Anal. Chem. 2001, 73, 5796-5804
Variations in Coupled Water, Viscoelastic Properties, and Film Thickness of a Mefp-1 Protein Film during Adsorption and Cross-Linking: A Quartz Crystal Microbalance with Dissipation Monitoring, Ellipsometry, and Surface Plasmon Resonance Study Fredrik Ho 1o 1 k* and Bengt Kasemo
Department of Applied Physics, Chalmers University of Technology, SE-412 96 Go¨teborg, Sweden Tommy Nylander
Department of Physical Chemistry, Lund University, SE-221 00 Lund, Sweden Camilla Fant, Kristin Sott, and Hans Elwing
Department of Molecular and Cell Biology, Go¨teborg University, Box 462, SE 405 30 Go¨teborg, Sweden
We have measured the time-resolved adsorption kinetics of the mussel adhesive protein (Mefp-1) on a nonpolar, methyl-terminated (thiolated) gold surface, using three independent techniques: quartz crystal microbalance with dissipation monitoring (QCM-D), surface plasmon resonance, and ellipsometry. The QCM-D and ellipsometry data shows that, after adsorption to saturation of Mefp-1, cross-linking of the protein layer using NaIO4 transforms it from an extended (∼20 nm), water-rich, and hydrogellike state to a much thinner (∼5 nm), compact, and less water-rich state. Furthermore, we show how quantitative data about the thickness, shear elastic modulus, and shear viscosity of the protein film can be obtained with the QCM-D technique, even beyond the Sauerbrey regime, if frequency (f) and energy dissipation (D) measurements measured at multiple harmonics are combined with theoretical simulations using a Voight-based viscoelastic model. The modeling result was confirmed by substituting H2O for D2O. As expected, the D2O substitution does not influence the actual adsorption behavior, but resulted in expected differences in the estimated effective density and shear viscosity. These results provide new insight and understanding about the adsorption kinetics and crosslinking behavior of Mefp-1. They also demonstrate how the above three techniques complement each other for biomolecule adsorption studies. The quartz crystal microbalance (QCM) technique has in recent years been shown to be a sensitive and practical tool for * Corresponding author: chalmers.se.
(fax) (+46)-31-7723134; (e-mail) fredrik@fy.
5796 Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
real-time measurements of macromolecule adsorption in several liquid-phase research applications.1-32 The two most common applications are in electrochemistry (EQCM) (see refs 1-4 and references therein), and biotechnology,5-8 to measure for example protein adsorption kinetics,9,11,12 antibody-antigen interactions,10,13-15 nucleotide hybridization and nucleotide-protein interactions,16-19 and lipid vesicle adsorption20,21 including spontaneous formation of supported bilayers,22-24 as well as bacteria and cell adsorption.25-27 The merit of the QCM technique has so far primarily been the simplicity and sensitivity (in the ng‚cm-2 range) by which an adsorbed mass, ∆m, can be deduced from a linear relation (the so-called Sauerbrey relation33) between adsorbed mass and measured changes in the resonant frequency, ∆f:
∆ms-QCM ) (CQCM/n)∆f
(1)
where CQCM ()17.7 ng‚cm-2‚Hz-1 at f ) 5 MHz) is the mass sensitivity constant and n () 1, 3, ...) is the overtone number. (The subscript s reminds us in the following that we are within the Saurebrey model.) For rigid, evenly distributed, and sufficiently thin adsorbed layers, eq 1 has indeed been shown to hold to a good approxima(1) Buttry, D. A.; Ward, M. D. Chem. Rev. 1992, 92, 1355-1379. (2) Kanazawa, K. K.; Melroy, O. R. IBM J. Res. Dev. 1993, 37, 157-171. (3) Reed, C. E.; Kanazawa, K. K.; Kaufman, J. H. J. Appl. Phys. 1990, 68, 19932001. (4) Takada, K.; Diaz, D. J.; Abruna, H. D.; Cuadrado, I.; Casado, C.; Alonso, B.; Moran, M.; Losada, J. J. Am. Chem. Soc. 1997, 119, 10763-10773. (5) Ward, M. D.; Buttry, D. A. Science 1990, 249, 1000-1007. (6) Ziegler, C.; Gopel, W. Curr. Opin. Chem. Biol. 1998, 2, 585-591. (7) Collings, A. F.; Caruso, F. Rep. Prog. Phys 1997, 60, 1397-1445. (8) Kasemo, B. Curr. Opin. Solid State Mater. Sci. 1998, 3, 451-459. (9) Rickert, J.; Weiss, T.; Gopel, W. Sens. Actuators, B1996, 31, 45-50. 10.1021/ac0106501 CCC: $20.00
© 2001 American Chemical Society Published on Web 11/16/2001
tion, as demonstrated for, for example, lipid mono- or bilayers24 or alkenethiols34 adsorbed from the liquid phase. However, for sufficiently nonrigid adsorbed films, this will not be the case. The physical explanation for this “failure” of the Sauerbrey relation10 derives from two sources. The first one is connected with the propagation of the shear acoustic wave in a viscoelastic film: a sufficiently thin and rigid adsorbed film acts as a “dead” mass on the piezoelectric oscillator (∆f ∝ ∆ms), while a viscoelastic or thicker film constitutes a coupled oscillator for which ∆f is not directly proportional to ∆ms. In other words, the effectively coupled mass depends on how the oscillatory motion of the crystal propagates into and through an adsorbed viscoelastic film.5,28-30 The second source of “failure” of the Sauerbrey relation is not a real failure, but rather related to the definition of mass. The latter is in solid-liquid interface studies frequently defined as molar (or dry) mass, i.e., excluding water. In QCM measurements, water (or any other liquid or solvent molecules) may couple as an additional mass via direct hydration, viscous drag, or entrapment in cavities in the adsorbed film. This means that the layer is essentially sensed as a viscoelastic “hydrogel” composed of the macromolecules and the coupled water. (We here only treat sufficiently thin layers so that no resonance or antiresonance occurs.) The typical amount of coupled water has in different systems been shown to vary significantly depending on the nature of the film, with mass-uptake estimations between a factor of 1.5 and 4 times larger than the molar mass.9,12,13,17 The effects of coupled water on the QCM response have been theoretically (10) Ho ¨o ¨k, F.; Rodahl, M.; Brzezinski, P.; Kasemo, B. Langmuir 1998, 14, 4 (4), 729-734. (11) Ho¨o ¨k, F.; Rodahl, M.; Kasemo, B.; Brzezinski, P. Proc. Natl. Acad. Sci. U.S.A. 1998, 95 (21), 12271-12276. (12) Caruso, F.; Furlong, D. N.; Kingshott, P. J. Colloid Interface Sci. 1997, 186, 129-140. (13) Muratsugu, M.; Ohta, F.; Miya, Y.; Hosokawa, T.; Kurosawa, S.; Kamo, N.; Ikeda, H. Anal. Chem. 1993, 65, 2933-2937. (14) Thompson, M.; Dhaliwal, G. K.; Arthur, C. L. Anal. Chem. 1986, 58, 1206. (15) Aberl, F.; Wolf, H.; Ko ¨sslinger, C.; Drost, S.; Woias, P.; Koch, S. Sens. Actuators, B 1994, 18-19, 271. (16) Niikura, K.; Matsuno, H.; Okahata, Y. J. Am. Chem. Soc. 1998, 120, 85378538. (17) Fawcett, N. C.; Craven, R. D.; Zhang, P.; Evans, J. A. Anal. Chem. 1998, 70, 2876-2889. (18) Okahata, Y.; Kawase, M.; Niikura, K.; Ohtake, F.; Furusawa, H.; Ebara, Y. Anal. Chem. 1998, 70, 1288-1296. (19) Su, H.; Chong, S.; Thompson, M. Langmuir 1996, 12, 2247-2255. (20) Tjarnhage, T.; Puu, G. Colloids Surf. B 1996, 8, 39-50. (21) Steinem, C.; Janshoff, A.; Wegener, J.; Ulrich, W. P.; Willenbrink, W.; Sieber, M.; Galla, H. J. Biosens. Bioelectron. 1997, 12, 787-808. (22) Zhdanov, V. P.; Keller, C. A.; Glasmastar, K.; Kasemo, B. J. Chem. Phys. 2000, 112, 900-909. (23) Keller, C. A.; Kasemo, B. Biophys. J. 1998, 75, 1397-1402. (24) Keller, C. A.; Glasmastar, K.; Zhdanov, V. P.; Kasemo, B. Phys. Rev. Lett. 2000, 84, 5443-5446. (25) Redepenning, J.; Schlesinger, T. K. Anal. Chem. 1993, 65, 3378-3381. (26) Janshoff, A.; Wegener, J.; Sieber, M.; Galla, H. J. Eur. Biophy. J. Biophys. Lett.. 1996, 25, 93. (27) Fredriksson, C.; Kihlman, S.; Rodahl, M.; Kasemo, B. Langmuir 1998, 14, 248-251. (28) Lucklum, R.; Behling, C.; Hauptman, P. Anal. Chem. 1999, 71, 2488-2496. (29) Rodahl, M.; Hook, F.; Fredriksson, C.; Keller, C. A.; Krozer, A.; Brzezinski, P.; Voinova, M.; Kasemo, B. Faraday Discuss. 1997, 107, 229-246. (30) Bandey, H. L.; Hillman, A. R.; Brown, M. J.; Martin, S. J. Faraday Discuss. 1997, 107, 105-122 (Acoustic Waves and Interfaces). (31) Urbakh, M.; Daikhin, L. Langmuir 1994, 10, 2836-2841. (32) Zhang, C.; Schranz, S.; Lucklum, R.; Hauptmann, P. IEEE Trans. Ultrason. Ferroelectr. Frequency Control 1998, 45, 1204-1210. (33) Sauerbrey, G. Z. Phys. 1959, 155, 206-222. (34) Schneider, T. W.; Buttry, D. A. J. Am. Chem. Soc. 1993, 115, 12391-12397.
modeled for stiff, rough,31 and textured 32 QCM-sensor surfaces and successfully applied to interpret experimental results. These factors, which at first sight may appear as an undesirable complications of the QCM technique, compared to, for example, optical methods such as ellipsometry or SPR, actually provide a platform for new information about adsorbed films, not obtainable by, for example, optical techniques, when the conventional QCM measurements are complemented with Q-factor or dissipation (D) measurements (QCM-D). With a theoretical platform that can treat the elastic and inelastic components of the shear-wave propagation through an adsorbed film, new information can be obtained through QCM-D measurements.5,28-30 In the present work, this is demonstrated for the first time for adsorbed biomolecules probed in an aqueous environment using a model system where the viscoelastic properties can be varied in situ, by (bio)chemical means; namely, the adsorption and subsequent polymerization of the mussel adhesive protein, Mytilus edulis foot protein, Mefp-1 (Mw 120 kD). Mefp-1, which is composed of 75-85 repeats of the decameric unit, NH2-Ala-LysPro-Ser-Tyr-Hyp-Hyp-Thr-L-DOPA-Lys-COOH,35 is especially attractive for such a study since it, as composed of repeating units of identical decapeptides, is chemically an unusually simple protein. Moreover, it has an open flexible conformation in solution36,37 that can be easily changed by cross-linking via chemical or enzymatic oxidation using NaIO4 and catechol oxidase (Mw ∼110 000), respectively, or via chelating using, for example, Cu ions.38-40 Specifically, NaIO4 (which was used as the crosslinking agent in the present work) initiates the formation of highly reactive o-quinones, the oxidation products of o-diphenols of the DOPA residues, including formation of di-DOPA cross-links within and probably also between Mefp-1, i.e., the same reaction as performed by the naturally occurring enzyme catechol oxidase.39,40 In the experiments described here, Mefp-1 was first adsorbed in a nativelike and flexible state on a nonpolar methyl-terminated surface and then cross-linked by NaIO4.41,42 The adsorption and cross-linking of Mefp-1 were probed using a recently introduced QCM technique which simultaneously measures changes in frequency, ∆f (related to mass uptake), and changes in the energy dissipation, ∆D (cf. viscoelastic properties), named QCM-D (see refs 10 and 43 for technical details). A Voightbased viscoelastic film model,30,44,45 describing the propagation and the damping of shear-bulk acoustic waves in a single uniform viscoelastic adsorbed film in contact with a semi-infinite bulk Newtonian liquid under no-slip conditions, was used to model the QCM-D response. The QCM-D data were compared with SPR and ellipsometry data (i) to obtain independent thickness determina(35) Waite, J. H.; Tanzer, M. L. Science 1981, 212, 1038-1040. (36) Williams, T.; Marumo, K.; Waite, J. H.; Henkens, R. W. Arch. Biochem. Biophys. 1989, 269, 415-422. (37) Deacon, M. P.; Davis, S. S.; Waite, J. H.; Harding, S. E. Biochemistry 1998, 37, 14108-14112. (38) Balla, J.; Kiss, T.; Jameson, R. F. Inorg. Chem. 1992, 31, 58-62. (39) Waite, J. H. Mar. Biol. Assoc. UK 1985, 65, 359-371. (40) Waite, J. H. Int. J. Adhes. Adhes. 1987, 7, 9-14. (41) Harder, P.; Grunze, M.; Waite, J. H. J. Adhes. 2000, 73, 161-177. (42) Fant, C.; Sott, K.; Elwing, H.; Ho ¨o ¨k, F. Biofouling 2000, 16, 119-132. (43) Rodahl, M.; Ho¨o ¨k, F.; Krozer, A.; Brzezinski, P.; Kasemo, B. Rev. Sci. Instrum. 1995, 66, 3924-3930. (44) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391-396. (45) Johannsmann, D. Macromol. Chem. Phys. 1999, 200, 501-516.
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
5797
tion,46,47 (ii) to compare the structural information contained in the QCM-D energy dissipation measurements with changes in the refractive index, n, of the protein films,47 and (iii) to derive the amount of coupled water sensed via ∆f. The latter is possible for the following reason: The mass-uptake estimation from SPR and ellipsometry is based on the difference in refractive index between the adsorbed protein molecules and water displaced by the proteins upon adsorption. Water associated with the protein film via, for example, the hydration shell, is therefore essentially not included in the mass determination by these techniques.48 In contrast, it is an inherent mass detected by the QCM technique. The amount of coupled water sensed via the QCM-D technique was also verified using D2O substitution measurements. THEORY QCM-D Technique. The QCM-D technique (Q-Sense AB), described in detail elsewhere,10,49,50 allowing simultaneous measurements of ∆f and ∆D at the first, third, etc., overtone (n ) 1, 3, ...; i.e., f ) 5 MHz, 15 MHz ,...) up to n ) 7, to obtain the resonant frequencies, f5, f15, etc., and the corresponding dissipation values, D5, D15, etc., with a repetition rate of ∼1 Hz. Since the linear relation between the adsorbed mass and the change in frequency (eq 1) is not necessarily valid for viscoelastic films, inducing additional energy dissipation and exhibiting a frequency (overtone)-dependent response, this type of information is critical. This is because if the system under investigation exhibits a frequency dependence in the measured interval (5-25 MHz in our case), measurements at several harmonics will allow the set of data (at n ) 1, 3, ...) to be compared with the theoretical representations (with several unknown parameters) that must be applied in such situations.44 We have in this work used the socalled Voight-based representation44 of a viscoelastic solid, in which the adsorbed film is represented by a (frequency-dependent) complex shear modulus according to
G ) G′ + iG′′ ) µf + i2πfηf ) µ f(1 + i2πfτf)
(2)
where µf is the elastic shear (storage) modulus, ηf the shear viscosity (loss modulus), f the oscillation frequency, and τf the characteristic relaxation time of the film. The adsorbed film is further represented by a uniform thickness, df-v (the subscript -v reminds us that we are within the Voight model) and a uniform density, Ff. The adsorbed film is situated between the QCM electrode and a semi-infinite Newtonian liquid under no-slip conditions, as depicted in Figure 1. In this case, the changes in the resonant frequency, ∆f, and the dissipation factor, ∆D become
∆f ) Im(β)/2πtqFq
(3)
and (46) Kull, T.; Nylander, T.; Tiberg, F.; Wahlgren, N. M. Langmuir 1997, 13, 5141-5147. (47) Lassen, B.; Malmsten, M. J. Colloid Interface Sci. 1996, 180, 339-349. (48) Jung, L. S.; Campbell, C. T.; Chinowsky, T. M.; Mar, M. N.; Yee, S. S. Langmuir 1998, 14, 5636-5648. (49) Rodahl, M.; Ho ¨o ¨k, F.; Kasemo, B. Anal. Chem. 1996, 68, 2219-2227. (50) Rodahl, M.; Kasemo, B. Rev. Sci. Instrum. 1996, 67, 3238-3241.
5798
Analytical Chemistry, Vol. 73, No. 24, December 15, 2001
Figure 1. Schematic illustration of the geometry and the parameters used to simulate the quartz crystal covered with a viscoelastic protein film with a thickness df(-v), in contact between the sensor surface and a semi-infinite Newtonian liquid. The film is represented by an elastic modules, µf, a viscosity, ηf, and a density, Ff. The bulk liquid is represented by a density, F, and a viscosity, η.
∆D ) -Re(β)/πftqFq
(4)
where
2πfηf - iµf 1 - R exp(2ξ1df) , 2πf 1 + R exp(2ξ1df)
β ) ξ1
ξ1 2πfηf - iµf +1 ξ2 2πfηl , ξ1 ) R) ξ1 2πfηf - iµf -1 ξ2 2πfηl ξ2 )
x
x
-
(2πf)2Ff , µf + i2πηf
2πfFl ηl
i
and where Fl and ηl are the bulk-liquid density and viscosity, respectively. SPR Technique. The surface plasmon resonance (SPR) technique is based on a collective electromagnetic motion that propagates along a metal surface, associated with which there is a localized evanescent wave with a decay length of ∼200 nm. For surface-sensing purposes, advantage is made of the fact that the excitation of the surface plasmon is very sensitive to changes in the refractive index of the medium sensed by the evanescent wave in close proximity to the gold surface. The SPR is excited using monochromatic and plane-polarized light that under total internal refraction conditions is directed through a quartz prism at the interface between the quartz and a thin (∼50 nm) layer of gold. At a certain angle, Θ, of incidence, SPR is excited, resulting in a sharp minimum in the intensity of the reflected light. If the refractive index (n) of the medium outside (decay length ∼200 nm) the gold surface changes, by, for example, protein adsorption, there is a proportional change in the angle, ∆Θ, at which SPR is
created. The SPR technique thus allows real-time measurements of the mass uptake of proteins, ∆mSPR, via the relation
∆mSPR ) CSPR∆RU
(5)
CSPR has been calibrated (using a large amount of different proteins) to be 6.5 × 10-2 ng‚cm-2 for adsorption on flat surfaces.51-53 ∆RU is the measured change in response units (a dimensionless quantity that is proportional to the change in refractive index, ∆n, at the interfacial region53). More advanced analysis of the SPR response curve, using more than one wavelength54 of the incident light or detailed line-shape analysis,55 has recently been shown to potentially make an estimation of both the refractive index and the optical thickness of thin films possible. However, to achieve the latter type of information, we have instead complemented our measurements using ellipsometry, which is more established for this purpose. Ellipsometry Technique. Ellipsometry is an optical method recording the change in polarization of elliptically polarized light, when it reflects on a sample surface. If the surface is optically modified, e.g., by protein adsorption, the associated change in polarization is detected. From the changes in the ellipsometric angles (∆, ψ), the refractive index, n, and the optical thickness, de (the subscript e means ellipsometry) of the film can be deduced. In the analyses of the ellipsometry data, the systems were treated as composed of four homogeneous and optically isotropic layers, located between the substrate and the surrounding solution as described elsewhere.56,57 Knowing the substrate properties from the measurements in different ambient media, the average refractive index, ne, and the mean optical thickness, de, of the adsorbed layer were calculated numerically from the ψ and ∆, with an equation derived for the optical four-layer model used to describe the system under study (see ref 56). The ne and de values were also used to calculate the amount adsorbed (∆m) according to the formula derived by Cuypers et al.,58 based on a ratio between molecular weight and molar refractivity and a partial specific volume of the protein of 4.01 g/mL and 0.742 mL/g, respectively. At a coverage above 100 ng‚cm-2, the relative error in the measured thickness and refractive index of the protein films are less than 5%.56,57 MATERIALS AND METHODS Measurement Techniques. The SPR measurements were done using a BIAcore 2000 system in a flow cell providing laminar flow (BIAcore AB), using a flow rate of 30 µL/min. The QCM-D measurements (Q-Sense AB) were done from a static solution, in a cell designed to provide a rapid ( 70 min, ∆f and ∆D for n ) 1 have increased and decreased, respectively, to values being about 50 (∼-45 Hz) and 9% (1.5 × 10-6), respectively, of their values prior to rinsing and crosslinking. In contrast, there is only a slight decrease (
