In Situ Investigation of Coadsorption of Myoglobin and Methylene Blue

The OWG spectroscopic investigation into the respective MB and myoglobin adsorptions shows that MB adsorption is reversible to some degree but that of...
3 downloads 0 Views 204KB Size
Download PDF
778

Langmuir 2004, 20, 778-784

In Situ Investigation of Coadsorption of Myoglobin and Methylene Blue to Hydrophilic Glass by Broadband Time-Resolved Optical Waveguide Spectroscopy Zhi-mei Qi,† Naoki Matsuda,*,† Akiko Takatsu,‡ and Kenji Kato‡ Nanoarchitectonics Research Center, National Institute of Advanced Industrial Science and Technology, Tsukuba 305-8565, Japan, and Metrology Laboratory of Japan, AIST, Tsukuba 305-8563, Japan Received August 18, 2003. In Final Form: November 4, 2003 Recently, we have developed a broadband optical waveguide (OWG) spectrometer by using commercially available glass plates of tens of micrometers in thickness as the substrate-free multimode waveguides (Qi et al. Opt. Lett. 2002, 27, 2001-2003). The spectrometer having a bandwidth from 360 to 800 nm is capable of simultaneously detecting the Soret-band absorption of heme proteins and the visible absorption of organic dyes. In this article, the spectrometer was used to in situ investigate coadsorption of methylene blue (MB) and myoglobin from the mixed aqueous solution onto bare glass. Both MB and myoglobin in the mixed solution are positively charged, which makes them not only avoid the chemical interaction between each other but also easy to adsorb to hydrophilic glass. It was found that the coadsorption of MB and myoglobin occurred just in the early stage and the glass surface was finally occupied by myoglobin. The OWG spectroscopic investigation into the respective MB and myoglobin adsorptions shows that MB adsorption is reversible to some degree but that of myoglobin is irreversible. It reveals that the electrostatic binding of myoglobin to bare glass is stronger than the case of MB. Therefore, the adsorbed MB can be substituted by myoglobin. Moreover, via the electrostatic repulsion the tightly immobilized myoglobin prevents bulk MB from occupying the empty surface sites. It is the reason MB is absent from the hydrophilic glass coated with a submonolayer of myoglobin. In the article, we explained both the strong dimerization of MB at the interface and a slow decrease with time of the Soret-band absorbance after its maximum was reached. We also estimated the myoglobin coverage based on the waveguide theory. The study shows the distinguished applicability of the broadband OWG spectroscopy for in situ, real-time monitoring of the dye-protein coadsorption to silica from the mixed solution.

1. Introduction Optical waveguide (OWG) spectroscopy is a rather new and powerful technique for surface monitoring.1-16 The technique takes advantage of the evanescent field, which penetrates less than a wavelength out of the waveguide * Corresponding author. † Nanoarchitectonics Research Center, AIST. ‡ Metrology Laboratory of Japan, AIST. (1) Bradshaw, J. T.; Mendes, S. B.; Saavedra, S. S. Anal. Chem. 2002, 74, 1751. (2) Bradshaw, J. T.; Mendes, S. B.; Armstrong, N. R.; Saavedra, S. S. Anal. Chem. 2003, 75, 1080. (3) Piraud, C.; Mwarania, E. K.; Yao, J.; O’Dwyer, K.; Schiffrin, D. J.; Wilkinson, J. S. J. Lightwave Technol. 1992, 10, 693. (4) Ross, S. E.; Seliskar, C. J.; Heineman, W. R. Anal. Chem. 2000, 72, 5549. (5) Qi, Z.-M.; Matsuda, N.; Santos, J.; Itoh, K.; Takatsu, A.; Kato, K. Langmuir 2003, 19, 214. (6) Qi, Z.-M.; Matsuda, N.; Takatsu, A.; Kato, K. Isago, H. Appl. Spectrosc. 2003, 57, 871. (7) Matsuda, N.; Zheng, J.; Qing, D.; Takatsu, A.; Kato, K. Appl. Spectrosc. 2003, 57, 100. (8) Kato, K. Takatsu, A.; Matsuda, N.; Azumi, R.; Matsumoto, M. Chem. Lett 1995, 437. (9) Matsuda, N.; Takatsu, A.; Kato, K. Chem. Lett. 1996, 105. (10) Stephens, D. A.; Bohn, P. W. Anal. Chem. 1989, 61, 386. (11) Qi, Z.-M.; Matsuda, N.; Yoshida, T.; Asano, H.; Takatsu, A.; Kato, K. Opt. Lett. 2002, 27, 2001. (12) Ohno, H.; Yoneyama, S.; Nakamura, F.; Fukuda, K.; Hara, M.; Shimomura, M. Langmuir 2002, 18, 1661. (13) Umemura, T.; Kasuya, Y.; Odake, T.; Tsunoda, K. Analyst 2002, 127, 149. (14) Tsunoda, K.; Umemura, T.; Ueno, H.; Okuno, E.; Akaiwa, H. Appl. Spectrosc. 2003, 57, 1273. (15) Santos, J.; Matsuda, N.; Qi, Z.-M.; Yoshida, T.; Takatsu, A.; Kato, K. Anal. Sci. 2003, 19, 199. (16) Qi, Z.-M.; Matsuda, N.; K.; Takatsu, A.; Kato, K. J. Phys. Chem. B, 2003, 107, 6873.

surface, to selectively respond to absorption of immobilized chemical and biological molecules over a given spectral bandwidth. The OWG spectroscopy has two typical applications. One application is to in situ investigate into the interfacial behaviors of specific analytes under the conditions of different surface properties and different solution characteristics. Another application is to examine the spectral properties of thin films such as the LB films that are too thin to be effectively studied by using a UVvis spectrophotometer. The OWG spectroscopy shows the following advantages: (1) It has a high sensitivity to monolayer absorption in contrast with the conventional transmission geometry. (2) It can be used to determine the surface coverage of adsorbed molecules based on the ray optics or the waveguide theory. (3) By use of the transverse electric (TE) and transverse magnetic (TM) modes in the waveguide, the technique enables the average orientation of the adlayer to be characterized. (4) It can provide useful information on the aggregation and isomerization of molecules at the liquid/solid interface since these phenomena are usually accompanied by a shift in electronic band of the analyte. Recently, the OWG spectroscopy has been connected with electrochemistry to result in a new technique referred to as spectroelectrochemistry, aimed at examining the effect of the redox process on the spectral properties of sub/monomolecular layers immobilized on the indium tin oxide (ITO) films deposited on the waveguides.2-4 Three types of planar waveguides have been used for OWG spectroscopy, including potassium ion-exchanged (PIE) glass waveguides,3-9 multimode waveguides of tensof-micrometers thick glass plates,10-16 and single-mode

10.1021/la035522h CCC: $27.50 © 2004 American Chemical Society Published on Web 12/17/2003

Broadband Optical Waveguide Spectroscopy

thin-film waveguides.1,2 PIE waveguides are robust, lowloss, and easy to fabricate, but they give a low sensitivity to monolayer absorption due to a weak evanescent field resulting from a graded-index profile. The thin glass plates acting as multimode waveguides are commercially available and can provide a broad bandwidth as well as a high sensitivity compared with PIE waveguides. Of these waveguides, the single-mode thin-film waveguides fabricated by sol-gel or sputtering technique are most sensitive yet considerably lossy. The use of such waveguides for broadband OWG spectroscopy gives rise to the challenge of effectively coupling polychromatic light in to and out of the thin films. Bradshaw et al.1 recently used a prism and a grating as the input and output couplers to make a single-mode waveguide of the TiO2-doped silica sol-gel film a broadband spectrometer. In our laboratory the OWG spectroscopy has been developed several years ago.8,9 We use this technique mainly for the first application noted above. The analytes investigated in the previous works focus on organic dyes including methylene blue (MB),7 bromothymol blue (BTB),5 rhodamine 6G (R6G),9 and copper tetra-tert-butylphthalocyanine (CuPct).6 PIE waveguides with a pair of LaSF8 prism couplers were used in these studies. To provide the broadband OWG spectroscopy with a sensitivity enhancement, we have recently improved the OWG spectrometer by using 30- and 50-µm-thick glass plates as substratefree multimode waveguides instead of PIE ones.11 Such thin glass plate waveguides with an air substrate have the unique feature of allowing the guided mode having the modal index, N, adjustable over a range from air index (nair ) 1) to glass index (ng ) 1.52).11,17 Note that N ) ng sin θ, where θ is the angle between the internally reflected light beam and the waveguide surface normal. The greater the mode order number, the smaller the modal index and the stronger the evanescent field. Under this condition, one can readily select a higher-order mode as the probe to make the thin glass plates more sensitive to monolayer absorption than PIE waveguides. The substrate-free structure also results in a simple fiber coupling of broadband light into the thin glass plates. The improved OWG spectrometer provides a spectral window from 360 to 800 nm.11 Such a broad bandwidth could greatly expand the number of fields in which the OWG spectroscopy is employed. For example, the apparatus can be easily used for kinetic study of heme protein adsorption to glass by monitoring the Soret-band absorption at about 410 nm.15,16 In this article, the broadband OWG spectrometer was applied to the in situ, real-time monitoring of the coadsorption of heme proteins and organic dyes from the mixed aqueous solution onto hydrophilic glass. It is usually difficult to do such an OWG spectroscopic investigation because the bandwidth of conventional OWG spectrometers is narrow compared with the wavelength range from the Soret band of heme proteins to the visible band of organic dyes. Therefore, the present study is believed to be the first application of the broadband OWG spectroscopy in the study of the heme protein-organic dye coadsorption. The respective adsorptions of heme proteins and organic dyes have been researched extensively by using a variety of techniques such as quartz crystal microbalance (QCM),18,19 second harmonic generation (SHG),20,21 and integrated optical waveguide attenuated total reflection (17) Qi, Z.-M.; Matsuda, N.; Santos, J.; Takatsu, A.; Kato, K. Opt. Lett. 2002, 27, 689. (18) Lvov, Y.; Ariga, K.; Ichinose, I.; Kunitake, T. J. Am. Chem. Soc. 1995, 117, 6117. (19) Ebara, Y.; Itakura, K.; Okahata, Y. Langmuir 1996, 12, 5165. (20) Salafsky, J. S.; Eisenthal, K. B. J. Phys. Chem. B 2000, 104, 7752

Langmuir, Vol. 20, No. 3, 2004 779

Figure 1. Schematic of the broadband OWG spectrometer: (1) 30-µm-thick glass plate; (2) silicone rubber cell; (3) glycerol drop; (4) optical fiber; (5) silicone rubber strip; (6) 1-mm-thick glass substrate; (7) output light beam to be detected; (8) quartz lens; (9) xenon-arc lamp; (10) multichannel CCD detector.

(IOW-ATR).22,23 Compared with these works, the interfacial behaviors of proteins and dyes in the case of coadsorption is seldom available so far. In this study, myoglobin and methylene blue (MB) were selected as the model adsorbates for coadsorption to bare glass and the coadsorption process was followed in real time by broadband OWG spectroscopy. An interesting phenomenon, that is, the competitive coadsorption of MB and myoglobin in the early stage and the resultant occupation of the glass surface by myoglobin, was observed. To understand this phenomenon, the respective MB and myoglobin adsorptions to glass were also examined. MB dimerization at the water/glass interface was explained and the myoglobin coverage was calculated on the basis of the waveguide theory. A conclusion was derived from the study that addition of a little amount of heme proteins in aqueous solutions of cationic dyes does not affect the chemical properties of dyes but can effectively prevent dyes from adsorbing on the wall of glass vessels charged with the dye solutions. 2. Experimental Section Figure 1 shows a schematic of the OWG spectrometer used in this study. The apparatus was fabricated by attaching a 30µm-thick glass plate (Matsunami Glass Industry, Ltd., Japan) to a pair of silicone rubber strips placed on a bulk glass supporter and then mounting a silicone rubber cell (2 × 1 × 0.5 cm) onto the glass plate. Prior to the cell attachment, the hydrophilic surface of the glass plate was cleaned with alcohol. The broadband light from a Xe-arc lamp (150 W) passing through a quartz fiber (200 µm in diameter with 100 µm core) was coupled into the glass plate by touching one end of the fiber to the plate area between the rubber strips (the area is 5 mm wide) and then covering the fiber end with a small drop of glycerol. Without lateral confinement, the forward propagation of the guided light was accompanied by lateral spreading. Both light beams were emitted out of an end face of the glass plate and one beam was directed to a multichannel charge-coupled device (CCD) detector (Hamamatsu Photonics K. K., Japan) via another fiber. Note that only the central part, being 5 mm wide, of the beam was collected for detection. A polarizer, if necessary, can be fixed in (21) Higgins, D. A.; Byerly, S. K.; Abrams, M. B.; Corn, R. M. J. Phys. Chem. 1991, 95, 6984. (22) Lee, J. E.; Saavedra, S. S. Langmuir 1996, 12, 4025. (23) Edmiston, P. E.; Lee, J. E.; Cheng, S.-S.; Saavedra, S. S. J. Am. Chem. Soc. 1997, 119, 560.

780

Langmuir, Vol. 20, No. 3, 2004

Qi et al.

Figure 2. UV-vis absorption spectra of the three aqueous solutions of 1 µM myoglobin, 2 µM MB, and the mixture (1 µM myoglobin and 2 µM MB). For comparison, an OWG spectrum (gray line) of the MB adlayer, taken from Figure 3a, is also shown here. front of the CCD detector to select the polarization state of the detected light beam. The CCD detector with a wavelength resolution of ∆λ ∼ 0.8 nm over a range 200-800 nm is capable of recording signal once per second and therefore enables the time-resolved spectroscopy. Owing to the guided mode dispersion, the modal index was found to distribute over a small range with a central value of 1.44, as determined by measuring the angle between two output light beams. The value of N ∼ 1.44 is larger than the refractive index of silicone rubber (n ) 1.40) but smaller than that of glycerol (n ) 1.47). The fiber coupling used here has a good stability for a long time and thus the time courses of adsorption and desorption of analytes can be readily obtained from the broadband, time-resolved OWG spectrometer. Horse heart myoglobin (MW ) 17 800, Sigma) and MB (Kanto Chemical Co, Japan) were used as received. Three aqueous solutions of myoglobin (1 µM), MB (2 µM), and the mixture (1 µM myoglobin and 2 µM MB) were prepared by using deionized water as solvent. The values of pH for three solutions are about 5. Since the isoelectric point of myoglobin is 7, the protein in the solutions is positively charged. This was also confirmed by electrophoresis. A Shimadzu 2100 UV-vis spectrophotometer was used to examine absorption spectra of the solutions. OWG spectroscopy of the three solutions was performed at room temperature by first injecting deionized water into the cell for acquisition of a reference spectrum. After the water in the cell is exchanged for a solution sample, the sample spectrum was recorded with a time interval of 1 s. The resulting absorption spectrum at a fixed time is determined from the equation A ) log [(IR - IB)/(IS - IB)], where IR, IS, and IB are the light intensities, at a fixed wavelength, in the reference, sample, and background spectra, respectively. For the OWG spectrometer used here, the noise causes an absorbance error of ∆A ≈ (0.006 in the wavelength range 360-800 nm.

3. Results and Discussion 3.1. Measurement of the Bulk Solution Spectra. Figure 2 shows UV-vis absorption spectra for the three aqueous solutions. The spectrum for the myoglobin solution shows the Soret band at 410 nm. Although the Q-band can also be observed at 505 nm, it is much weaker than the Soret band. Therefore, the Soret-band absorption measurement enables myoglobin adsorption to be sensitively monitored by OWG spectroscopy. The spectrum for the MB solution includes a visible band at 664 nm and a short-wavelength shoulder. This is a typical electronic spectrum of monomeric MB. The lack of MB dimers and higher aggregates in the solution is evidenced by the fact that the absorbance at a fixed wavelength obeys the Lambert-Beer law over the concentration range 0-6 µM. [For comparison, an OWG spectrum of the MB adlayer, taken from Figure 3a, was included in Figure 2. It can be regarded as the electronic spectrum of dimeric MB. The

Figure 3. (a) OWG spectra of the MB adlayer obtained during the course of the dye adsorption and (b) adsorption time dependences of A598nm and A664nm as well as the ratio of A664nm to A598nm.

higher aggregates of MB absorb light mainly at λ e 570 nm.24,25] The spectrum for the mixed solution appears to be a simple overlapping of two spectra for the dye and protein solutions, and neither a new band nor a band shift was observed. It suggests absence of the chemical interaction between MB and myoglobin in the mixed solution. MB is a cationic dye and it should be difficult to make the chemical interaction with positively charged myoglobin due to the electrostatic repulsion between each other. Absence of the chemical interaction between the dye and protein in the mixed solution facilitates our analyses of OWG spectra for the mixed adlayer by comparison with those for both the dye and protein adlayers. 3.2. Monitoring of MB Adsorption to Bare Glass. Figure 3a shows OWG spectra of the dye adlayer obtained up to 250 s after the sample injection. These spectra are significantly different in shape from the bulk solution spectrum. The spectrum achieved at the first second has two bands at 664 and 609 nm that are dominated by absorptions of monomeric and dimeric MB, respectively. Both bands change with time, and finally the band at 664 nm becomes a shoulder. In the meantime, the band at 609 nm slowly moves to shorter wavelength and locates at 586 nm at 250 s. A small shift of the band from λ ) 609 to 586 nm may reveal the formation of a small amount of higher aggregates at the interface. Figure 3b shows two absorbances at 598 and 664 nm (A598nm and A664nm) and the ratio of A664nm to A598nm versus the adsorption time. Note that A598nm, mainly arising from absorption of MB dimers, is selected since it is very close to the peak (24) Jacobs, K. Y.; Schoonheydt, R. A. J. Colloid Interface Sci. 1999, 220, 103. (25) Jockusch, S.; Turro, N. J.; Tomalia, D. A. Macromolecules 1995, 28, 7416.

Broadband Optical Waveguide Spectroscopy

absorbances in the most spectra. A598nm increases with the adsorption time, indicating an increase in the number of MB dimers at the interface. Owing to the lack of MB dimers in the bulk solution, the adsorbed dimers result from MB dimerization at the water/glass interface. The value of A664nm/A598nm rapidly decreases with increasing adsorption time up to 50 s, revealing that MB dimerization at the interface is faster than the monomer adsorption and thus the initially adsorbed monomers are quickly converted into dimers. A rapid dimerization causes the number of adsorbed monomers to continuously decrease after 50 s, as indicated by a slow decrease of A664nm/A598nm with time. When the adsorption time exceeds 100 s, A664nm/ A598nm is almost fixed at 0.470, indicating an equilibrium between the numbers of MB monomers and dimers at the interface. Because the OWG spectra obtained at the equilibrium are extremely similar to the electronic absorption spectrum of dimeric MB, we consider that most of the adsorbed monomers are converted into dimers (dimerization constant for MB at the interface is therefore very large). However, a further increase of A598nm or A664nm with time after 100 s is still dependent on the monomer adsorption since it is the controlling step. Many studies on MB adsorption to glass focus on determination of the molecular orientation at the interface and investigation into the effects of the surface chemistry and solution properties on the adsorption behavior of the dye.7,13,14,21 Here the mechanism of MB dimerization at the interface was discussed on the basis of the experimental data above and reported previously. MB molecules have the π-electronic structure, and the self-organization of H-aggregates arises from both the π-π and hydrophobic interactions between MB molecules. The predominant species of MB in aqueous solutions, especially in dilute solutions, is the monomer because the electrostatic repulsion between the positive charges on the dye molecules prevents MB from aggregating (dimerization constant of 4 × 103 M-1 for MB in aqueous solution).25,26 When MB monomers adsorb from aqueous solution onto hydrophilic glass through the electrostatic attraction, their positive charges are neutralized with -OH groups on the glass surface. The charge neutralization eliminates the electrostatic repulsion between the adsorbed monomers and between the adsorbed and dissolved monomers. Without the repulsive force to be against the π-π and hydrophobic interactions between the dye molecules, the adsorbed monomers could easily self-assemble into dimers or higher aggregates by adjusting their orientations. The adsorbed monomers could also attract the bulk ones to form dimers or higher aggregates at the interface (considering the molecular orientation requirement for MB dimerization at the interface, the latter case should occur more easily than the former case). From this point of view, MB dimerization at the interface should be useful to improve the monomer adsorption. MB can also adsorb to hydrophobic glass. Tsunoda and co-workers13,14 show that the MB species adsorbed from the aqueous solution on the alkylsilane-modified glass substrates is dominated by the dye monomer other than its dimer. This can be understood with the following reason. Silanization results in a neutral glass surface that cannot cause the charge neutralization for adsorbed MB monomers. Therefore, the MB monomers on the silanized glass surface still electrostatically repel each other and also repel those in the solution, which prevents MB from dimerizing at the interface. (26) Antonov, L.; Gergov, G.; Petrov, V.; Kubista, M.; Nygren, J. Tatanta 1999, 49, 99.

Langmuir, Vol. 20, No. 3, 2004 781

Figure 4. (a) OWG spectra of the MB adlayer obtained during the course of the dye desorption and (b) dependences of A598nm and A664nm as well as the ratio of A664nm to A598nm on the desorption time.

The OWG spectroscopic investigation shows that MB adsorption to bare glass is reversible to some extent. Figure 4a shows OWG spectra of the MB adlayer achieved up to 250 s after replacement of the dye solution in the cell with deionized water. The visible-band absorption slowly weakens, reflecting a slow desorption of MB from the glass surface. MB desorption is ascribed to its solvability in water. Figure 4b shows A598nm and A664nm as well as the ratio of A664nm to A598nm as a function of desorption time. With increasing time, both absorbances slowly decrease but the value of A664nm/A598nm slightly increases. It indicates that the number of dimers at the interface decreases while the relative number of monomers to dimers slightly increases with time. The findings suggest that MB desorption takes place in a manner that MB molecules involved in the adsorbed dimers remove one by one from the surface into the water. If the desorption unit is MB dimer other than its monomer, the ratio of A664nm to A598nm should not change with time. On the basis of this analysis, MB adsorption and desorption can be expressed by the following reversible reactions (M and D represent MB monomers and dimers): ad

M(sol) y\ z M(ad) des ad

M(sol) + M(ad) y\ z D(ad) des It is noteworthy that the spectra in Figures 3a and 4a are based on the same reference spectrum obtained prior to the dye adsorption. These smooth spectra with regular time dependence demonstrate good stability of the OWG spectrometer during measurements. 3.3. Investigation of Myoglobin Adsorption to Bare Glass. By use of a new thin-film glass plate as the waveguide, the in situ, real-time monitoring of myoglobin adsorption to bare glass was performed. Figure 5a shows OWG spectra obtained with the 1 µM myoglobin solution.

782

Langmuir, Vol. 20, No. 3, 2004

Figure 5. (a) OWG spectra of the myoglobin adlayer obtained up to 250 s after the solution sample injection. (b) Time dependence of the absorbance at 409 nm.

The Soret band is located at 409 nm in the spectra. The black line in Figure 5b shows the time dependence of the peak absorbance (A409nm). A409nm rapidly increases with increasing time up to 20 s, indicating an initially quick adsorption of myoglobin to bare glass. After reaching a maximum of 0.320 at 50 s, A409nm slowly decreases with time. It means that myoglobin adsorption is not in equilibrium after yielding the largest coverage. This is different from hemoglobin adsorption from the 1 µM aqueous solution that reaches equilibrium 20 s after the sample injection.11 Evidently, such a time course of the absorbance change does not satisfy the Langmuir isotherm: A ) A0(1 - e-t/τ), where A0 is the absorbance at equilibrium and τ was defined in ref 16. A slow decrease of A409nm with time, corresponding to a slow reduction in the number of adsorbed molecules, is not a fortuitous result and it was observed repeatedly in different measurements with different protein concentrations. To our knowledge, this result and its explanation are not available in the previous studies on myoglobin adsorption.15,18,22,27 Protein adsorption at the interface is complicated and a subtle change in the OWG spectra of the protein adlayer may contain important information on the interfacial behavior of the protein. For a slow decrease with time of A409nm from its maximum, we offer a possible reason below. Myoglobin lacks symmetry and has a molecular size of 2.5 × 3.5 × 4.5 nm.18 When myoglobin in the acidic aqueous solution comes into contact with the hydrophilic glass surface, the electrostatic attraction between the protein and glass depends on the number of positively charged amino acid residue groups in the contact region of the protein surface. Providing a uniform distribution of charged groups on the protein surface, the attractive force (27) Arai, T.; Nord, W. Colloids Surf. 1990, 51, 1.

Qi et al.

between the protein and glass would be determined by the contact area of the protein surface: the larger the contact area, the stronger the attractive force. With such a model, the large asymmetry of myoglobin would result in both strongly and weakly bound protein molecules. In the early study, we have observed a slow desorption of cytochrome c (molecular size 2.5 × 2.5 × 3.7 nm) from bare glass into pure water.16 Therefore, the weakly bound myoglobin having a small contact area (e.g., 2.5 × 3.5 nm without considering the conformation change) should also be able to slowly remove from the substrate via the hydrophilic interaction and the laterally electrostatic repulsion of adjacent protein molecules. The surface sites left by removal of the weakly bound molecules could be occupied again by myoglobin via a strong adsorption. A strong adsorption requires the protein to occupy a larger area than those weakly bound. Therefore, with the substitution of strongly bound molecules for those weakly bound, the number of myoglobin on the glass surface would show a tendency to slowly decease with time. Because desorption of the weakly bound myoglobin is much slower than the protein adsorption, a slow decrease with time of the number of adsorbed molecules could be observed by OWG spectroscopy after the number reached the maximum. The experimental investigation shows that this is indeed the case: A409nm slowly decreases with time after reaching the maximum of 0.320 at t ) 50 s. It was noted above that, for hemoglobin adsorption to bare glass, the Soret-band absorbance, after reaching its maximum, was detected no longer to change with time within the detection limit of the device. This probably arises from the large size of hemoglobin (5 × 5.5 × 6.5 nm)18 that results in the tightly electrostatic binding of the protein to hydrophilic glass based on the above model (i.e., the lack of weakly adsorbed hemoglobin on the bare glass). A comparison of myoglobin adsorption with that of hemoglobin suggests that a slow decrease with time of A409nm from its maximum could not be simply ascribed to the unfolding of myoglobin at the interface that even indeed occurred (otherwise, the unfolding of hemoglobin at the interface would also cause the absorbance to decrease). In addition, the respective measurements with the TE and TM modes also show the similar decrease with time of A409nm, which confirmed that such phenomenon did not arise from the reorientation of adsorbed myoglobin (for the time courses of the polarized OWG absorbance changes at 409 nm, see Supporting Information). After the myoglobin solution in the cell was replaced with deionized water, absorption of the protein adlayer was continuously monitored by OWG spectroscopy (spectra not shown). The Soret band was observed to be similar in shape and location to those in Figure 5a. The gray line in Figure 5b shows A409nm versus the monitoring time. A nearly steady value of A409nm ) 0.250 was observed up to 250 s after the water injection. This value is smaller than the maximum of A409nm ) 0.320 achieved during the protein adsorption, which is considered to arise from the reduction in the number of myoglobin induced by substitution of strongly bound molecules for those weakly bound (the contribution of the bulk absorption to the OWG absorbance is small enough to be neglected). A steady-state absorbance of 0.250 reflects that the resulting adlayer of myoglobin is stable in water and no desorption can be detected within the detection limit of the spectrometer. A comparison between the myoglobin and MB adsorptions shows that the electrostatic binding of myoglobin to hydrophilic glass is much stronger than the case of MB. The myoglobin coverage (CS) in moles per square centimeter can be estimated by use of the pseudo-Beer

Broadband Optical Waveguide Spectroscopy

Langmuir, Vol. 20, No. 3, 2004 783

law:16,28 A ) η(CS/d)L, where η is the optical power fraction in the adlayer with a thickness d,  is the molar absorptivity, and L is the path length. In this study, L ) 2 cm and A409nm ) 0.250. With a spectrophotometer  was determined to be 1.259 × 105 M-1 cm-1 at 409 nm. η as a function of d can be calculated at 409 nm on the basis of the waveguide theory. For example, with a specific TE mode η can be obtained from the following equation:

η)

∫0dE2(x) dx/∫-∞+∞E2(x) dx

where E(x) is the transverse electric field distribution in a four-layer planar waveguide including the water superstate (n ) 1.33), the protein adlayer (n ∼ 1.40), the 30-µm glass plate (n ) 1.53), and the air substrate (n ) 1) (see Supporting Information for details). As noted in section 2, the modal index is ∼1.44. Without the use of a polarizer in the measurement, η ) (ηTE + ηTM)/2 (ηTE and ηTM correspond to the TE and TM modes). For d ) 2.5 nm, ηTE and ηTM at 409 nm are calculated to be 7.640 × 10-5 and 6.580 × 10-5. Thus, CS is determined to be 3.492 × 10-12 mol/cm2, occupying 23.3% of a full monolayer (∼15 × 10-12 mol/cm2 for a full monolayer of native myoglobin).22 With d ) 3.5 and 4.5 nm, η is calculated to be 9.930 × 10-5 and 12.740 × 10-5, respectively, and CS is obtained to be 3.499 × 10-12 and 3.506 × 10-12 mol/cm2, respectively. It is evident that the calculated CS is insensitive to d because of a linear dependence of the calculated η on d. This means that the protein coverage obtained by calculating the optical power fraction in the adlayer is almost independent of both orientations and conformational changes of individual molecules. Under this condition, the coveragecalculating method above can be applied to those analytes with unknown molecular sizes but known refractive indexes. Note that in the above calculation the myoglobin adlayer is assumed not to have dichroism. The polarized OWG spectroscopy also confirmed that the myoglobin adlayer was almost isotropic. 3.4. Characterization of the Coadsorption of Myoglobin and MB. After the investigation of myoglobin adsorption, the sample cell was detached and the glass surface used was cleaned repeatedly with alcohol-soaked tissue paper. The complete removal of myoglobin from the glass surface was confirmed with a reproducible detection of the OWG transmission spectrum. The cleaned glass surface was still hydrophilic and was used again for coadsorption of myoglobin and MB from the mixed solution. Figure 6a shows OWG spectra obtained up to 250 s after the mixed solution was introduced into the cell. The spectra detected in the early stage include the Soret band of myoglobin and the visible band of dimeric MB, evidencing coadsorption of the protein and dye at the interface. However, with the mixed solution the spectrum achieved at a fixed time is different in shape from the sum of those obtained at the same time with the individual solutions, suggesting that the interaction between MB and myoglobin at the interface affects their adsorption behaviors. Figure 6b shows the time courses of two absorbance changes at 598 and 409 nm. A598nm drastically increases, attains the maximum of 0.158 at 15 s, and then decreases. After 100 s A598nm is almost a constant of 0.065, ascribed to the sum of a slight absorption of myoglobin ( ) 3400 M-1 cm-1) and a baseline change because in this case the visible band of dimeric MB disappears from the OWG spectra. It reflects that after a rapid increase the number of adsorbed MB dimers decreases with time until (28) DeGrandpre, M.; Burgess, L. W. Anal. Chem. 1988, 60, 2582.

Figure 6. (a) OWG spectra of the mixed aqueous solution of myoglobin (1 µM) and MB (2 µM) obtained up to 250 s after the sample injection. (b) Time courses of the absorbance changes at 409 and 598 nm.

nearly zero. On the other hand, A409nm monotonically increases with time before 100 s and then tends to be a steady value of 0.318. The time course of the change in A409nm, compared with that for myoglobin adsorption from the individual solution (black line in Figure 5b), shows that the protein adsorption from the mixed solution is quite slow. This is ascribed to the obstacles of MB dimers adsorbed in the early stage. In addition, myoglobin adsorption from the individual solution resulted in a slow decrease with time of A409nm from its maximum. However, no decrease in A409nm could be observed for the protein adsorption from the mixed solution, giving two possible explanations. One explanation is that the production of weakly bound myoglobin is avoided in a slow adsorption process. It is easily understood that a slow adsorption makes the protein at the interface have sufficient time and space to change their orientations and conformations for a stable immobilization. Another explanation is that the weakly bound myoglobin still exists but a slow decrease with time of the protein coverage, induced by the substitution of strongly bound molecules for those weakly bound, cannot be detected by OWG spectroscopy due to a slow adsorption process. With A409nm ) 0.318 the protein coverage is estimated to be CS ) 4.450 × 10-12 mol/cm2, equivalent to 30% of a full monolayer of the folded protein. Moreover, refs 15 and 22 show that the formation of a full monolayer of myoglobin on the hydrophilic glass substrate requires the bulk concentration to be greater than 1 µM. Therefore, the glass surface used here is believed to still have space available for MB to adsorb. However, the lack of MB on the surface was confirmed with the disappearance of the absorption band of the dye from the OWG spectra obtained in the final stage. This result is considered to arise from the fact that the positive charges on the tightly immobilized protein molecules yield an electrostatic repulsive force for bulk MB that is stronger than the attractive force between the substrate and the dye. In other words, the heme protein adsorption results in a

784

Langmuir, Vol. 20, No. 3, 2004

positively charged submonolayer that prevents cationic dyes from adsorbing on the substrate. This conclusion was also judged by step-by-step adsorptions of myoglobin and MB from the individual solutions. We found that when the protein adlayer was first formed on the bare glass plate, the subsequent adsorption of MB on the substrate became very weak relative to the dye adsorption on a clean surface. The electrostatic interaction acts over a long range, which should be a reason for the result that the adsorbed myoglobin of 30% monolayer coverage effectively hindered MB adsorption. Another possible factor contributing to the absence of MB from the interface is that via unfolding the adsorbed myoglobin occupies a larger area of the glass surface than that estimated with the native protein (myoglobin does not have disulfide bonds and is relatively easy to unfold at the interface).29 Both MB and myoglobin are amphiphilic and the hydrophobic bonds could be formed between each other under the appropriate conditions. In the present case, the electrostatic repulsion as the basic interaction between MB and myoglobin suppresses the formation of hydrophobic bonds between them. Coadsorption of MB and myoglobin on the glass surface was observed in the early stage when the protein coverage was too low to prevent the dye from adsorbing. With increasing protein coverage, the number of empty surface sites decreases yet the electrostatic repulsion between the adsorbed protein and bulk MB is enhanced, which make MB adsorption gradually slow down and finally stop. As noted in section 3.2, MB adsorption to bare glass is reversible to some extent. When MB adsorption becomes slower than its desorption, the number of MB dimers at the interface decreases with time. From Figure 6b it can be seen that up to 15 s after the sample injection the myoglobin coverage is below 10% of a full monolayer. In this period, MB adsorption is faster than its desorption, causing the number of adsorbed MB dimers to increase with time. After 15 s the protein adsorption yields more than 10% monolayer coverage, which makes MB adsorption slower than its desorption, as reflected by a decrease of A598nm with time. In addition to suppressing adsorption of MB, the tightly immobilized myoglobin improves desorption of MB, which is evidenced with a comparison of the time course of the change in A598nm between Figures 6b and 4b. The comparison shows that with the mixed adlayer the MB desorption is faster than that with a pure dye adlayer (e.g., in the time range from 25 to 75 s the average value of ∆A/∆t is -1.445 × 10-3 s-1 in Figure 6b and -7.082 × 10-4 s-1 in Figure 4b). The above findings show that the major effect of MB adsorption on that of myoglobin arises from changes in the number of surface sites while myoglobin adsorption affects that of MB basically through the electrostatic repulsion between the (29) Min, D. J.; Winterton, L.; Andrade, J. D. Colloid Interface Sci. 1998, 197, 43.

Qi et al.

protein and dye. Although coadsorption of myoglobin and MB onto hydrophobic glass has not been investigated, the similar results could also be obtained with hydrophobic substrates because the binding of myoglobin to hydrophobic glass has been demonstrated to be stronger than that to hydrophilic one.22 4. Conclusion We have demonstrated the outstanding applicability of the broadband OWG spectroscopy for in situ, real-time investigation into the coadsorption of heme proteins and organic dyes. By use of a 30-µm-thick glass plate as the substrate-free, multimode planar waveguide, the competitive coadsorption of myoglobin and MB on the hydrophilic glass was observed in the early stage while the resultant occupation of the glass surface by myoglobin was confirmed. The lack of MB on the glass surface is ascribed to a combination of the substitution of myoglobin for the initially adsorbed MB and the electrostatic repulsion of tightly immobilized myoglobin for bulk MB that prevents the dye from adsorbing on the substrate. The electrostatic repulsion between MB and myoglobin also suppresses the possible chemical interaction between them in the mixed aqueous solution. Therefore, the present study suggests an effective method for hindering cationic organic dyes from adsorbing on the wall of the vessels charged with the dye solutions, that is, addition of a small amount of heme proteins in the dye solutions. Such a method can be used to decrease the effect of dye adsorption on the bulk concentration in the case that the dye concentration is required to be precisely controlled. In addition, the selective adsorption of myoglobin to glass can be used to separate the protein from the mixed solution. The article also shows both a simple but clear explanation of the strong dimerization of MB on the hydrophilic glass surface and the estimation of the myoglobin coverage based on calculating the optical power fraction in the adlayer. The calculated coverage was found to be almost independent of orientations and conformational changes of individual protein molecules. We also found that myoglobin adsorption from the individual solution resulted in a slow decrease with time of the coverage from its maximum. We postulate that myoglobin adsorption produces both strongly and weakly bound protein molecules and that a slow decrease of the protein coverage arises from the substitution of strongly bound protein molecules for those weakly bound. Acknowledgment. Z.-M.Q. thanks Professor S. S. Saavedra and Professor K. Tsunoda for useful discussions on the adsorption behavior of MB on glass surfaces. Supporting Information Available: Figures showing side view of a four-layer planar waveguide and time dependence of the polarized OWG absorbance. This information is available free of charge via the Internet at http://pubs.acs.org. LA035522H