A Comparison of Covalent Immobilization and Physical Adsorption of

Applied and Plasma Physics, School of Physics (A28), The University of Sydney, Sydney, NSW 2006, Australia. Langmuir , 0, (), ... E-mail: s.hirsh@phys...
16 downloads 4 Views 5MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

A Comparison of Covalent Immobilization and Physical Adsorption of a Cellulase Enzyme Mixture S. L. Hirsh,* M. M. M. Bilek, N. J. Nosworthy, A. Kondyurin, C. G. dos Remedios, and D. R. McKenzie Applied and Plasma Physics, School of Physics (A28), The University of Sydney, Sydney, NSW 2006, Australia Received May 18, 2010. Revised Manuscript Received July 20, 2010 This paper reports the first use of a linker-free covalent approach for immobilizing an enzyme mixture. Adsorption from a mixture is difficult to control due to varying kinetics of adsorption, variations in the degree of unfolding and competitive binding effects. We show that surface activation by plasma immersion ion implantation (PIII) produces a mildly hydrophilic surface that covalently couples to protein molecules and avoids these issues, allowing the attachment of a uniform monolayer from a cellulase enzyme mixture. Atomic force microscopy (AFM) showed that the surface layer of the physically adsorbed cellulase layer on the mildly hydrophobic surface (without PIII) consisted of aggregated enzymes that changed conformation with incubation time. The evolution observed is consistent with the existence of transient complexes previously postulated to explain the long time constants for competitive displacement effects in adsorption from enzyme mixtures. AFM indicated that the covalently coupled bound layer to the PIII-treated surface consisted of a stable monolayer without enzyme aggregates, and became a double layer at longer incubation times. Light scattering analysis showed no indication of aggregates in the solution at room temperature, which indicates that the surface without PIII-treatment induced enzyme aggregation. A model for the attachment process of a protein mixture that includes the adsorption kinetics for both surfaces is presented.

Introduction Invertase was the first enzyme mixture to be immobilized on a surface1 while retaining its biological function. Since then, the study of enzyme immobilization has generated interest in many different fields, including biomedical, biofouling, and bioreactors. The kinetics of the attachment process for a single enzyme have been studied and models have been generated for attachment, desorption, and spreading on the surface.2 Hydrophilic surfaces have been found to retain the function of the enzymes better than hydrophobic surfaces, which encourage enzyme spreading and unfolding.3 However, spreading has also been associated with increased irreversibility of the attachment,4 which is also desired. These reports suggest that the optimal attachment condition would be irreversible binding to a hydrophilic surface. The kinetics of the immobilization of an enzyme mixture has been found to be more complex than that for a single enzyme. Larger and more strongly binding enzymes can displace those previously adsorbed.5 It has been suggested that this process may proceed by the *Corresponding author. E-mail: [email protected]. (1) Nelson, J. M.; Griffin, E. G. J. Am. Chem. Soc. 1916, 38, 1109-1115. (2) (a) Ramsden, J. J. Protein Adsorption Kinetics. In Biopolymers at Interfaces; Malmsten, M., Ed.; Marcel Dekker, Inc: New York, 2003; pp 199-220. (b) Tilton, R. Mobility of Biomolecules at Interfaces. In Biopolymers at Interfaces; Malmsten, M., Ed.; Marcel Dekker, Inc: New York, 2003; Vol. 110, pp 221-257. (c) vanEijk, M. C. P.; Stuart, M. A. C. Langmuir 1997, 13, 5447-5450. (d) Van Tassel, P. R.; Guemouri, L.; Ramsden, J. J.; Tarjus, G.; Viot, P.; Talbot, J. J. Colloid Interface Sci. 1998, 207, 317-323. (3) Karlsson, M.; Ekeroth, J.; Elwing, H.; Carlsson, U. J. Biol. Chem. 2005, 280, 25558-25564. (4) (a) VanTassel, P. R.; Viot, P.; Tarjus, G. J. Chem. Phys. 1997, 106, 761-770; (b) Macritch., F. J. Colloid Interface Sci. 1972, 38, 484; (c) Feng, L.; Andrade, J. D. Biomaterials 1994, 15, 323-333. (5) (a) Vroman, L.; Adams, A. L. J. Colloid Interface Sci. 1986, 111, 391-402. (b) Norde, W. Adv Colloid Interface 1986, 25, 267-340. (c) Norde, W.; Anusiem, A. C. I. Colloids Surf. 1992, 66, 73-80. (6) (a) Dejardin, P.; Le, M. T. Langmuir 1995, 11, 4008-4012. (b) Huetz, P.; Ball, V.; Voegel, J. C.; Schaaf, P. Langmuir 1995, 11, 3145-3152. (c) Ball, V. Mechanism of Interfacial Exchange Phenomena for Proteins Adsorbed at Solid-Liquid Interfaces. In Biopolymers at Interfaces; Malmsten, M., Ed. Marcel Dekker: New York, 2003; Vol. 110, pp 295-320. (d) Schaaf, P.; Dejardin, P.; Schmitt, A. Langmuir 1987, 3, 1131-1135.

14380 DOI: 10.1021/la1019845

formation of a “transient complex” on the surface, though this structure has not been directly observed.6 There is no well-accepted model for the attachment kinetics of an enzyme mixture and little is understood regarding surface layer conformation changes over time.6c There are conflicting reports on the effectiveness of the covalent attachment of enzymes from a mixture. To date, the covalent immobilization of enzyme mixtures has been achieved using specific linker chemistry methods. Some studies suggest that covalent binding is the optimal immobilization method since it may prevent removal of enzyme from the surface during the reaction and improve reusuability.7 The covalent immobilization of cellulase and invertase has been shown to improve stability with respect to pH, temperature, reuse, and storage than free enzyme in solution.7,8 However, other studies have indicated that covalent attachment may result in a loss of activity compared to other immobilization techniques such as physical adsorption (physisorption).9 In addition, Tebeka et al.9c reported very different surface roughness and cross-sectional characteristics for the physisorbed and covalently immobilized cellulase. Cellulase is a mixture of enzymes that act synergistically to convert cellulose into glucose. Cellulase is of great interest because of its potential to convert cellulosic biomass into glucose for ethanol fuel production. Cellulase extracted from the fungus, Trichoderma reesei, is used by food, brewing, textile, and pulp and paper industries10 and is a mixture of at least two cellobiohydrolases, five endoglucanases, and β-glucosidase.11 These enzymes hydrolyze (7) Garcia, A.; Oh, S.; Engler, C. R. Biotechnol. Bioeng. 1989, 33, 321-326. (8) (a) Mao, X. P.; Guo, G. J.; Huang, J. F.; Du, Z. Y.; Huang, Z. S.; Ma, L.; Li, P.; Gu, L. Q. J. Chem. Technol. Biot. 2006, 81, 189-195. (b) Cadena, P. G.; Jeronimo, R. A. S.; Melo, J. M.; Silva, R. A.; Lima, J. L.; Pimentel, M. C. B. Bioresour. Technol. 2010, 101, 1595-1602. (9) (a) Li, C. Z.; Yoshimoto, M.; Fukunaga, K.; Nakao, K. Bioresour. Technol. 2007, 98, 1366-1372. (b) Saville, B. A.; Khavkine, M.; Seetharam, G.; Marandi, B.; Zuo, Y. L. Appl. Biochem. Biotechnol. 2004, 113-16, 251-259. (c) Tebeka, I. R. M.; Silva, A. G. L.; Petri, D. F. S. Langmuir 2009, 25, 1582-1587. (d) Kotwal, S. M.; Shankar, V. Biotechnol. Adv. 2009, 27, 311-322. (10) Dincer, A.; Telefoncu, A. J. Mol. Catal. B: Enzym 2007, 45, 10-14. (11) Zhang, Y. H. P.; Lynd, L. R. Biotechnol. Bioeng. 2004, 88, 797-824.

Published on Web 08/12/2010

Langmuir 2010, 26(17), 14380–14388

Hirsh et al.

Article

cellulose in a succession of endo-exo synergistic mechanisms.12 Endoglucanases make random cuts along the cellulose fibers to yield more chain ends for the cellobiohydrolases to produce cellobiose and oligosaccharides. In the final step, the β-glucosidase (cellobiase) converts cellobiose to glucose. Cellulase enzymes are modular proteins that typically contain at least three distinguishable domains:12c a catalytic region; a linker region; and a cellulose binding domain. The cellulose binding domain can attach to the insoluble cellulose and the linker region enables the cellulose to be held close to the catalytic region so that it can be acted upon.13 This paper reports the first use of the linker-free covalent attachment of an enzyme mixture and achieves irreversible attachment on a hydrophilic surface. Covalent immobilization is achieved with plasma immersion ion implantation (PIII), which has recently been reviewed.14 However, this method has only been used to covalently immobilize single enzymes, including horseradish peroxidase,15 catalase,16 and soybean peroxidase.17 This method of immobilization has been shown to produce more densely packed layers of enzyme than are adsorbed on the untreated surfaces17,18 and the layers on the treated surfaces appeared to be monolayers.18 PIII is a promising technique for many applications because it improves bioactivity and activity retention after repeated washing cycles and after freeze-drying.19 In this paper, we compare the structural properties and the behavior over time of adsorbed enzyme layers from a cellulase enzyme mixture when covalently attached versus physically adsorbed on PIII-treated and untreated polystyrene, respectively. Polystyrene was chosen because it is an inexpensive carbon-based polymer that is used in many applications involving protein adsorption.20 AFM is used to provide direct images of the structures present on the surfaces and their evolution as a function of incubation time.

Materials and Methods 1. PIII-Treated Polystyrene and Cellulase Immobilization. Polystyrene films, 0.25 mm thick (Goodfellow Cambridge Ltd.), were placed into a plasma chamber18 on a conductive sample holder with a tungsten mesh placed approximately 5 cm in front and connected electrically to the sample holder to prevent arcing. Nitrogen was introduced into the plasma chamber at a pressure of 0.267 Pa (2  10-3 Torr). A radio frequency plasma was created in the chamber using an inductively coupled antenna operating at 13.56 MHz and was used as the source of ions for implantation. Ions were accelerated to the surface by the application of 20 kV, 20 microsecond, bias pulses to the sample holder at a frequency of 50 Hz. The polystyrene films were PIII-treated for 400 s. After the treatment, the surface appeared uniformly dark brown in color. (12) (a) Stahlberg, J.; Johansson, G.; Pettersson, G. Biochim. Biophys. Acta 1993, 1157, 107-113. (b) Teeri, T.; Salovuori, I.; Knowles, J. Bio-Technol. 1983, 1, 696-699. (c) Goyal, A.; Ghosh, B.; Eveleigh, D. Bioresour. Technol. 1991, 36, 37-50. (13) Zhao, X.; Rignall, T. R.; McCabe, C.; Adney, W. S.; Himmel, M. E. Chem. Phys. Lett. 2008, 460, 284-288. (14) Bilek, M. M. M.; McKenzie, D. Biophys. Rev. 2010, 2, 55-65. (15) Gan, B. K.; Nosworthy, N. J.; McKenzie, D. R.; dos Remedios, C. G.; Bilek, M. M. M. J. Biomed. Mater. Res. A 2008, 85A, 605-610. (16) Nosworthy, N. J.; Ho, J. P. Y.; Kondyurin, A.; McKenzie, D. R.; Bilek, M. M. M. Acta Biomater 2007, 3, 695-704. (17) MacDonald, C.; Morrow, R.; Weiss, A. S.; Bilek, M. M. M. J. R. Soc. Interface 2008, 5, 663-669. (18) Gan, B. K.; Kondyurin, A.; Bilek, M. M. M. Langmuir 2007, 23, 2741-2746. (19) Nosworthy, N. J.; McKenzie, D. R.; Bilek, M. M. Biomacromolecules 2009, 10, 2577-2583. (20) (a) Norde, W.; Lyklema, J. J. Colloid Interface Sci. 1978, 66 (2), 257-265. (b) Gessner, A.; Paulke, B. R.; Muller, R. H. Electrophoresis 2000, 21, 2438-2442. (c) Engel, M. F. M.; Visser, A. J. W. G.; van Mierlo, C. P. M. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 11316-11321.

Langmuir 2010, 26(17), 14380–14388

Pieces of PIII-treated and untreated polystyrene sheet 0.25 mm thick of dimensions 20  30 mm were placed in a 15 mL of 10 mM sodium phosphate buffer solution (PO4) with 150 μL of cellulase solution (Sigma No. C2730) and placed on a rocker for various incubation times. The surfaces were washed five times in phosphate buffer (PO4), pH 7, for 10 min cycles to remove loosely bound protein. Spin coated polystyrene (PS) films on silicon were used for atomic force microscopy analysis. These were more suitable for AFM imaging than the PS sheet because of their superior smoothness and flatness. A solution of PS (Austrex 400, Polystyrene Australia Pty. Ltd.) was dissolved in toluene (Sigma No. 34866, purity >99.9%) at 1 g L-1 was applied to a SCS G3P-8 Spincoater at 2000 rpm. These surfaces (silicon with ∼100 nm layer of polystyrene) were treated using the same method as described above for the polystyrene sheet. 2. Cellulase Solution Analysis. Light scattering analysis was carried out on the cellulase solution (1/100 dilution) in pH 7 phosphate buffer at the National Measurement Institute (NMI), Sydney Australia with a Malvern Zetasizer Nano series Nano-ZS machine. The machine uses backscattered laser light at 173° scattering angle to determine the size of the particles in the solution. The analysis was conducted with temperature ramping at 20 and 37 °C and from 45 to 66 °C in 3 °C increments. Five scans, each of approximately 2 min, were conducted at each temperature. The Malvern software was used for the analysis. The viscosity of the solution was approximated with water and the refractive index of the cellulase enzymes was estimated at 1.45. 3. Surface Layer Analysis. Atomic force microscopy (AFM) was used to determine the coverage of the cellulase on the surface. AFM images (topographical and phase) were collected of spin coated polystyrene, PIII-treated polystyrene, and cellulase attached to PIII-treated and untreated polystyrene surfaces at room temperature using a PicoSPM instrument (Australian Centre for Microscopy and Microanalysis, The University of Sydney) in tapping mode. Cellulase incubation times of 1 min, 60 min, and 24 h cellulase incubation times were used. All samples were washed five times in PO4 (pH 7) for 10 min cycles, rinsed in Milli-Q water to remove salt and then dried in Petri dishes stored at room temperature prior to AFM measurement in air. Scans were taken over a range of areas from 0.5  0.5 μm, 1  1 μm, and 2  2 μm at three different points on the sample. The images were analyzed and the root-mean-square roughness characteristics were studied using WSxM software (version 3, Nanotech Electronica, S.L., Spain).21 The rms for the untreated polystyrene and PIII-treated surfaces with no enzyme was found to be 0.2 ( 0.05 nm. The roughness features on the surfaces without enzyme were found to be on a longer length scale than those found with the enzyme. To isolate the rms changes due to changes in the enzyme layer, the rms values were calculated after applying a parabolic flattening filter and the average rms values of the 0.5  0.5 images for each sample are reported. The 60 min incubation samples were stored for a few weeks prior to measurement; however, no change is expected during storage since the samples were dry. For each sample, all the images were found to be consistent, which suggests uniformity across the sample. ATR-FTIR spectra were recorded on cellulase adsorbed on treated and untreated polystyrene sheets using a Digilab FTS 7000 FTIR spectrometer. The optical density of spectral peaks associated with the characteristic amide I (1650 cm-1), amide II (1540 cm-1), and amide A (3300 cm-1) bond vibrations for (21) Horcas, I. et al. Rev. Sci. Instrum. 2007, 78, 013705.

DOI: 10.1021/la1019845

14381

Article

Hirsh et al.

Figure 1. AFM topography images of cellulase immobilized on untreated polystyrene at incubation times of 1 min (left), 60 min (middle), and 24 h (right). The 1 min and 24 h images were taken on the same day, whereas the 60 min image was taken at a later date with a different tip. These images are consistent with transient complexes that facilitate enzyme exchange postulated in Ball et al.6c The left image may be showing the attachment of enzymes which form the transient complex, the middle image may be showing the complex after turning to expose previously adsorbed enzymes to solution for desorption and the right image may be showing the structure of the surface layer after their desorption.

proteins were used to quantify the relative amounts of cellulase attached to polystyrene sheets. Using the Bouger-Lambert-Beer law, the ratio of the absorbance intensity of the protein amide peaks to the intensity of the PS absorption at 1452 cm-1 is proportional to the concentration of the amide functional groups. Thus, this ratio can be used to analyze the relative amounts of cellulase attached to the surfaces. The relative amount of cellulase attached to PIII-treated and untreated polystyrene surfaces was determined as a function of cellulase incubation time (1, 5, 10, 20, and 60 min; 8, 21, 24, 30, and 48 h). The ATR-FTIR spectra were recorded before and after sodium dodecyl sulfate (SDS) washing to determine the proportion protein on the surface that was covalently bound. The SDS wash was conducted with 2% SDS at 70 °C for 1 h. The use of SDS washing as a test of covalent binding is discussed by Bilek and McKenzie.14 SDS is a detergent that is used to unfold proteins.22 SDS interferes with the physical forces that are responsible for the physisorption of proteins onto surfaces but does not attack covalent bonds leaving the protein’s primary structure intact. SDS washing has been used as a method to test whether biological molecules are covalently attached to surfaces23-25 and to detect covalently bound drug-protein adducts.26 In some situations, steric hindrance may

prevent the SDS from accessing all of the sites where physical forces bind the protein to the surface. An example of such a situation may be where there is a thick coverage of strongly denatured and aggregated protein completely blocking access to the interface at the surface. Since our PIII-treated surfaces are relatively hydrophilic compared to untreated polystyrene surface control from which our SDS wash successfully removed all of the protein, it is unlikely that steric hindrance could be responsible for the SDS-resistant binding observed on the PIII-treated surfaces.We therefore propose that the high proportion of enzyme still adsorbed after SDS cleaning implies that the PIII-treated surfaces have sites capable of covalently binding protein. Such a covalent coupling capability has previously been observed on a range of PIII-treated polymers as well as for plasma polymerized layers with similar surface chemistry and structure.14,16,17,27 Likely mechanisms for this linker-free covalent coupling are discussed in the literature.14,28 Kruss DS10 equipment was used to measure the water contact angle using the sessile drop technique as a function of time from immediately after PIII treatment at 24 h intervals for 1 week. The biological activity of the immobilized cellulase enzymes on the PIII-treated and untreated surfaces were measured to ensure that the enzymes remain functional after attachment. The enzymes

(22) Laemmli, U. K. Nature 1970, 227 (5259), 680. (23) Shlyakhtenko, L. S.; Gall, A. A.; Weimer, J. J.; Hawn, D. D.; Lyubchenko, Y. L. Biophys. J. 1999, 77, 568-576. (24) Vandenberg, E.; Elwing, H.; Askendal, A.; Lundstrom, I. J. Colloid Interface Sci. 1991, 143, 327-335. (25) Hodneland, C. D.; Lee, Y. S.; Min, D. H.; Mrksich, M. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 5048-5052. (26) Zhou, S. F. J. Chromatogr. B 2003, 797, 63-90.

(27) (a) Kondyurin, A.; Nosworthy, N. J.; Bilek, M. M. M. Acta Biomater 2008, 4, 1218-1225. (b) Yin, Y. B.; Nosworthy, N. J.; Gong, B.; Bax, D.; Kondyurin, A.; McKenzie, D. R.; Bilek, M. M. M. Plasma Process Polym. 2009, 6, 68-75. (c) Kondyurin, A.; Naseri, P.; Fisher, K.; McKenzie, D. R.; Bilek, M. M. M. Polym. Degrad. Stab. 2009, 94, 638-646. (28) Yin, Y. B.; Bilek, M. M. M.; McKenzie, D. R.; Nosworthy, N. J.; Kondyurin, A.; Youssef, H.; Byrom, M. J.; Yang, W. R. Surf. Coat. Technol. 2009, 203, 1310-1316.

14382 DOI: 10.1021/la1019845

Langmuir 2010, 26(17), 14380–14388

Hirsh et al.

Article

Figure 2. AFM topography images of cellulase immobilized on PIII-treated polystyrene at incubation times of 1 min, 60 min, and 24 h. The 1 min and 24 h images were taken on the same day, whereas the 60 min image was taken at a later date with a different tip. The 60 min image shows some drift.

were immobilized (1/100 dilution) at room temperature, pH 7 for 24 h on polystyrene surfaces (7 cm 10 cm). The surfaces were washed in buffer five times for 10 min intervals to remove any loosely bound protein and then were placed in a 50 mL tube with 0.15 g of Carboxymethylcellulose (CMC) (Sigma No. 419273) dissolved in 30 mL of 0.05 M sodium acetate buffer (pH 5). The mixture was reacted in a water bath at 37 °C. Aliquots of 0.5 mL were measured for glucose at 24, 48, and 96 h. Glucose is a product of the hydrolysis of cellulose and was measured with a glucose oxidase assay kit (Sigma No. GAGO20). The kit is an enzymatic assay involving a three-step reaction. First, the glucose is oxidized to gluconic acid and hydrogen peroxide. The hydrogen peroxide then reacts with o-dianisidine in the presence of peroxidase. The oxidized o-dianisidine then reacts with 12N sulfuric acid to form a pink colored product. The optical density of this product was determined at 540 nm (Beckman DU530 spectrophotometer).

Results 1. AFM-Room Temperature Cellulase Incubation. Selected atomic force microscope images, corresponding rms values, and height profiles of untreated polystyrene surfaces incubated with cellulase for a range of incubation times are shown in Figure 1. Analysis of these images reveals that the enzyme surface layer changes with incubation time. At an incubation of 1 min the layer contains large features, approximately 100 nm in size, suggesting enzyme aggregates and smaller features of approximately 15 nm in size, consistent with individual enzyme molecules. With further incubation time the surface roughness and the height profile increases. For example, at 60 min, the height profile is typically between 2 and 9 nm and has an rms value of 1.5 nm. The neighboring enzyme Langmuir 2010, 26(17), 14380–14388

aggregates appear to be merging to form a network structure with voids. With a further incubation time of 24 h the rms value is 1.7 nm and the height profile ranges from 4 to 11 nm. A rod-like structure appears to develop within the network with a feature size of 20 nm x 100 nm. Selected AFM images with corresponding rms values and height profiles of the enzyme layers on PIII-treated surfaces at various incubation times are shown in Figure 2. These images show a very different structure from the enzymes on the untreated surface. The enzymes maintain a height profile ranging between 1 and 3 nm and the rms values for each incubation time are significantly lower than those for the untreated polystyrene surface, indicating a more uniform enzyme layer. The features in the image are 10-15 nm in size and are interpreted as individual enzyme molecules. The rms and surface height profile for 1 and 60 min incubation times remain similar; however, the rms at 24 h is twice that of the shorter incubation times and the height profile in some locations is twice the height profile of the shorter incubation times. 2. Light Scattering Measurements of Cellulase Solution. Figure 3A shows the light scattering results for the cellulase solution conducted at room temperature (20 °C). The intensity distribution reproducibly showed a single peak with an average diameter value of approximately 8 ( 0.14 nm. The peak covered the range from 3 to 15 nm. This wide distribution is a result of the mixture of enzymes in the cellulase solution and the size distribution is consistent with the sizes of the enzymes obtained from the crystalline structures.29 At 37 and 45 °C, no change was observed in the intensity distribution. However, Figure 3B shows that at 48 °C, the solution becomes unstable and begins to form aggregates which increase in size with time and temperature. At 51 °C, the intensity DOI: 10.1021/la1019845

14383

Article

Hirsh et al.

Figure 3. A-C show the distribution of light scattering intensity as a function of particle size. Each plot contains 5 scans, repeated at approxi-

mately 2 min intervals. (A) At room temperature a single reproducible peak is dominant. (B) At 48 °C the solution becomes unstable, forming aggregates of increasing size. (C) At 60 °C the solution contains large aggregates. (D) The distribution of particle sizes by volume at 60 °C is shown. The large aggregates are still in the minority by volume. The peak shift observed for the smaller enzymes indicates that they are also forming aggregates.

Figure 4. Plot of the average of the particle size weighted by intensity illustrates the aggregation of cellulase with temperature and time. The vertical lines indicate points at which measurements commenced at a new temperature after stabilization at that temperature.

distribution continues to change, indicating that the solution is unstable and the aggregates are continuing to grow. At 54 °C, the rate of growth of aggregates begins to decrease and the average size of the enzyme aggregates is approximately 22.5 ( 1.8 nm. From 54 to 66 °C, there are two distinct peaks representing the enzymes originally present at room temperature together with the enzyme aggregates. At 60 °C, the average size of the aggregates is 25 ( 2.0 nm (Figure 3C). Figure 3D shows the percentage by volume of enzyme aggregates at 60 °C. While the aggregates showed a consistent increase with time and temperature, up to this point, the percentage of enzyme aggregates is only 13% ( 1.8% by volume. Figure 4 is a plot of the temperature ramping and the average particle size (29) (a) (UniProtKB), P. K. B. CBHI. http://www.uniprot.org/uniprot/P62694. (b) (UniProtKB), P. K. CBHII. http://www.uniprot.org/uniprot/P07987. (c) (UniProtKB), P. K. EG1. http://www.uniprot.org/uniprot/P07981. (d) (UniProtKB), P. K. EG2. http://www.uniprot.org/uniprot/P07982. (e) (UniProtKB), P. K. B. EG3. http://www.uniprot.org/uniprot/O00095. (f) (UniProtKB), P. K. EG4. http:// www.uniprot.org/uniprot/O14405. (g) (UniProtKB), P. K. EG5. http://www. uniprot.org/uniprot/P43317. (h) (UniProtKB), P. K. B. Beta-glucosidase. http:// www.uniprot.org/uniprot/O93785.

14384 DOI: 10.1021/la1019845

determined by the intensity distribution. It illustrates that the aggregation process begins at 48 °C, accelerates at 51 °C, and approaches an asymptote in average particle size as determined by the intensity profile at approximately 25 nm. 3. AFM: Heated Incubation of Cellulase. Figure 5 shows AFM images of cellulase immobilized to the PIII-treated surface during incubation at 60 °C for 1 h (right) compared to PIIItreated surfaces with cellulase immobilized at room temperature for 20 h (left). The rms roughness value for the room temperature sample is consistent with a single layer of enzymes, while the rms value for the heated sample is higher. AFM images of cellulase immobilized to the PIII-treated surface during incubation at 60 °C for 1 h revealed larger enzyme structures than the treated enzyme layer immobilized at room temperature for 20 h. The size of the enzymes immobilized at room temperature can be estimated at 15-20 nm from the AFM topography and phase images (Figure 5). The phase images show changes in mechanical response as a function of position and allow a better measurement of the size because the resolution is not as limited by tip radius. The size of the enzymes immobilized at 60 °C appears to be 30-40 nm, approximately twice that of the enzymes immobilized at room temperature. 4. ATR-FTIR and Contact Angle Analysis. The ATRFTIR results in Figure 6A indicate that the amount of protein immobilized to both surfaces (PIII-treated and untreated) was approximately the same for incubation times up to 1 h. However, with further incubation, the cellulase on the PIII-treated surface gradually increased linearly with incubation time. The cellulase on the untreated polystyrene surface saturated and remained steady with incubation time. After SDS washing, no protein was detected on the untreated polystyrene surface, indicating that none of the protein was bound covalently. Figure 6B shows the measured amount of protein before and after SDS washing on the PIII-treated surface. It can be observed that the amount of protein remaining after SDS washing did not increase with incubation time. Up to incubation times of 8 h, almost all of the protein was bound covalently to the surface and for further incubation times increasing amounts of protein were physisorbed. The results show that all of the covalent binding Langmuir 2010, 26(17), 14380–14388

Hirsh et al.

Article

Figure 5. (A) AFM topography image of cellulase immobilized on PIII-treated polystyrene at room temperature; rms roughness = 0.4 nm ( 0.05 nm. (B) AFM topography image of cellulase immobilized on PIII-treated polystyrene while incubated at 60 °C; rms roughness = 0.6 nm ( 0.05 nm. (C) AFM phase image of the same region as in part A. (D) AFM phase image of the same region as in part B.

occurs within the first hour of incubation. Further investigation of the incubation time dependence of the binding showed that the amount found to be covalently bound was reached within the first minute of incubation (Figure 7). Similar amounts of protein are observed on both surfaces after the first minute of incubation. Some of the small variations observed may result from simultaneous adsorption and desorption processes. Figure 8A compares the contact angle of the untreated polystyrene surface with the contact angle of the PIII-treated surface as a function of time after treatment while stored in laboratory ambient conditions. The untreated polystyrene surface is relatively hydrophobic with a contact angle of 83.6 ( 1.2 deg. After PIII-treatment, the surface becomes hydrophilic with a contact angle of 49.1 ( 4.2 deg. Figure 8B shows the surface energy, and its polar and dispersive components, calculated from the data in Figure 8A. An increase in surface hydrophilicity of the PIII-treated surface compared to its untreated counterpart has been previously reported.14 It is believed to be due to the oxidation of the highly reactive treated surface on exposure to atmosphere.14,27c The hydrophilic nature of the treated surface is stable with time. Since hydrophobic surfaces are known to denature proteins by exposing their hydrophobic core,3 the PIII-treated surface is more likely than the untreated surface to maintain protein conformation during attachment. Langmuir 2010, 26(17), 14380–14388

5. Biological Activity Tests. The biological activity (Figure 9) of the immobilized enzymes on PIII-treated and untreated polystyrene surfaces was determined using CMC as the substrate. The PIII-treated surface consistently showed higher activity throughout the reaction. At 96 h, the PIII-treated surface had produced 0.97 mg of glucose, while the untreated surface had only produced 0.63 mg (65% of the PIII-treated surface value).

Discussion The surface attachment process of a single enzyme solution can be thought of as a series of five steps: (1) transport to the surface; (2) attachment to the surface; (3) rearrangement and spreading on the surface; (4) desorption from the surface; (5) transport away from the surface.5b Solutions to the diffusion equation (eq 1) provide the protein flux to the surface.30 !1=3 6q - 1 - 1=3 Jp ¼ ½Γð4=3Þ 9 Dp Cp ð1Þ b2 wlDp Here, Jp is the protein flux to the surface (kg m-2 s-1), Dp is the diffusion coefficient of the protein (m2 s-1), Cp is the protein mass (30) Kallay, N. Interfacial Dynamics; Marcel Dekker, Inc.: New York, 2000.

DOI: 10.1021/la1019845

14385

Article

Hirsh et al.

Figure 7. Relative amount of protein as determined by the normalized amide peak intensities on untreated and PIII-treated surfaces as a function of cellulase incubation time (from 1 min to 1 h) determined by ATR-FTIR spectroscopy. The PS peak at 1452 cm-1 was used for normalization. Figure 6. (A) Relative amount of protein determined by normalized amide peak intensity on untreated and PIII-treated surfaces as a function of cellulase incubation time (from1 to 48 h) as determined by ATR-FTIR spectroscopy. (B) Relative amount of protein determined by normalized amide peak intensity on the PIII-treated surface as a function of incubation time before and after SDS washing as determined by ATR-FTIR spectroscopy.

concentration, l (m) is the distance of rectangular flow channel, b (m) is the thickness of the flow channel, w (m) is the width of the flow channel, and q is the volume flow rate of solution (m3 s-1). The rate of surface accumulation will be governed by the enzyme’s rate constants for adsorption and desorption at the surface31 dðΘACs Þ ¼ ka Co ð1 - ΘÞA - kd ΘACs dt

ð2Þ

where d(Θ)ACs)/dt is the rate of surface accumulation, ka is the adsorption rate constant, C0 is the protein bulk concentration, Φ is the fractional surface coverage, A is the total surface area, kd is the desorption rate constant, and Cs is the protein surface concentration. As a result of energy minimization, the enzymes will also spread and rearrange themselves on the surface, which may result in attached enzymes being displaced from the surface. The desorption rate constant kd includes contributions from spontaneous desorption and desorption triggered by spreading of neighboring molecules. The residence time of the adsorbed molecule is the inverse of kd. The above model, however, does not take into account the competition for binding sites between different proteins in the case of a mixture of proteins adsorbing to a surface. In an enzyme mixture, such as the cellulase mixture (Table 1), the protein flux to the surface of a given molecule is related to its concentration in the mixture and to its diffusion coefficient.32 The diffusion coefficients depend on the molecular weight and molecular conformation. The composition of the adsorbed layer may evolve with time as proteins that arrive later and have a higher affinity for the surface displace earlier bound proteins.5c,33 When a molecule changes conformation on the surface it can also spread, displacing neighboring molecules.4a These processes will impact on the (31) Toth, J., Adsorption Theory, Modeling, Analysis; Marcel Dekker, Inc: New York, 2002; Vol. 107, p 871. (32) Latour, R. Biomaterials: Protein-surface Interactions. In Encyclopedia of Biomaterials and Biomedical Engineering; Taylor & Francis: Philadelphia, PA, 2005. (33) (a) Norde, W.; Favier, J. P. Colloids Surf. 1992, 64, 87-93. (b) Elgersma, A. V.; Zsom, R. L. J.; Lyklema, J.; Norde, W. J. Colloid Interface Sci. 1992, 152, 410-428.

14386 DOI: 10.1021/la1019845

Figure 8. (A) Contact angles of untreated polystyrene and PIIItreated polystyrene as a function of time after treatment as measured using the sensile drop method. (B) Total surface energy (T), the polar component of the surface energy (P), and the dispersive component of the surface energy (D) of the untreated polystyrene and the PIIItreated polystyrene as a function of time after treatment.

desorption rate coefficients (eq 2) for the competing species in the mixture and therefore impact on their residence times on the surface. The final surface composition and conformation strongly depend on the surface energy because energy minimization drives competition and spreading. When covalent binding occurs, desorption is restricted by the strength of the binding, which reduces the possibility of competitive effects. In addition, the number of conformations possible from spreading effects may be reduced by fixing a single residue in place with a covalent bond. The differences between the adsorption behavior of the treated and untreated surfaces can be explained by differences in the attachment mechanisms as well as differences in the surface energy. The processes involved are illustrated in a schematic, Figure 10. For the untreated surface, within the first minute of incubation, the final adsorbed mass has essentially been reached (Figure 6A). The AFM data (Figure 1) at 1 min incubation shows the presence of large aggregates, 100 nm in size, on the surface. These aggregates must be Langmuir 2010, 26(17), 14380–14388

Hirsh et al.

Article

Table 1. Composition of the T. reesei Cellulase Mixture Showing Two Cellobiohydrolases, Five Endoglucanases, and β-Glucosidase with Their Respective Molecular Weights and Respective Concentrations enzyme name

mass (Da)

concentration (%)

CBH I CBHII EG I, II, III, IV, V β-glucosidase

54.129a 49.729b 24.4-48.229c-g 52.229h

6012c 2012c 19 134

Figure 9. Biological activity of cellulase immobilized to PIII-treated and untreated polystyrene surfaces was determined by measuring the glucose produced after the hydrolysis of carboxymethylcellulose (CMC).

Figure 10. Schematic model for protein adsorption on untreated (left) and PIII-treated (right) polystyrene. After 1 min incubation (top) the untreated surface has adsorbed protein that has unfolded and formed into transient aggregate structures. Newly arrived enzymes are being incorporated into the aggregates. After 60 min incubation, the transient aggregate structure rotates and exposes some of the earlier bound enzymes to solution. After 24 h structure the aggregate structure has evolved further into rod-like structures. On the hydrophilic PIII-treated surface (right) a covalently bound monolayer is adsorbed within the first minute and retains enzyme conformation (top). After 60 min incubation, the monolayer is unchanged (middle). After 24 h incubation, partial coverage by a second layer is observed.

induced by the surface since the solution at room temperature shows no aggregates (Figure 3A). This indicates that the molecules adsorbing are being destabilized by their interaction with the surface and unfold in response to the adverse (hydrophobic) surface energy (Figure 8), which encourages aggregation on the surface of adsorbed molecules. These enzyme aggregates may correspond to the “transient complexes” that have been described in the literature.6c,d In this postulated model for the competitive effects, the formation of transient complexes is the rate limiting step. The binding on the untreated surface is not covalent, which allows for desorption, (34) Chauve, M.; Mathis, H.; Huc, D.; Casanave, D.; Monot, F.; Ferreira, N. L. Comparative kinetic analysis of two fungal beta-glucosidases. Biotechnol Biofuels 2010, 3, 3.

Langmuir 2010, 26(17), 14380–14388

competition and spreading effects that will then modify the enzyme layer over time, as observed in Figure 1. We have observed that with a 60 min incubation time, the aggregate enzyme structure evolves on the surface to form a network structure. This may correspond to the postulated “turning” of the transient complexes that exposes previously adsorbed enzymes in the complex to the solution, allowing them to desorb.6c,d Our observed time-scale, on the order of 1 h, for the evolution of the structure is consistent with the time-scale previously observed in the literature for competitive enzyme replacement.6b After a 24 h incubation time, the enzyme network after desorption, evolves to become rod-like, with a greater height profile and roughness (Figure 1). On the PIII-treated surface, the adsorbed proteins bind to form stable monolayer coverage within the first minute (Figure 7). An SDS wash after an incubation of 1 h showed this monolayer to be covalently attached (Figure 6B). Since the binding is covalent, competition and desorption effects are no longer significant and the surface attachment can be modeled in the transport limited regime, where the initial protein flux to the surface largely reflects the composition of the bound layer and this layer does not change with time. At room temperature, no aggregates are present in the solution (Figure 3A) and no aggregates are observed on the PIIItreated surface (Figure 2). The enzymes on the PIII-treated surface are approximately 10-15 nm in size, which is consistent with the dimensions of their structures as determined by X-ray crystallography.29 Unfolding and aggregation is not encouraged because the surface energy conditions favor the native conformation (Figure 8). At long incubation times, a second layer coverage forms (illustrated in Figure 10). This can be observed by the increased surface roughness and height profile characteristics of the 24 h incubation sample in Figure 2. This interpretation is further supported by the linear increase in the amount of protein attached to the surface with incubation time from 1 to 48 h (Figure 6A). The second layer is not covalently attached (Figure 6B) and is believed to form as a result of protein-protein interactions. Slow forming second enzyme layers that grow linearly with time have been observed in the literature.3 To investigate the transport-limited nature of the binding on the treated surface, we analyzed the results of the adsorption in a solution at 60 °C. The flux to the surface now consists of aggregate molecules of an average size of 25 ( 2.0 nm as well as unaggregated molecules. The aggregates comprise approximately 13% of the volume of enzyme in solution. The size distribution on the surface reflects this change in transport flux because aggregates are now observed, which were not present at room temperature (Figure 5). Taking into account the small drift in the images by measuring normal to the drift direction in the AFM phase image, the average size of the aggregates is estimated to be 23 ( 5 nm, which agrees with the size of the aggregates measured in solution. The percentage of aggregates (greater than 15 nm) observed on the surface heated during incubation is estimated to be 60%. Since the aggregates have a larger volume the percentage by volume on the surface consisting of aggregates will be greater than 60%. The percentage of aggregates on the surface is much greater than would be expected from the initial flux of particles to the surface, especially given that the diffusion coefficient for larger particles is smaller than that for small particles. The apparent enhancement of larger particles on the PIIItreated surface incubated at 60 °C could be explained by the following argument based on the mechanism of covalent binding. A suggested mechanism for the covalent binding on PIII-treated surfaces is a reaction between radical groups created in the surface by the treatment and protein side chains.14 The collision cascades of the ion implantation process are known to create free radicals DOI: 10.1021/la1019845

14387

Article

to depths of hundreds of nanometers. The time to form a covalent bond would depend on the migration time of an unpaired electron in the modified zone to the surface. If the time constant for this process is short compared to the characteristic time for transport to the surface and compared to the residence time of physisorbed protein, the enzymes in the initial flux to the surface would occupy the available sites for covalent binding and this would mean that few aggregates would be present on the surface—certainly no greater proportion than that in solution. Conversely, if the time constant for covalent binding is long then the aggregates will have an increased probability of forming a covalent bond because of their increased surface area and hence, increased likelihood of a reaction with an available radical group on the surface prior to desorption. The dominance of aggregates on the PIII-treated surface incubated at 60 °C may therefore reflect the kinetics of the covalent binding reaction. The covalent binding time constant must be longer than the residence time of the small mostly nonaggregated enzymes. Although they diffuse to the surface more slowly, the larger aggregated enzymes would be expected to have a longer residence time on the surface and this coupled with their larger interaction area would significantly increase the probability of covalent bond formation.

Conclusions The structure of the cellulase layer when physisorbed on untreated polystyrene is very different to that when covalently immobilized on PIII-treated polystyrene, which is a mildly hydrophilic surface activated to covalently couple proteins directly. On the untreated polystyrene, the structure of the physisorbed layer strongly depends on the competitive binding properties of the enzymes in the cellulase mixture and the spreading induced by their interactions with the surface upon adsorption. Cellulase coverage was found to saturate within the first minute of incubation and contained surface-induced enzyme aggregates. The untreated surface is mildly hydrophobic and causes enzyme unfolding and aggregation. The enzyme aggregate surface layer was found to change structure with further incubation time. These changes are a likely result of the competition between different enzymes in the mixture for physisorption sites and by the spreading of the adsorbed enzymes. Observed enzyme aggregates and conformational changes in the enzyme layer as a function of incubation time in the AFM images may be evidence to support the postulated existence of transient complexes, which facilitate the exchange of enzymes adsorbed from a mixture. On the hydrophilic PIII-treated surface, a covalently attached monolayer with no enzyme aggregates was observed. This

14388 DOI: 10.1021/la1019845

Hirsh et al.

monolayer was found to have saturated within the first minute of incubation and does not change with incubation time. A second physisorbed layer slowly forms at long incubation times. At room temperature, the composition and structure of the enzyme layer reflects the initial enzyme flux to the surface. The enzymes in the initial flux are covalently attached to the surface, their conformation does not appear to change and the layer structure is constant with further incubation time. At 60 °C incubation temperature, the proportion of enzyme aggregates on the surface is higher than their proportion in solution. This indicates that the characteristic time for formation of a covalent bond is long compared to residence times of the unaggregated molecules. The enzyme aggregates formed in solution are preferentially covalently attached to the surface because of their increased surface area and increased residence time, both of which increase the probability of reaction with an unpaired electron on the surface, a likely mechanism for covalent attachment by PIII treatment. The structural evolution of the enzyme layers may have implications for the observed activity of the covalently immobilized (PIII-treated surface) and physisorbed (untreated surface) cellulase enzyme layers. Activity on the PIII-treated surface was found to be higher than the activity of the untreated surface. The activity on the untreated surface may be lower since enzyme aggregates are generally not as active as the same number of unaggregated enzymes. Aggregates affect the accessibility to cellulose substrate and also adversely affect the conformation. Since structural conformation of the attached cellulase is preserved on the PIII-treated surface, the enzyme functionality is likely to be preserved. However, the close proximity of the enzymes may affect the accessibility of the molecules to cellulose substrate. In addition, the ratio of the each of the enzymes in the cellulase mixture may be different on both surfaces, which could have large implications for their synergistic activity. It may be possible to control the relative proportions of each enzyme on the PIII-treated surface by adjusting their concentration in solution, dilution, PIII-treatment time, and surface energy. Acknowledgment. The authors of this paper acknowledge the Australian Research Council for financial support and the National Measurement Institute of Australia and Dr. Asa Jamting for assistance with the light scattering measurements. Software provided by WSxM software was used for AFM analysis. We also acknowledge the Australian Centre for Microscopy and Microanalysis at the University of Sydney and its personnel for assistance with the AFM measurements.

Langmuir 2010, 26(17), 14380–14388