Protein-Associated Water and Secondary Structure Effect Removal of

Dec 23, 2010 - While Pt and Au surfaces were oxide-free (fairly inert elements), W, Ti, and SS substrates were associated with native oxide. Differenc...
0 downloads 9 Views 3MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

Protein-Associated Water and Secondary Structure Effect Removal of Blood Proteins from Metallic Substrates Gaurav Anand, Fuming Zhang, Robert J. Linhardt, and Georges Belfort* The Howard P. Isermann Department of Chemical and Biological Engineering, and the Center for Biotechnology and Interdisciplinary Studies, Rensselaer Polytechnic Institute, Troy, New York 12180, United States Received October 18, 2010. Revised Manuscript Received November 27, 2010 Removing adsorbed protein from metals has significant health and industrial consequences. There are numerous protein-adsorption studies using model self-assembled monolayers or polymeric substrates but hardly any highresolution measurements of adsorption and removal of proteins on industrially relevant transition metals. Surgeons and ship owners desire clean metal surfaces to reduce transmission of disease via surgical instruments and minimize surface fouling (to reduce friction and corrosion), respectively. A major finding of this work is that, besides hydrophobic interaction adhesion energy, water content in an adsorbed protein layer and secondary structure of proteins determined the access and hence ability to remove adsorbed proteins from metal surfaces with a strong alkaline-surfactant solution (NaOH and 5 mg/mL SDS in PBS at pH 11). This is demonstrated with three blood proteins (bovine serum albumin, immunoglobulin, and fibrinogen) and four transition metal substrates and stainless steel (platinum (Pt), gold (Au), tungsten (W), titanium (Ti), and 316 grade stainless steel (SS)). All the metallic substrates were checked for chemical contaminations like carbon and sulfur and were characterized using X-ray photoelectron spectroscopy (XPS). While Pt and Au surfaces were oxide-free (fairly inert elements), W, Ti, and SS substrates were associated with native oxide. Difference measurements between a quartz crystal microbalance with dissipation (QCM-D) and surface plasmon resonance spectroscopy (SPR) provided a measure of the water content in the protein-adsorbed layers. Hydrophobic adhesion forces, obtained with atomic force microscopy, between the proteins and the metals correlated with the amount of the adsorbed protein-water complex. Thus, the amount of protein adsorbed decreased with Pt, Au, W, Ti and SS, in this order. Neither sessile contact angle nor surface roughness of the metal substrates was useful as predictors here. All three globular proteins behaved similarly on addition of the alkaline-surfactant cleaning solution, in that platinum and gold exhibited an increase, while tungsten, titanium, and stainless steel showed a decrease in weight. According to dissipation measurements with the QCM-D, the adsorbed layer for platinum and gold was rigid, while that for the tungsten, titanium, and stainless steel was much more flexible. The removal efficiency of adsorbed-protein by alkaline solution of SDS depended on the water content of the adsorbed layers for W, Ti, and SS, while for Pt and Au, it depended on secondary structural content. When protein adsorption was high (Pt, Au), protein-protein interactions and protein-surface interactions were dominant and the removal of protein layers was limited. Water content of the adsorbed protein layer was the determining factor for how efficiently the layer was removed by alkaline SDS when protein adsorption was low. Hence, protein-protein and protein-surface interactions were minimal and protein structure was less perturbed in comparison with those for high protein adsorption. Secondary structural content determined the efficient removal of adsorbed protein for high adsorbed amount.

Introduction Protein adsorption at interfaces is important in both medical and nonmedical applications. Understanding of protein adsorption is of critical importance in many fields like tissue compatibility, surgical implants such as with catheters and heart valves, and at the cellular interface with solid substrates. It is also relevant for biofilm formation and in bioseparations with membrane filtration and adsorptive or chromatographic columns. Researchers have proposed various mechanisms for protein adsorption, and it appears that protein adsorption is similar to adsorption of synthetic polymers from dilute solutions onto solid substrates. However there are differences due to the folded and cross-linked nature of proteins.1 Protein adsorption generally involves a fast early stage with monolayer adsorption with almost flat orientation followed by a slow conformational (and orientational) *Corresponding author: e-mail [email protected]; Ph (518) 276-6948. (1) Jamadagni, S. N.; Godawat, R.; Dordick, J. S.; Garde, S. How Interfaces Affect Hydrophobically Driven Polymer Folding. J. Phys. Chem. B 2009, 113(13), 4093–4101. (2) Lee, C. S.; Belfort, G. Changing activity of ribonuclease A during adsorption: a molecular explanation. Proc. Natl. Acad. Sci. U.S.A. 1989, 86(21), 8392–6.

1830 DOI: 10.1021/la1041794

rearrangement of previously adsorbed molecules.2,3 Many physical interactions at the interface influence both initial protein attachment and subsequent conformational rearrangement on the surface. Broadly, many factors affect protein adsorption, such as electrostatic interactions, i.e., surface charge, isoelectric point of the surface and the protein, ionic strength, and double-layer effects, specific chemical interactions such as metal binding ligands, hydrophobic interactions, intra- and intermolecular protein interactions, and temperature. Also, the physical features of the surface such as roughness, rheology, and morphology may also affect protein adsorption.4-6 Immobilized metal ion chromatography (3) O’Shaughnessy, B.; Vavylonis, D. Irreversibility and polymer adsorption. Phys. Rev. Lett. 2003, 90(5), 056103. (4) Vonrecum, A. F.; Vankooten, T. G. The influence of micro-topography on cellular-response and the implications for silicone implants. J. Biomater. Sci., Polym. Ed. 1995, 7(2), 181–198. (5) Tamerler, C.; Oren, E. E.; Duman, M.; Venkatasubramanian, E.; Sarikaya, M. Adsorption kinetics of an engineered gold binding peptide by surface plasmon resonance spectroscopy and a quartz crystal microbalance. Langmuir 2006, 22(18), 7712–7718. (6) Han, M.; Sethuraman, A.; Kane, R. S.; Belfort, G. Nanometer-scale roughness having little effect on the amount or structure of adsorbed protein. Langmuir 2003, 19(23), 9868–9872.

Published on Web 12/23/2010

Langmuir 2011, 27(5), 1830–1836

Anand et al.

(IMAC) employs resins that are used for protein purification due to specific affinity between histidines and metal.7-9 Over the years, researchers have studied protein adsorption on well-defined model surfaces with well-characterized chemistries and on heterogeneous solid substrates. These include self-assembled monolayers (SAM),10,11 polymers,12,13-17 carbon nanotube,18-21 semiconductors,19 lipids,22 and metals.23-25 Currently, it is believed that the conformational changes that occur during protein adsorption at solid interfaces (inorganic or organic surfaces) play a critical role in determining their adhesion mechanism. Such conformational transitions are also thought to underlie the transfection of conformational diseases, such as (7) Garcia, A. A.; Kim, D. H.; Miles, D. R. Immobilization of silver and platinum ions for metal affinity-chromatography. React. Polym. 1994, 23(2-3), 249–259. (8) Agarwal, S.; Garcia, A. A.; Miles, D. Comparison of retention and binding behavior of dUTP and biotin-conjugated dUTP using an immobilized silver ion chromatography support. Sep. Sci. Technol. 1998, 33(1), 1–18. (9) Serafica, G. C.; Belfort, G.; Pimbley, J. Protein fractionation using fast flow immobilized metal chelate affinity membranes. Biotechnol. Bioeng. 1994, 43(1), 21–36. (10) Ostuni, E.; Chapman, R. G.; Liang, M. N.; Meluleni, G.; Pier, G.; Ingber, D. E.; Whitesides, G. M. Self-Assembled Monolayers That Resist the Adsorption of Proteins and the Adhesion of Bacterial and Mammalian Cells. Langmuir 2001, 17(20), 6336–6343. (11) Sethuraman, A.; Han, M.; Kane, R. S.; Belfort, G. Effect of surface wettability on the adhesion of proteins. Langmuir 2004, 20(18), 7779–88. (12) Sethuraman, A.; Vedantham, G.; Imoto, T.; Przybycien, T.; Belfort, G. Protein unfolding at interfaces: slow dynamics of alpha-helix to beta-sheet transition. Proteins 2004, 56(4), 669–78. (13) Zoungrana, T.; Findenegg, G. H.; Norde, W. Structure, stability, and activity of adsorbed enzymes. J. Colloid Interface Sci. 1997, 190(2), 437–448. (14) Giacomelli, C. E.; Norde, W. Influence of hydrophobic Teflon particles on the structure of amyloid beta-peptide. Biomacromolecules 2003, 4(6), 1719–1726. (15) Vermeer, A. W. P.; Giacomelli, C. E.; Norde, W. Adsorption of IgG onto hydrophobic Teflon. Differences between the F-ab and F-c domains. Biochim. Biophys. Acta 2001, 1526(1), 61–69. (16) Maste, M. C. L.; Norde, W.; Visser, A. J. W. G. Adsorption-induced conformational changes in the serine proteinase savinase: A tryptophan fluorescence and circular dichroism study. J. Colloid Interface Sci. 1997, 196 (2), 224–230. (17) Norde, W.; Giacomelli, C. E. Conformational changes in proteins at interfaces: from solution to the interface, and back. Macromol. Symp. 1999, 145, 125–136. (18) Asuri, P.; Bale, S. S.; Pangule, R. C.; Shah, D. A.; Kane, R. S.; Dordick, J. S. Structure, Function, and Stability of Enzymes Covalently Attached to SingleWalled Carbon Nanotubes. Langmuir 2007, 23(24), 12318–12321. (19) Clare, T. L.; Clare, B. H.; Nichols, B. M.; Abbott, N. L.; Hamers, R. J. Functional monolayers for improved resistance to protein adsorption: Oligo(ethylene glycol)-modified silicon and diamond surfaces. Langmuir 2005, 21(14), 6344–6355. (20) Karajanagi, S. S.; Vertegel, A. A.; Kane, R. S.; Dordick, J. S. Structure and function of enzymes adsorbed onto single-walled carbon nanotubes. Langmuir 2004, 20(26), 11594–11599. (21) Shim, M.; Kam, N. W. S.; Chen, R. J.; Li, Y. M.; Dai, H. J. Functionalization of carbon nanotubes for biocompatibility and biomolecular recognition. Nano Lett. 2002, 2(4), 285–288. (22) Litt, J.; Padala, C.; Asuri, P.; Vutukuru, S.; Athmakuri, K.; Kumar, S.; Dordick, J.; Kane, R. S. Enhancing Protein Stability by Adsorption onto Raftlike Lipid Domains. J. Am. Chem. Soc. 2009, 131(20), 7107–7111. (23) Williams, D. F.; Askill, I. N.; Smith, R. Protein adsorption and desorption phenomena on clean metal surfaces. J. Biomed. Mater. Res. 1985, 19(3), 313–320. (24) Kokh, D. B.; Huang, B.; Wade, R. C.; Winn, P. J. Modeling of Protein Adsorption on a Metal Surface: Brownian Dynamics Simulations. Biophys. J. 2009, 96(3, Suppl. 1), 298a–299a. (25) Clark, G. C. F.; Williams, D. F. The effects of proteins on metallic corrosion. J. Biomed. Mater. Res. 1982, 16(2), 125–134. (26) Miller, D. M.; Youkhana, I.; Karunaratne, W. U.; Pearce, A. Presence of protein deposits on ’cleaned’ re-usable anaesthetic equipment. Anaesthesia 2001, 56(11), 1069–72. (27) Brown, P.; Preece, M.; Brandel, J. P.; Sato, T.; McShane, L.; Zerr, I.; Fletcher, A.; Will, R. G.; Pocchiari, M.; Cashman, N. R.; d’Aignaux, J. H.; Cervenakova, L.; Fradkin, J.; Schonberger, L. B.; Collins, S. J. Iatrogenic Creutzfeldt-Jakob disease at the millennium. Neurology 2000, 55(8), 1075–1081. (28) Weissmann, C.; Enari, M.; Klohn, P. C.; Rossi, D.; Flechsig, E. Transmission of prions. J. Infect. Dis. 2002, 186(Suppl 2), S157–65. (29) Weissmann, C.; Enari, M.; Klohn, P. C.; Rossi, D.; Flechsig, E. Transmission of prions. Proc. Natl. Acad. Sci. U.S.A. 2002, 99(Suppl. 4), 16378–83. (30) Flechsig, E.; Hegyi, I.; Enari, M.; Schwarz, P.; Collinge, J.; Weissmann, C. Transmission of scrapie by steel-surface-bound prions. Mol. Med. 2001, 7(10), 679–84.

Langmuir 2011, 27(5), 1830–1836

Article

Creutzfeldt-Jakob disease (CJD), through the strong adhesion of proteins to surgical instruments.26-30 Although self-assembled monolayers are excellent model systems to study protein adsorption for mechanistic insights, they do not have practical relevance. Proteins interact directly with various metallic surfaces in medical and nonmedical applications. Most surgical instruments are metallic. Hardly any comprehensive protein-adsorption studies on metallic substrates have been reported. Proteins have the capacity to bind and release metal ions, and it has previously been reported that metals tend to corrode in the presence of proteins due to release of metal ions in physiological fluids like blood and sweat.23,25 Since little has been reported on proteins at metallic interfaces, and because of the known transmission of prion disease with metallic surgical instruments,28-30 we have chosen to study the adsorption behavior of three blood proteins (bovine serum albumin (BSA), immunoglobulin (IgG), and fibrinogen (FIB)) on five transition metal substrates (platinum (Pt), gold (Au), tungsten (W), titanium (Ti), and 316 grade stainless steel (SS)). These transition metals lie between groups 4 and 11 in the periodic table and have high coordination number values. Also, proteins are multivalent.31 Therefore, strong interaction between proteins and metallic surfaces is likely. Transition metals are commonly used in industry and specifically for medical applications. Serum albumin, immunoglobulin, and fibrinogen were chosen here because they are the main components of blood and come into contact with various metallic medical devices during surgical procedures. To track the adsorption of these blood proteins on the five metallic surfaces, several different experimental methods were used. These include a quartz crystal microbalance with dissipation (QCM-D), surface plasmon resonance spectroscopy (SPR), and atomic force microscopy (AFM) in order to obtain an in-depth understanding of protein adsorption and desorption on metal substrates.

Experimental Section Materials. All materials and reagents were used as received.

Bovine serum albumin (BSA), γ-globulin (IgG), fibrinogen (FIB), sodium dodecyl sulfate (SDS), and phosphate buffer saline (PBS) were obtained from Sigma-Aldrich Chemicals (St. Louis, MO). Atomically smooth AT-cut piezoelectric quartz sensor crystals coated with platinum, gold, tungsten, titanium, and stainless steel (316 grade) were purchased from Q-Sense, Goteborg, Sweden. Gold sensor chip for SPR measurements were from GE Healthcare, Uppsala, Sweden.

Methods. Surface Characterization. Metal Deposition on QCM Sensors. Sensor crystals coated with different metals were purchased from Q-Sense, Goteborg, Sweden. The detailed chemical composition of the metal surfaces, XPS survey reports, and metal deposition methods of the metal coating on the sensor crystals are provided in the Supporting Information (SI). Details on water content estimation, XPS reports of various metallic substrates used in this study and plots of adsorbed protein mass versus time for all the three proteins can be found in the SI. Contact Angle. Contact angle measurements were conducted using a contact angle goniometer and tensiometer (Rame-Hart Instrument Co., Netcong, NJ). Static sessile drop contact angle of distilled water drops in air were measured on clean metallic sensor crystal surfaces using the contact angle tool in DROPimage software (Rame-Hart). Reported values are the mean and variance of 10 contact angle measurements. Surface Roughness. 1 μm  1 μm sized high-resolution images of clean sensor crystals were acquired using the MFP-3D atomic force microscope (Asylum Research, Santa Barbara, CA), and the (31) Sriram, S. M.; Banerjee, R.; Kane, R. S.; Kwon, Y. T. MultivalencyAssisted Control of Intracellular Signaling Pathways: Application for UbiquitinDependent N-End Rule Pathway. Chem. Biol. 2009, 16(2), 121–131.

DOI: 10.1021/la1041794

1831

Article roughness was calculated using the built-in toolbar in IGOR software (WaveMetrics Inc., Portland, OR). The surface roughness for all the crystals was similar and varied between 2 and 3 nm. The roughness of the gold surface of the SPR chip was similar (∼3 nm) to that of the gold-coated sensor crystal and to that reported by other researchers.32 Multimolecular Force Spectroscopy. We used multimolecular force spectroscopy to characterize adhesion properties of metallic surfaces with the hydrophobic groups. The atomic force microscope (AFM) cantilever tip was modified with a 10 μm diameter gold-coated borosilicate sphere and was functionalized with a hydrophobic undecanethiol (HS(CH2)10CH3) self-assembled monolayer (SAM). The spring constant of each cantilever was recalibrated before measuring the adhesion forces using a two-step procedure. First, the slope of the contact region during force-distance measurements was used to calculate the sensitivity of the lever in nm/V, and then a “thermal tune” was performed to determine the resonant frequency of the cantilever. An algorithm in IGOR (Wavemetrics Inc.) computed the spring constant using the equipartition theorem.33 Using the Deriaguin approximation to convert adhesion forces, Fa, into energy, Ea, of interaction, between two flat surfaces (large sphere of radius, R, and flat substrate), the measured forces, Fa, were normalized by the radius, R (5 μm), of the silica sphere, such that Ea = Fa/R.34 More than 100 adhesion curves were generated at different points on the substrates, and mean values of adhesion energy and variance were calculated from Gaussian profiles fitted to a histogram of adhesion energy measurements.

Adsorption Measurements. Quartz Crystal Microbalance with Dissipation (QCM-D). To follow the amount of protein adsorbed per unit area on the various chemically modified substrates, a quartz crystal microbalance with dissipation (QCM-D) (D300 System, Q-Sense AB, G€ oteborg, Sweden) was used. QCMD is a high-resolution mass weighing device which can detect adsorbed mass to the resolution of less than a ng/cm2. Thin piezoelectric disks of quartz coated with different metals like platinum (pI = 3.335), gold (pI = 4.136), tungsten (pI = 0.5037-39), titanium (pI = 5.338,40), and alloy stainless steel (pI ∼ 3.0-4.036) were made to oscillate (fundamental frequency ∼5 MHz) mechanically in shear mode at resonant frequency by means of an oscillating electric field across the crystal. The lateral amplitude of the vibrating crystal was 1-2 nm. In a flow configuration, the sample flows radially from the center of the cell (stagnation point flow)41 to the exit at the circumference of the cell comprising a volume of about 100 μL. When the mass adsorbs on the crystal, it induced changes in the frequency, Δf, which was then related to the mass of deposited material, Δm. QCM is more than a mass weighing device as it can also be used to characterize viscoelastic properties and the structural state of (32) Edgar, C. D.; Gray, D. G. Smooth model cellulose I surfaces from nanocrystal suspensions. Cellulose 2003, 10(4), 299–306. (33) Hutter, J. L.; Bechhoefer, J. Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 1993, 64(7), 1868–73. (34) Derjaguin, B. V. Untersuchungen u€ber die Reibung und Adh€asion (Investigations over the friction and adhesion). Kolloid Z. 1934, 69, 155–164. (35) Kallay, N.; Torbic, Z.; Golic, M.; Matijevic, E. Determination of the isoelectric points of several metals by an adhesion method. J. Phys. Chem. 1991, 95 (18), 7028–7032. (36) Lefevre, G.; Cerovic, L.; Milonjic, S.; Fedoroff, M.; Finne, J.; Jaubertie, A. Determination of isoelectric points of metals and metallic alloys by adhesion of latex particles. J. Colloid Interface Sci. 2009, 337(2), 449–455. (37) Fernandes, C. M.; Senos, A. M. R.; Vieira, M. T. Particle surface properties of stainless steel-coated tungsten carbide powders. Powder Technol. 2006, 164(3), 124–129. (38) Kosmulski, M. Chemical Properties of Material Surfaces; Marcel Dekker: New York, 2001; Vol. 102. (39) Parks, G. A. Isoelectric points of solid oxides solid hydroxides and aqueous hydroxo complex systems. Chem. Rev. 1965, 65(2), 177. (40) Hsu, J.-P.; Chang, Y.-T. An experimental study of the stability of TiO2 particles in organic-water mixtures. Colloids Surf., A 2000, 161(3), 423–437. (41) Dabros, T.; van de Ven, T. G. M. A Direct Method for Studying Particle Deposition onto Solid Surfaces. Colloid Polym. Sci. 1983, 261, 694–707.

1832 DOI: 10.1021/la1041794

Anand et al. the adsorbing films. Sauerbrey derived the following relationship with no slip.42 Δm ¼ - C

Δfn n

ð1Þ

where C = 17.7 ng cm-2 Hz-1, n is the overtone number, n = 1, 3, 5, 7, ..., and fn is the frequency of the overtone. Four separate resonant frequencies (overtones, n = 1, 3, 5, and 7) were used to detect the oscillation of the shear wave through the crystal at 5, 15, 25, and 35 MHz, respectively. The data from the seventh overtone is reported, as it exhibited minimum noise. Five different metallic surfaces as AT-cut QCM-D sensor crystals were tested for adsorption from 1 μM protein solutions in PBS buffer at physiological pH of 7.40. All the sensors were newly purchased and were thoroughly cleaned by immersion in a 1:1:5 mixture of H2O2 (30%), NH3 (25%), and distilled water at 65-70 °C for 20 min followed by washing with a distilled water/ethanol mixture (50:50) and blow drying with pressurized nitrogen gas. The crystals were then irradiated with UV-ozone (Novascan Technologies, PSD-UVT, Ames, IA) for 10 min and were again rinsed with ethanol and dried under nitrogen. Crystals were first equilibrated inside the chamber in air followed by the equilibration with PBS buffer. Equilibration of the crystal with buffer was indicated by cessation of any further decrease in frequency with time. Protein solution was then introduced (6 μL/min) into the QCM chamber (chamber volume ∼100 μL),41 and a decrease in frequency due to protein adsorption onto the substrate-covered crystal was recorded with time. The duration of adsorption was kept constant at 40 min for all the substrates followed by washing with buffer for 15 min to remove any nonadsorbed and loosely attached protein molecules from the crystal surface. Alkaline solution (NaOH) of sodium dodecyl sulfate (5 mg/mL, pH 11) in PBS was then introduced to desorb surface-adsorbed protein. It was previously shown that alkaline solution of SDS at a concentration of 5 mg/mL and pH ∼ 11 was most effective in removing adsorbed protein layers from surfaces irrespective of operating temperature.43 Desorption time was kept constant at 40 min for all the protein-surface combinations. After 40 min of desorption the chamber was flushed with PBS buffer solution to remove any loosely attached protein molecules. Control experiments to monitor the adsorption of only buffer and SDS on various substrates were also performed and negligible adsorption of either was observed for all the surfaces. This assured that the change in mass was solely due to the adsorption of proteins. The small drop in mass after washing steps at t = 40 min and t = 95 min was due to the removal of excess and loosely attached (nonspecifically adsorbed) protein from the top of the crystal. Surface Plasmon Resonance Spectroscopy (SPR). SPR experiments were performed on a Biacore 3000 instrument (GE Healthcare, Uppsala, Sweden) using an Au sensor chip. Adsorption of BSA, IgG, and FIB in PBS buffer at pH 7.40 was carried out for 40 min followed by washing with PBS buffer and desorption by alkaline solution of SDS. The flow rate of protein solutions as well as alkaline SDS was maintained consistent with the QCM-D experiments, i.e., 6 μL/min. The microfluidic instrument was equipped with four flow channels (physical dimensions: length, L = 2.4 mm, width, W = 0.5 mm, height, H = 0.05 mm in each chip) with a provision for continuous flow. An automatic sample needle delivers buffer (42) Sauerbrey, G. Venvendung von Schwingquarzen sur Wagung Dunner Schichten und zur Mikrowagung -Translation - Use of vibrating quartz for weighing of thin layers and for microweighing. Z. Phys. A: Hadrons Nucl. 1959, 155(2), 206–222. (43) Karlsson, C. A. C.; Wahlgren, M. C.; Tragardh, A. C. The removal of betalactoglobulin from stainless steel surfaces at high and low temperature as influenced by the type and concentration of cleaning agent. J. Food Process Eng. 1998, 21(6), 485–501.

Langmuir 2011, 27(5), 1830–1836

Anand et al.

Article

Figure 1. Static sessile drop images of 20 μL distilled water on (a) Pt, (b) Au, (c) W, (d) Ti, and (e) stainless steel (SS) and (f ) the contact angle values on different metallic substrates. Error bar is the variance of 10 different measurements.

Figure 3. Adhesion energy between hydrophobic (undecanethiol functionalized) gold-coated 10 μm diameter borosilicate sphere and metallic substrates determined by AFM force spectroscopy in liquid (PBS buffer) mode. Each bar represents the mean obtained by a Gaussian profile fitted to a histogram of adhesion energy measurements between CH3-SAM probe and the metallic surface in PBS buffer in a set of ∼100 measurements at different points on the surface, and the error is the standard deviation. The inset shows a schematic of the experimental set up.

Figure 2. (a) High-resolution AFM image (1 μm  1 μm) obtained by height trace in contact mode of clean gold surface of QCM-D sensor crystal. (b) Roughness profile along a line chosen in arbitrary direction on the gold surface and is shown by the diagonal line in (a).

and sample to the sensor chip surface. Constant analyte concentration can be maintained by ensuring continuous flow during the measurements. All the concentrations, adsorption and desorption times, flow rates, and the entire experimental protocol were kept the same as those used for the QCM-D adsorption experiments discussed above.

Results and Discussion Surface Characterization. Contact Angle. Static sessile drop contact angles of 20 μL distilled water drops on various metallic substrates are reported in Figure 1. The sensor crystals are highly polished surfaces, and the contact angle Langmuir 2011, 27(5), 1830–1836

measurements vary from θ ∼ 35° to 55° with an error of (2°. Surface Roughness. Metal-coated sensor crystal surfaces were imaged using atomic force microscope, and the roughness for the 5 metals was of the order of 2-3 nm. Figure 2 shows raw data for the gold-coated sensor crystal. The top panel depicts the high resolution surface image generated by the height trace acquisition by the AFM, and the diagonal line is drawn arbitrarily across the surface. The surface profile (height variation) obtained from the diagonal line is given in Figure 2b. No matter which direction the line was drawn, the height values fluctuated between 2 and 3 nm. Also, the roughness toolbar of the software was used to obtain the root-mean-square value of roughness for the entire image, and it was found out to be in the same range (2-3 nm). Therefore, the metal-coated sensor crystals were highly polished and planarized surfaces, and there was no significant difference between their roughness values. Multimolecular Force Spectroscopy. Hydrophobic interactions were directly obtained by measuring the adhesion energy between an AFM cantilever tip modified with a 10 μm diameter DOI: 10.1021/la1041794

1833

Article

Figure 4. Mass of proteins adsorbed (highest, FIB; intermediate, IgG; lowest, BSA) versus time on (a) gold and (b) titanium substrates measured by QCM-D. Protein adsorption period from 0 to 40 min, washing with PBS buffer for 40-55 min, cleaning with NaOH þ SDS at pH 11 for 55-95 min, and washing with PBS for 95-110 min at pH 7.40. (c) Dry mass of protein adsorbed (lowest, BSA; intermediate, IgG; highest, FIB) on gold as a function of time as measured by surface plasmon resonance (SPR) spectroscopy.

gold-coated sphere and functionalized with hydrophobic undecanethiol SAM and the underlying flat metal-coated sensor crystal surface in wet mode. The adhesion energy was measured in PBS buffer at pH 7.40, which was also used in every experiment to prepare the protein solutions. Figure 3 shows the results of these force measurements, and it can be clearly seen that platinum and gold exhibit highest adhesion energies as compared with tungsten, titanium, and stainless steel. Adsorption. QCM-D. Typical adsorption curves for three primary blood proteins (BSA, IgG, and FIB) on gold and Ti surfaces are shown in parts a and b of Figure 4, respectively. The process involved a 40 min adsorption period followed by washing with PBS buffer (15 min), desorption with addition of alkaline solution of SDS (40 min), and again washing with PBS buffer (15 min). From the protein adsorption data in Figure 4a,b, we observe that Au and Ti behaved differently with respect to addition of the cleaning solution (alkaline SDS). For Au, a very 1834 DOI: 10.1021/la1041794

Anand et al.

Figure 5. (a) Mass of proteins (BSA, IgG, and FIB) adsorbed on different metallic substrates before addition of SDS þ NaOH (black bars) and after addition of SDS þ NaOH at pH 11 (checkered bars) as determined by QCM-D. Group of metals (W, Ti, and SS) where the protein layer desorbed by SDS þ NaOH are shaded in gray. (b) Percentage of adsorbed (slanted bars) BSA, (solid bars) IgG, and (checkered bars) FIB mass removed by alkaline SDS solution from different metallic sensor surfaces.

large increase in mass occurred on addition of alkaline SDS, while with Ti the opposite occurred; most of the adsorbed protein layer was removed from the surface. Although the adsorbed amount of each protein was different on each surface (Au and Ti), all the three proteins (Fib, IgG, and BSA) behaved similarly. Figure 5a summarizes the adsorbed amount after the first (before SDS cleaning) and last (after SDS cleaning) PBS washing step for different proteins on five metallic substrates. The black and checkered bars show the total adsorbed amount (ng/cm2) after the 40 min adsorption period and after 40 min treatment with alkaline SDS solution, respectively. Different adsorption behavior emerged for two groups of metals. One group (Pt, Au) exhibited an increase in adsorbed amount after alkaline surfactant addition, while the other group (W, Ti, SS) showed the reverse behavior. For the three proteins (BSA, IgG, and Fib), the percentage of surface-adsorbed protein removed with alkaline surfactant is shown in Figure 5b. Increasing positive removals were observed for the three proteins (BSA, IgG, and Fib) on tungsten, titanium, and stainless steel while negative removals were observed for gold and platinum. Normalized dissipation, ΔD/(Δf/n), reflects the rigidity of surface adsorbed films and has been extensively used in QCM-D studies by many researchers to characterize the viscoelastic properties of the surface Langmuir 2011, 27(5), 1830–1836

Anand et al.

Article

Figure 6. Percentage of adsorbed BSA, IgG, and FIB mass removed from gold (4), platinum (O), tungsten (g), titanium (crossed O), and stainless steel (!) surfaces versus normalized dissipation (flexibility, ΔD/(Δf/n)) of the film. n=7.

adsorbed films.44-51 Rigid films have a low value of ΔD/Δf/n while floppy or flexible films have a higher value of ΔD/(Δf/n).45-48 Figure 6 shows a plot of percentage removal efficiency of adsorbed protein films on different metallic substrates, and it appears that the removal efficiency of protein films increase with an increase in flexibility (ΔD/(Δf/n)) value for each adsorbed protein layer. However, the ease of removal was highest for FIB followed by IgG and BSA. The molar concentration of each protein in the original solution was kept constant at 1 μM, and therefore the total number of molecules of protein FIB (MW 340 kDa, pI = 5.8052) was less than that of IgG (MW 150 kDa, pI = 6.552), which in turn was less than the number of BSA (MW 66 kDa, pI = 4.7052,53) molecules. SPR. In an attempt to delineate the amount of adsorbed protein from that of coupled water, adsorption experiments using QCM-D (protein, SDS, and coupled water) and SPR (protein, SDS) were undertaken with Au and the three proteins (BSA, IgG, and Fib) (Figure 4a,c). A detailed comparison of the results for QCM-D (Figure 4a) and SPR (Figure 4c) shows that, after SDS washing between 90 and 100 min, the protein films are partially removed from the surface by SDS, and there was a net decrease in dry mass after the second PBS wash. The apparent increase in wet mass seen in Figure 4a was due to increased water content. (44) Boujday, S.; Bantegnie, A.; Briand, E.; Marnet, P. G.; Salmain, M.; Pradier, C. M. In-depth investigation of protein adsorption on gold surfaces: Correlating the structure and density to the efficiency of the sensing layer. J. Phys. Chem. B 2008, 112(21), 6708–6715. (45) Dutta, A. K.; Belfort, G. Interactions between polycationic and polyanionic layers: Changes in rigidity, charge and adsorption kinetics. Sens. Actuators, B 2009, 136(1), 60–65. (46) Dutta, A. K.; Nayak, A.; Belfort, G. Reversibly Controlling the Rigidity of Adsorbed Polycations. Macromolecules 2007, 41(2), 301–304. (47) Hook, F.; Rodahl, M.; Kasemo, B.; Brzezinski, P. Structural changes in hemoglobin during adsorption to solid surfaces: Effects of pH, ionic strength, and ligand binding. Proc. Natl. Acad. Sci. U.S.A. 1998, 95(21), 12271–12276. (48) Jordan, J. L.; Fernandez, E. J. QCM-D sensitivity to protein adsorption reversibility. Biotechnol. Bioeng. 2008, 101(4), 837–842. (49) Mingarro, I.; Abad, C.; Braco, L. Interfacial activation-based molecular bioimprinting of lipolytic enzymes. Proc. Natl. Acad. Sci. U.S.A. 1995, 92(8), 3308–3312. (50) Zelander, G. QCM-D real-time monitoring of structural changes in an adsorbed protein layer. Nature Methods 2006, 41–42. (51) Anand, G.; Sharma, S.; Kumar, S. K.; Belfort, G. Stability of Tethered Proteins. Langmuir 2009, 25(9), 4998–5005. (52) Ortega-Vinuesa, J. L.; Tengvall, P.; Lundstr€om, I. Molecular packing of HSA, IgG, and fibrinogen adsorbed on silicon by AFM imaging. Thin Solid Films 1998, 324(1-2), 257–273. (53) Bakhshayeshi, M.; Zydney, A. L. Effect of solution pH on protein transmission and membrane capacity during virus filtration. Biotechnol. Bioeng. 2008, 100(1), 108–117.

Langmuir 2011, 27(5), 1830–1836

Figure 7. Percentage of dry mass of adsorbed protein removed through addition of SDS þ NaOH from a gold surface as measured by SPR (2) and percentage of wet mass removal from gold (4), platinum (O), tungsten (g), titanium (crossed O), and stainless steel (!) surfaces as measured by QCM-D versus the composite index, n.

It has been shown earlier that SDS forms a complex with preadsorbed protein layers and fails to remove the protein layer if it is too tightly bound.54 The associated water percentages for FIB, IgG, and BSA on gold were calculated here to be 64, 59, and 50%, respectively (see Supporting Information). It is probably due to the higher associated water content of adsorbed FIB molecules that the layer was removed by SDS with greater ease than with IgG, which in turn was more easily removed than BSA. Sethuraman et al.11 showed that the propensity for proteinprotein interactions was much increased when the R-helical and random content of the secondary structural components of a protein was higher. A dimensionless parameter composite index, n, was computed by summing the percentage of R-helical and random content of a protein’s secondary structure and normalized by the molecular weight of the largest protein in the group. They showed that the tendency of proteins to destabilize by interaction with a similar surface-bound protein (and the surface) was higher for proteins with a higher value of the composite index, n. Figure 7 shows a plot of dry mass (SPR) and wet mass (QCM) removal efficiency of the three proteins (BSA, IgG, and FIB) by SDS þ NaOH from gold (solid triangles for SPR and open triangles for QCM-D) and the wet mass (open circles) removal efficiency from platinum, and it can be seen that on gold and platinum surfaces the proteins with a higher value of composite index are more difficult to remove. As a previous study suggested,11 proteins with a higher value of a composite index, like FIB, interact both with other like molecules and with the surface, preferably destabilize other like molecules, and tend to form a more aggregated and possibly dense network multilayer film as compared with proteins characterized by a lower value of the composite index, n. More rigid protein films (lower dissipation) were formed on Pt and Au as compared with the other surfaces (Figure 6), and therefore the adsorbed protein molecules had a higher probability of selfinteraction with neighboring molecules and interacting with the surface. As the composite index n is a measure of self-interaction (and interaction with the surface), we notice a correlation of (54) Green, R. J.; Su, T. J.; Lu, J. R.; Penfold, J. The Interaction between SDS and Lysozyme at the Hydrophilic Solid^aˆ’Water Interface. J. Phys. Chem. B 2001, 105(8), 1594–1602.

DOI: 10.1021/la1041794

1835

Article

Anand et al.

removal with composite index of each protein on Pt and Au (and W) surfaces. While the correlation is rather weak for W, there is no noticeable effect for Ti and stainless steel (Figure 7). For the adsorption time for 40 min, the adsorbed protein molecules seem to be well-hydrated and more native-like on W, Ti, and stainless steel surfaces, and therefore the correlation with composite index was small or negligible. Thus, secondary structure matters with metal surfaces that exhibit bind proteins strongly (Au, Pt) but less so (W) or not at all (Ti, SS) for proteins that bind proteins weakly. Therefore, removal efficiency of adsorbed protein layers by alkaline SDS solution was affected by (i) destabilization of protein secondary structure (composite index) and therefore resulted in increased protein-protein and protein-surface interactions on Pt and Au surfaces (more adsorption) and (ii) the water content absorbed onto the W, Ti, and stainless steel surfaces. Thus, the surface character of the metals, the absorbed water content of the adsorbed protein, and the structural stability of the protein (as measured by the inverse of the composite index) all affect binding and removal of proteins from metal surfaces.

Conclusions As the isoelectric point (pI) of the three proteins and the five surfaces was lower than the pH of the working PBS buffer, all the proteins and surfaces were effectively negatively charged, and therefore electrostatic attractive interactions were not expected to dominate for any of the protein-surface combinations. We have previously shown that protein adsorption was higher on metals when compared with strongly hydrophobic self-assembled monolayers.55 This was probably due to enhanced affinity between aromatic amino acid R groups (tryptophan, tyrosine, and phenylalanine) on the protein’s exterior and the metal.56-61 The reduced adsorption on hydrophobic self-assembled monolayers when compared with clean metal surfaces was attributed to the entropic elasticity of alkane chains.62,63 (55) Anand, G.; Sharma, S.; Dutta, A. K.; Kumar, S. K.; Belfort, G. Conformational Transitions of Adsorbed Proteins on Surfaces of Varying Polarity. Langmuir 26 (13), 10803-10811. (56) Johnson, R. D.; Todd, R. J.; Arnold, F. H. Multipoint binding in metalaffinity chromatography II. Effect of pH and imidazole on chromatographic retention of engineered histidine-containing cytochromes c. J. Chromatogr., A 1996, 725(2), 225–235. (57) Karlsson, M.; Carlsson, U. Protein Adsorption Orientation in the Light of Fluorescent Probes. Mapping of the Interaction between Site-Directly Labeled Human Carbonic Anhydrase II and Silica Nanoparticles. 2005, 88(5), 3536–3544. (58) Sigal, G. B.; Bamdad, C.; Barberis, A.; Strominger, J.; Whitesides, G. M. A self-assembled monolayer for the binding and study of histidine-tagged proteins by surface plasmon resonance. Anal. Chem. 1996, 68(3), 490–7. (59) Appleton, T. G.; Pesch, F. J.; Wienken, M.; Menzer, S.; Lippert, B. Linkage isomerism in square-planar complexes of platinum and palladium with histidine and derivatives. Inorg. Chem. 1992, 31(21), 4410–4419. (60) Caubet, A.; Moreno, V.; Molins, E.; Miravitlles, C. Methionine and histidine Pd(II) and Pt(II) complexes - crystal-structures and spectroscopic properties. J. Inorg. Biochem. 1992, 48(2), 135–152. (61) Ge, R.; Zhang, Y.; Sun, X.; Watt, R. M.; He, Q.-Y.; Huang, J.-D.; Wilcox, D. E.; Sun, H. Thermodynamic and Kinetic Aspects of Metal Binding to the Histidine-rich Protein, Hpn. J. Am. Chem. Soc. 2006, 128(35), 11330–11331. (62) Jeon, S. I.; Andrade, J. D. Protein--surface interactions in the presence of polyethylene oxide: II. Effect of protein size. J. Colloid Interface Sci. 1991, 142(1), 159–166. (63) Jeon, S. I.; Lee, J. H.; Andrade, J. D.; De Gennes, P. G. Protein--surface interactions in the presence of polyethylene oxide: I. Simplified theory. J. Colloid Interface Sci. 1991, 142(1), 149–158.

1836 DOI: 10.1021/la1041794

Ti, W, and SS metal layers used here were associated with the natively charged oxide layers while Pt and Au layers were oxidefree (XPS results, Supporting Information). However, the isoelectric points between pure metals and metals oxides do not change significantly.35 While titanium can have a stable titanium oxide layer, native tungsten trioxide dissolves easily in water and is washed away.64 Platinum and gold are inert and are not prone to oxidation at room temperature. Therefore, the experimental results are not expected to be significantly affected by the presence of an oxide layer. When the proteins interact with metals in bioprocessing and other operations, the metals have similar chemical composition under normal conditions (with or without oxide) to that studied here. Additionally, the XPS results also verify that the metal surfaces were free of other contaminants like carbon and sulfur. The wettability and surface roughness of the five different metallic sensor crystals were almost the same within error and therefore were not considered deciding factors during proteinadsorption measurements. However, hydrophobic interactions show distinguishably high values for platinum and gold surfaces as compared with the three other metals. Adsorbed proteins formed rigid films on Au and Pt. They were difficult to remove with an alkaline solution of SDS, despite its well-known ability to chemically remove adsorbed protein layers. The proteins adsorbed much more on Pt and Au and self-interacted. Secondary structure (and propensity to self-interact and interact with the surface) as measured by a dimensionless parameter, the composite index, n, was important for adsorption onto Au and Pt, less so onto W, and had no effect on Ti and SS (Figure 7). Proteins with a larger value of n, such as Fib, were more difficult to remove from Au and Pt as compared with proteins having lower value of n, such as IgG and BSA. For W, Ti, and SS, the water content of the adsorbed protein layers appeared to correlate with removal efficiency of the protein layers using alkaline SDS cleaning solution. This was perhaps due to enhanced solubility and diffusivity of alkaline/SDS in the protein film and thereby an increased efficiency to break the dominant noncovalent interactions among the protein molecules and between them and the metal surfaces. Also, not surprisingly, protein secondary structure, and hence inherent stability, influences the ability of the cleaning SDS/ NaOH solution to remove proteins from metal surfaces. Acknowledgment. We acknowledge the support of U.S. Department of Energy, DOE (DE-FG02-90ER14114 and DOE DE-FG02-07ER46429), and the National Science Foundation (Grant CTS-94-00610). We also thank Prof. Joel Plawsky and Arya Chatterjee of Rensselaer Polytechnic Institute for help with the contact angle measurements. G.A. is grateful for the Howard P. Isermann fellowship. Supporting Information Available: Details on water content estimation, XPS reports of various metallic substrates used in this study and plots of adsorbed protein mass versus time for all the three proteins. This material is available free of charge via the Internet at http://pubs.acs.org. (64) Lassner, E.; Schubert, W.-D. Tungsten: Properties, Chemistry, Technology of the Element, Alloys, and Chemical Compounds; Kluwer Academic: New York, 1999.

Langmuir 2011, 27(5), 1830–1836