank Professor Jan Skov Pedersen, Aarhus University, for providing access to his laboratory for the viscosity measurements. LA703366K
