Polymeric Brushes as Functional Templates for Immobilizing

Dec 13, 2007 - Polymeric Brushes as Functional Templates for Immobilizing Ribonuclease A: Study of Binding Kinetics and Activity ... We employed poly(...
0 downloads 0 Views 180KB Size
Download PDF
Langmuir 2008, 24, 913-920

913

Polymeric Brushes as Functional Templates for Immobilizing Ribonuclease A: Study of Binding Kinetics and Activity Sean P. Cullen,† Xiaosong Liu,‡ Ian C. Mandel,§ Franz J. Himpsel,‡ and Padma Gopalan*,§ Chemical and Biological Engineering, Department of Physics, and Materials Science and Engineering, UniVersity of WisconsinsMadison, Madison, Wisconsin 53706 ReceiVed August 13, 2007. In Final Form: October 17, 2007 The ability to immobilize proteins with high binding capacities on surfaces while maintaining their activity is critical for protein microarrays and other biotechnological applications. We employed poly(acrylic acid) (PAA) brushes as templates to immobilize ribonuclease A (RNase A), which is commonly used to remove RNA from plasmid DNA preparations. The brushes are grown by surface-anchored atom-transfer radical polymerization (ATRP) initiators. RNase A was immobilized by both covalent esterification and a high binding capacity metal-ion complexation method to PAA brushes. The polymer brushes immobilized 30 times more enzyme compared to self-assembled monolayers. As the thickness of the brush increases, the surface density of the RNase A increases monotonically. The immobilization was investigated by ellipsometry, X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis (TGA), and near-edge X-ray absorption fine structure spectroscopy (NEXAFS). The activity of the immobilized RNase A was determined using UV absorbance. As much as 11.0 µg/cm2 of RNase A was bound to PAA brushes by metal-ion complexation compared to 5.8 µg/cm2 by covalent immobilization which is 30 and 16 times the estimated mass bound in a monolayer. The calculated diffusion coefficient D was 0.63 × 10-14 cm2/s for metal-ion complexation and 0.71 × 10-14 cm2/s for covalent immobilization. Similar values of D indicate that the binding kinetics is similar, but the thermodynamic equilibrium coverage varies with the binding chemistry. Immobilization kinetics and thermodynamics were characterized by ellipsometry for both methods. A maximum relative activity of 0.70-0.80 was reached between five and nine monolayers of the immobilized enzyme. However, the relative activity for covalent immobilization was greater than that of metal-ion complexation. Covalent esterification resulted in similar temperature dependence as free enzyme, whereas metal-ion complexation showed no temperature dependence indicating a significant change in conformation.

Introduction Immobilization of biomolecules on solid supports is of great interest for applications in biosensors and biotechnology in general.1-18 In the past few years, research in this area has examined a range of binding chemistries, substrates, and * Corresponding author. E-mail: [email protected]. † Chemical and Biological Engineering. ‡ Department of Physics. § Materials Science and Engineering. (1) Akgol, S.; Bayramoglu, G.; Kacar, Y.; Denizli, A.; Arica, M. Y. Polym. Int. 2002, 51, 1316. (2) Barrias, C. C.; Martins, C. L.; Miranda, C. S.; Barbosa, M. A. Biomaterials 2005, 26, 2695. (3) Battistel, E.; Bianchi, D.; Rialdi, G. Pure Appl. Chem. 1991, 63, 1483. (4) Bayramoglu, G.; Akgol, S.; Bulut, A.; Denizli, A.; Arica, M. Y. Biochem. Eng. J. 2003, 14, 117. (5) Becker, M. L.; Remsen, E. E.; Pan, D.; Wooley, K. L. Bioconjugate Chem. 2004, 15, 699. (6) Bencina, M.; Bencina, K.; Strancar, A.; Podgornik, A. J. Chromatogr., A 2005, 1065, 83. (7) Castner, D.; Ratner, B. D. Surf. Sci. 2002, 500, 28. (8) Christman, K. L.; Requa, M. V.; Enriquez-Rios, V. D.; Ward, S. C.; Bradley, K. A.; Turner, K. L.; Maynard, H. D. Langmuir 2006, 22, 7444. (9) Cosnier, S. Biosens. Bioelectron. 1999, 14, 443. (10) Dai, J. H.; Bao, Z. Y.; Sun, L.; Hong, S. U.; Baker, G. L.; Bruening, M. L. Langmuir 2006, 22, 4274. (11) Danisman, T.; Tan, S.; Kacar, Y.; Ergene, A. Food Chem. 2004, 85, 461. (12) DeLouise, L. A.; Miller, B. L. Anal. Chem. 2005, 77, 1950. (13) Feng, C.; Zhang, Z.; Forch, R.; Knoll, W.; Vancso, G.; Schonherr, H. Biomacromolecules 2005, 6, 3243. (14) Grazu, V.; Abian, O.; Mateo, C.; Batista-Viera, F.; Fernandez-Lafuente, R.; Guisan, J. M. Biotechnol. Bioeng. 2005, 90, 597. (15) Hu, Y.; Das, A.; Hecht, M.; Scoles, G. Langmuir 2005, 21, 9103. (16) Lee, C. S.; Lee, S. H.; Park, S. S.; Kim, Y. K.; Kim, B. G. Biosens. Bioelectron. 2003, 18, 437. (17) Liu, X. S.; Jang, C. H.; Zheng, F.; Jurgensen, A.; Denlinger, J. D.; Dickson, K. A.; Raines, R. T.; Abbott, N. L.; Himpsel, F. J. Langmuir 2006, 22, 7719. (18) Tugulu, S.; Arnold, A.; Sielaff, I.; Johnsson, K.; Klok, H. A. Biomacromolecules 2005, 6, 1602.

techniques for immobilizing mainly small molecules such as biotin and proteins such as enzymes and antibodies.8-10,19 Immobilized enzymes are particularly of interest due to their broad spectrum of use in numerous biotechnology processes as they can be easily removed from the reaction mixture preventing cross contamination.20,21 The critical considerations when employing bound enzymes are stability, activity, and concentration. The stability of enzymes is known to be enhanced by covalent binding, or cross-linking to a substrate or template.14 However, the activity is generally found to decrease in an immobilized enzyme compared to the free enzyme.20 The amount immobilized can be increased by moving from a two-dimensional (2D) platform such as a self-assembled monolayer (SAM) to a high surface area three-dimensional (3D) scaffold such as hydrogels, polymer coatings, or polymeric brushes.10,17 Polymer brushes offer certain advantages over other materials as they are covalently anchored to the substrate providing excellent mechanical stability and present a high surface area template with functionality controllable by monomer type and brush length. The three main reactions that are commonly used for immobilization of enzymes are esterification, amidation, and nitrilotriacetate (NTA)-Cu2+ complexes.10 Hence poly(acrylic acid),10 poly(glycidyl methacrylate),20 and poly(methacrylic acid)22 which provide the required carboxylic acid or epoxide functionality have been the natural choices as templates. Poly(acrylic acid) (PAA) has been used to immobilize biomolecules (19) Bayramoglu, G.; Yalcin, E.; Arica, M. Y. Colloids Surf., A 2005, 264, 195. (20) Xu, F. J.; Cai, Q. J.; Li, Y. L.; Kang, E. T.; Neoh, K. Biomacromolecules 2005, 6, 1012. (21) Bayramoglu, G.; Kaya, B.; Arica, M. Y. Food Chem. 2005, 92, 261. (22) Harris, B. P.; Kutty, J. K.; Fritz, E. W.; Webb, C. K.; Burg, K. J. L.; Metters, A. T. Langmuir 2006, 22, 4467.

10.1021/la702510z CCC: $40.75 © 2008 American Chemical Society Published on Web 12/13/2007

914 Langmuir, Vol. 24, No. 3, 2008

by both ionic interactions and covalent attachment.23 For example, PAA brushes were used to immobilize model proteins such as, hen egg-white lysozyme and bovine serum albumin (BSA), which have a net positive and negative charge, respectively, at neutral pH.10 More recently, among enzymes, glucose oxidase has been immobilized by Xu et al. on poly(glycidyl methacrylate) brushes with a relative activity of 60%.20 Most of these reports in the literature on immobilization of other proteins on polymeric brushes have not reported the temperature dependence of activity or relative activity as a function of binding chemistry. Ribonuclease A (RNase A) is a single subunit enzyme with a molecular weight of 14.7 kDa used to degrade RNA in many biotechnology applications.3 RNase A catalyzes cleavage of the phosphodiester bond between the 5′-ribose of a nucleotide and the phosphate group attached to the 3′-ribose of an adjacent pyrimidine nucleotide forming a 2′,3′-cyclic phosphate that is then hydrolyzed to the corresponding 3′-nucleoside phosphate. One of the most important uses of RNase A is in the removal of RNA from DNA preparations. This is a critical step in the downstream processing of biopharmaceutical plasmid DNA that is used as an active pharmaceutical ingredient in gene therapy and DNA vaccination. RNA removal is achieved by addition of RNase A to the sample, and subsequently, the enzyme is removed by an additional purification step. A good alternative would be the use of immobilized RNase A on a solid support allowing for easy separation of this enzyme from the plasmid DNA mixture. Current methods to purify plasmid DNA such as agarose gel electrophoresis extraction, column chromatography, and selective adsorption of DNA using solid-phase support or glass beads yield pure DNA, but the yield is poor due to multiple steps and often leads to incompatibility of the purified DNA with the subsequent reagents. RNase A has also been immobilized in polymer matrix such as styrene, nylon, and inorganic solid supports such as silica and glass.24-26 In this paper, we examine the feasibility of using polymeric brushes as 3D scaffolds to covalently immobilize RNase A with high relative activity. We specifically studied the use of PAA brushes synthesized by atom-transfer radical polymerization (ATRP) to immobilize RNase A by covalent immobilization and a high-capacity NTA-Cu2+ complexation. The effect of brush length and binding chemistry on the kinetics of binding, further activity, and temperature dependence of activity of the bound enzyme was systematically analyzed by a combination of ellipsometry, UV absorbance, X-ray photoelectron spectroscopy (XPS), and near-edge X-ray absorption fine structure spectroscopy (NEXAFS). We further show that these types of immobilization are transferable to glass microbeads for easy use in plasmid DNA procedures. Materials and Methods tert-Butyl acrylate, N,N,N′,N′,N′-pentamethyldiethylenetriamine (PMDETA), copper(I) bromide, ethyl-2-bromoisobutyrate, N-hydroxysuccinimide (NHS), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide (EDC), trizma hydrochloride, RNase A from bovine pancreas (E.C. 3.1.27.5), ribonucleic acid (RNA) from baker’s yeast (S. cereVisiae), copper(II) sulfate, NR′NR-bis(carboxymethyl)-L-lysine hydrate (NTA), tetrahydrofuran (THF), ethanol, toluene, and methanol were purchased from Sigma Aldrich Co. tert-Butyl acrylate (t-BA) was purified by passing through a column of basic alumina (23) Hollmann, O.; Czeslik, C. Langmuir 2006, 22, 3300. (24) Baek, W. O.; Vijayalakshmi, M. A. Biochim. Biophys. Acta 1997, 1336, 394. (25) Horak, D.; Rittich, B.; Safar, J.; Spanova, A.; Lenfeld, J.; Benes, M. J. Biotechnol. Prog. 2001, 17, 447. (26) Younus, H.; Owais, M.; Rao, D. N.; Saleemuddin, M. Biochim. Biophys. Acta 2001, 1548, 114.

Cullen et al. and stored at -10 °C. PMDETA and ethyl-2-bromoisobutyrate were purified by vacuum distillation. Chloro(dimethyl)silyl]propyl 2-bromo2-methylpropanoate (CPBMP) was synthesized using a previously reported procedure.27 All other materials were used as received. The term substrate is used to describe both silicon wafers and beads throughout the Materials and Methods. Preparation of Silicon Substrates.27 Silicon wafers (100, p-type) were diced into 1 cm2 pieces and cleaned by sonicating in toluene for 5 min followed by rinsing with acetone and absolute ethanol and drying under a nitrogen stream. The substrates were further cleaned with piranha wash (3:1 H2SO4/H2O2) for 1 h and rinsed with copious amounts of deionized water, acetone, and absolute ethanol. They were then dried in a stream of nitrogen. The substrates were placed in a reactor in nitrogen atmosphere, followed by the addition of 50 mL of anhydrous toluene and dropwise addition of 25 µL of CPBMP. CPBMP was allowed to immobilize on SiOx for 16 h. The samples were then removed and rinsed sequentially with toluene, acetone, and absolute ethanol. The functionalized silicon substrates were dried in a stream of nitrogen and stored in a desiccator until further use. Bead Preparation. Glass beads purchased from Polysciences, Inc. were washed in 10% HNO3 for 1 h at 100 °C.28 The beads were then washed with deionized water and dried in vacuum at 100 °C. The ATRP initiator was immobilized by suspending the beads in 10 mL of toluene and 50 µL of CPBMP under nitrogen for 16 h. The beads were then separated by centrifugation, washed with THF and ethanol, and dried under reduced pressure at room temperature. Polymerizations. Copper(I) bromide and the functionalized substrate (glass beads or silicon substrate) were added to a clean dry air-free polymerization flask. The flask was then evacuated and back-filled three times with nitrogen. tert-Butyl acrylate, acetone, PMDETA, and ethyl-2-bromoisobutyrate were then added via an air-free syringe, and the flask was degassed with three freezepurge-thaw cycles. The polymerization was then allowed to proceed at 80 °C for 24 h. The Si-g-P(t-BA) substrate was removed, rinsed with THF, acetone, and ethanol, dried under nitrogen, and stored in a desiccator. Pyrolysis of Polymer Brushes. The substrate was placed in an oven preheated to 200 °C for 30 min to pyrolize Si-g-P(t-BA) to Si-g-PAA. The substrate was allowed to cool before it was rinsed with deionized water and ethanol and dried in a stream of nitrogen. Activation and Immobilization of RNase A to Si-g-PAA. Si-g-PAA was immobilized with RNase A by two methods: method A, covalent immobilization, and method B, metal-ion complexation. Method A: CoValent Immobilization. The substrate was placed into dry vials, 00.1 M EDC and 0.1 M NHS were added, and the reaction was allowed to proceed for 30 min. After activation the substrate was rinsed and reacted in a 1 mg/mL RNase A solution in Tris-NaCl buffer for 24 h at room temperature. The substrate was rinsed sequentially with buffer, water, and ethanol and dried with a stream of nitrogen. Method B: Metal-Ion Complexation. The substrate was placed in a dry vial, 0.1 M EDC and 0.1 M NHS were added, and the reaction was allowed to proceed for 30 min. After activation, the substrate was rinsed and placed in a 0.1 M aminobutyl-NTA solution at pH ∼10. The reaction was allowed to proceed for 1 h. Following the conjugation with aminobutyl-NTA, the substrate was rinsed with deionized water and placed in a 50 mM CuSO4 solution for 2 h. After activation, the substrate was rinsed with deionized water and reacted in a 1 mg/mL RNase A solution in Tris-NaCl buffer for 24 h at room temperature. The substrate was rinsed sequentially with buffer, water, and ethanol and dried with a stream of nitrogen. Enzyme Binding Kinetics. To monitor the binding of RNase A to the PAA brush, the activated substrate was placed into a 1 mg/mL RNase A solution in Tris-NaCl buffer. The substrate was removed (27) Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immaraporn, B.; Gelman, A.; Luokala, B. B.; Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.; Pakula, T. Macromolecules 1999, 32, 8716. (28) Gomez, J. L.; Bodalo, A.; Gomez, E.; Bastida, J.; Hidalgo, A. M.; Gomez, M. Enzyme Microb. Technol. 2006, 39, 1016.

Polymeric Brushes for Immobilizing RNase A

Langmuir, Vol. 24, No. 3, 2008 915

at various time intervals, rinsed sequentially with buffer, water, and ethanol, and dried with a stream of nitrogen. The surface density of RNase A was estimated by measuring the thickness of dry brushes by ellipsometry using eq 1. σ ) Fhd

(1)

σ is the surface density of RNase A (g/cm2), F is the density of RNase A (g/cm3), and hd is the height change due to the immobilization of RNase A (cm).22 Enzyme Activity Measurements. To monitor the activity of the immobilized RNase A, the immobilized enzyme (1 cm2 silicon wafer or 5 mg of beads) was reacted with a RNA (1 mg/mL; Tris-NaCl) solution for 1 h at 37 °C. The temperature was varied from 25 to 80 °C to study the temperature dependence of enzymatic activity. The immobilized enzyme was then removed (beads were removed by centrifugation), and the absorbance of the solution due to remaining RNA was measured using a Cary50 UV-vis spectrometer at 300 nm. The activity (ar) and relative activity (RA) were determined as follows: ar )

A0 - Af A0

(2)

A0 ) absorbance of digestion mixture at t ) 0 min Af ) absorbance of digestion mixture at t ) 60 min RA )

aimmobilized r afree r

(3)

aimmobilized ) activity of immobilized RNase r afree ) activity of free RNase r Characterization. Ellipsometry. The thickness of the polymer brushes grafted on the silicon substrate was determined by a Rudolph Auto EL null ellipsometer. Measurements were made at three wavelengths: 632.8, 546, and 405 nm, at an angle of incidence of 70°. Three spots on each sample were measured, and four replicates were used. The reported thicknesses are for the dry brushes. The thickness was determined using FilmEllipse software version 1.1 (Scientific Company Intl.) from measurements of the ellipsometric angles Ψ and ∆ made at all three wavelengths. ThermograVimetric Analysis (TGA). The weight loss was determined using a thermogravimetric analyzer Q50 from TA Instruments. Approximately 10 mg of beads was used for each measurement. Unmodified, PAA, PAA-RNase A, and PAA-NTA-Cu2+-RNase A beads were investigated. The weight loss was measured in nitrogen atmosphere in the temperature range of 25-500 °C with a heating rate of 20 °C/min. XPS Measurement. The surface chemical composition of the polymer brushes and immobilized enzyme were determined by XPS and monitored through all the steps of the synthesis. The measurements were performed on a Perkin-Elmer 5400 ESCA spectrometer: Phi model 5400 using a Mg X-ray source (300 W, 15 kV) at a take off angle of 0°. The hemispherical energy analyzer was operated in the hybrid mode with a 1 mm × 3.5 mm selected area aperture. Survey spectra were collected at constant pass energy of 89.45 eV with a scan step size of 1.0 eV. High-resolution spectra were collect with pass energy of 35.75 eV and a step size of 0.05 eV. NEXAFS Measurement. NEXAFS spectroscopy was performed at the C 1s, N 1s, and S 1s edges, both in the total electron yield and the fluorescence yield mode. For the electron yield of the C 1s and N 1s edges, we used the 10-mTGMmonochromator at the Synchrotron Radiation Center (SRC) of the University of Wisconsins Madison. The S 1s edge data in both the electron and fluorescence yield mode were taken at the double-crystal monochromator (DCM) of the Canadian Synchrotron Radiation Facility (CSRF) at the SRC, using InSb(111) crystals and a nine-element solid-state Ge detector. Fluorescence yield data of the N 1s edge were acquired at beamline

Figure 1. C(1s) XPS spectra of Si-g-PAA (open circles) and Si-g-PtBA (filled squares) with a brush thickness of 50 nm. C-H peak is located at 286 eV, C-O peak is at 288.5 eV, and CdO is at 290 eV. 8.0 of the Advanced Light Source (ALS) in Berkeley using a channel plate with an Al filter.

Results and Discussion Immobilization of RNase A. In comparison to ionic interactions, covalent immobilization of proteins offers distinct advantages in terms of stability toward subsequent processing steps. We choose to immobilize RNase A to PAA brushes synthesized by pyrolysis of poly(tert-butyl acrylate) (PtBA). PAA is attractive as it swells to more than 4 times its initial thickness in aqueous medium which is likely to facilitate binding of large biomolecules10 and also presents a high density of the desired -COOH groups for further functionalization. The PtBA brushes were grown from ATRP initiators immobilized on SiOx. A trichlorosilane derivative CPBMP was self-assembled as the surface-anchored ATRP initiator.29 The growth of the brushes was monitored by both XPS and ellipsometry. Brush thicknesses in the range of 5-80 nm were synthesized by adjusting the ratio of t-BA to ethyl-2-bromoisobutyrate from 100 to 2000. The XPS analysis of the brushes confirmed the surface chemical composition. The ratio of the C-H and CdO peaks was approximately 9:1, which is consistent with the composition of Pt-BA brushes. The Pt-BA brushes were converted to PAA by heating the substrate to 200 °C for 30 min. Other methods such as acid hydrolysis can attack the ester group on the ATRP initiator and cleave the brush off the surface, whereas thermal deprotection of tert-butyl groups has been shown to be quite effective.30 The XPS C(1s) spectra of a PtBA and PAA brushes are shown in Figure 1. The effectiveness of pyrolysis was confirmed by the reduction of the C-O side peak and decrease in the C/O ratio in XPS of the resulting brush, which is consistent with the expected composition of PAA brush. The thickness of the polymer brush decreased to nearly half (e.g., 30 nm for PtBA brush to 16 nm for PAA brush) of the original brush length after pyrolysis due to the loss of the tert-butyl group, and the contact angle of the surface decreased from 90° to 45°. The change in XPS spectra along with the thickness and contact angle change confirm the pyrolysis of the brush to PAA. RNase A was immobilized on the PAA brushes by two different methods to evaluate the effect of binding chemistry on the amount immobilized and the relative activity. Method A is a conventional EDC coupling using NHS as an activator, which reacts with the pendant -NH2 on the amino acid residue lysine (Scheme 1). Method B uses metal-ion complexes bound to PAA brushes. PAA brushes can be converted to metal-ion complex containing (29) Ayres, N.; Cyrus, C. D.; Brittain, W. J. Langmuir 2007, 23, 3744. (30) Treat, N. D.; Ayres, N.; Boyes, S. G.; Brittain, W. J. Macromolecules 2006, 39, 26.

916 Langmuir, Vol. 24, No. 3, 2008

Cullen et al.

Scheme 1. Immobilization of RNase A by Metal-Ion Complexation and Covalent Immobilization

films by treatment with EDC/NHS, aminobutyl-NTA, and Cu2+. The metal-ion complex strongly coordinates with ligands on proteins, such as histidine (Scheme 1). This method has been used widely for metal affinity chromatography and was recently reported by Dai et al.10 as an attractive high-capacity binding chemistry for proteins. The main advantage of the method B is to minimize the competitive hydrolysis of NHS-activated carboxylic acid by water during protein immobilization. However, the effect of utilizing the histidine residues on the activity of the enzyme or the temperature dependence of activity was not reported in the paper. We were interested in examining the applicability of this method to binding of RNase A and on the temperature dependence of activity. Both methods allowed immobilization of RNase A as confirmed by XPS. The nitrogen peaks originate from the bound protein, whereas the carbon peak has contributions from both protein and the polymer brush. Figure 2a is a representative C(1s) spectrum showing the introduction of the C-NH peak from RNase A immobilized by method A to a 50 nm thick PAA brush. The immobilization of RNase A by both methods resulted in similar XPS spectra, and the C(1s) peaks are indistinguishable. Figure 2b shows the emergence of a new N(1s) peak from the bound RNase A for both types of binding and clearly shows the increase in the N(1s) signal due to the immobilization. Although XPS data confirms the chemistry, it is not possible to quantitatively estimate the amount of bound enzyme using XPS data alone. The penetration depth of XPS allows only the surface of the brush to be examined; therefore, we only see differences in the spectra at short brush lengths. Furthermore, nitrogen from NHS is displaced by RNase A after binding by method A, whereas the nitrogens from (NTA)-Cu2+ complexes are retained after binding by method B (Scheme 1). Hence, we used ellipsometry to quantify the binding of RNase A. Surface Density (SD) of Immobilized RNase A by Ellipsometry. To examine how the binding scales with brush length, the SD was measured for dry brush thicknesses in the 5-80 nm range. As expected, method B bound more RNase A than method A. Figure 3 shows the SD at various brush thicknesses for both binding methods. SD increases monotonically with brush thickness for both immobilization methods. These results indicate that RNase A diffuses into the polymer brush and binds along

the length of the chain. Due to the extended conformation of the polymer brush in aqueous media the enzyme is able to penetrate throughout the brush. Furthermore, the surface density did not level off at the brush lengths examined. Poly(acrylic acid) is known to swell considerably in water, resulting in an extended chain structure on the substrate in deionized water.10 On the basis of these studies, a dry 80 nm PAA brush can swell to 400 nm in water. The resulting thickness change of 320 nm in aqueous solution on protein immobilization would translate into 42 µg/cm2 of immobilized RNase A. The maximum mass of immobilized RNase A in our experiments is ∼11 and ∼6 µg/cm2 for method B and A, respectively. These surface densities are 30 and 16 times the estimated mass bound for a monolayer of RNase A (0.37 µg/cm2 for a thickness approximately 2.8 nm).31 This is in good agreement with other reports that have shown that increasing brush thickness translates into increased amount of bound protein.10 Greater amount of RNase A can be bound to the surface by synthesizing longer brushes. However, it may not be beneficial to bind more RNase A due to the effect crowding may have on activity as discussed in the next section. We examined RNase A immobilized by method A to PAA with NEXAFS to determine if there is a preferred orientation. Figure 4 shows the N (1s) NEXAFS spectra in the total electron yield (TEY) mode at normal incidence for RNase A immobilized by method A to a PAA brush with 13 nm length, in comparison with the RNase A immobilized via a SAM in a previous study.17 Analogous N (1s) spectra of the Si substrate and Si-g-PAA (not shown) do not exhibit any of the features shown in Figure 4, confirming that RNase A was successfully immobilized on the surfaces and that N 1s spectra are characteristic of the adsorbed protein molecules. The raw spectra are normalized by subtracting linear preedge background, allowing for a semiquantitative determination of the coverage from the spectral intensity. The height of the first peak in Figure 4 is significantly greater for the Si-g-PAA-RNase A than for SAM-RNase A. NEXAFS spectra were also taken at angle 30° from grazing incidence. No polarization effect was observed, indicating that the RNase A (31) Dimilla, P. A.; Folkers, J. P.; Biebuyck, H. A.; Harter, R.; Lopez, G. P.; Whitesides, G. M. J. Am. Chem. Soc. 1994, 116, 2225.

Polymeric Brushes for Immobilizing RNase A

Langmuir, Vol. 24, No. 3, 2008 917

Figure 4. N1s NEXAFS spectra of RNase A immobilized on a silver substrate via a self-assembled monolayer (ref 17) (blue) and on a silicon substrate via a 13 nm long PAA brush (red). By normalizing the NEXAFS intensity to the preedge background, one obtains a semiquantitative spectrum.

Figure 2. XPS spectra of Si-g-PAA-RNase A with an initial PAA brush thickness of 50 nm. (a) C(1s) spectra of Si-g-PAARNase A (open circles): C-H peak is located at 286 eV, C-NH is at 287 eV, C-O is at 288.5 eV, and CdO is at 290 eV. (b) Comparison of the Si-g-PAA-NTA-Cu2+ (unfilled triangles), Si-g-PAA-NTA-Cu2+-RNase A (unfilled circles), and Si-gPAA-RNase A (filled squares) N(1s) spectrum.

Figure 5. N1s NEXAFS spectra of RNase A immobilized via PAA brushes with various brush thicknesses: 13 nm (blue), 28 nm (green), 39 nm (red). When normalized to the preedge background, the NEXAFS intensity increases monotonically with the brush thickness.

molecules were immobilized without any preferred orientation. This is consistent with the immobilization method, which connects the RNase A at the 10 lysine residues. This type of attachment also produced random orientation in a SAM as reported in our previous study.17 Figure 5 shows the N 1s NEXAFS spectra of RNase A immobilized by method A with various brush thicknesses of 13, 28, and 39 nm. The intensity of the N 1s absorption increases monotonically with brush thickness, which is consistent with the ellipsometric results in Figure 3 (full squares). The intensity does not increase linearly with brush length, as one might expect naively. This can be explained by the finite probing depth of NEXAFS spectra in the electron yield mode32 (about 5 nm),

which is only a fraction of the chain length (data not shown). Method B also did not show any preferred orientation, either. This is expected since binding can occur at any of the four histidine residues on the enzyme (data not shown). The NEXAFS spectrum of RNase A has been characterized previously17 for a SAM. Similar types of NEXFAS measurements on RNase immobilized at a surface by SAMs showed an orientation effect comparable to that expected for properly folded RNase.17 After about 12 h in vacuum, this orientation effect disappears, indicating denaturation. The spectra shown in Figures 4 and 5 were taken well before this effect takes place. Similar to the SAM-RNase A, the Si-g-PAA-RNase A shows a dominant peak at 401.5 eV in Figure 4. It is assigned to a transition from the N 1s core level into the π* orbital that is delocalized over the peptide bond connecting two adjacent amino acids in polypeptides and proteins.17,32,33 The region of the σ* orbitals shows a peak at ∼405.4 eV, which is assigned to the transition to the N-C σ* orbital. A high-lying peak at ∼411.0 eV has been attributed to N 1s to σ*C-NH2, σ*C-NH, and σ*CONH transitions.33 Kinetics and Thermodynamics of RNase A Immobilization. The immobilization kinetics of RNase A to Si-g-PAA was estimated using thickness measurement of the enzyme-im-

(32) Cooper, G.; Gordon, M.; Tulumello, D.; Turci, C.; Kaznatcheev, K.; Hitchcock, A. R. J. Electron Spectrosc. Relat. Phenom. 2004, 137-40, 795.

(33) Gordon, M. L.; Cooper, G.; Morin, C.; Araki, T.; Turci, C. C.; Kaznatcheev, K.; Hitchcock, A. P. J. Phys. Chem. A 2003, 107, 6144.

Figure 3. Surface density of RNase A bound by the two methods to a Si-g-PAA brush as a function of initial dry thickness of the brush. Open circles correspond to method B, and filled squares correspond to method A.

918 Langmuir, Vol. 24, No. 3, 2008

Cullen et al.

Figure 6. Comparison of RNase A surface density for method B (unfilled circles) and method A (filled squares) on a 50 nm thick PAA brushes as a function of time.

mobilized substrate by ellipsometry. The polymer brush thickness was fixed at 50 nm for this study. Figure 6 shows the kinetic binding curve of RNase A for both binding schemes. The rate of binding for both methods is low as the time constant is of the order of hours. Both binding methods reached a saturation point around 15-16 h and immobilized similar amount of enzyme up to 4 h. After 4 h the more efficient method B binds nearly double the RNase A compared to method A. Using Fick’s Law for linear diffusion into a thin film, we calculated the diffusion coefficient for RNase A for both binding methods.10 We fit our data in Figure 6 with the solution to Fick’s second law, which is represented by eq 4 where Γ is the protein coverage (µg/cm2) at time t (s), Γm is the equilibrium coverage (µg/cm2), D is the diffusion coefficient (cm2/s), and l is the film thickness (cm).

Γ Γm

∞

)1-

∑

n)0(2n

[

]

-D(2n + 1)2π2t

8 + 1)2π2

exp

4l2

ΓmC Kd + C

Figure 8. Relative activity and activity as a function of the surface density of RNase A. The maximum for method A (filled squares) occurs near five monolayers of enzyme, whereas for method B (unfilled circles) it occurs near nine monolayers.

(4)

The equilibrium coverage for a 50 nm thick brush for method B was estimated from experiments to be 6.9 µg/cm2 (D ) 0.63 × 10-14 cm2/s) and 3.9 µg/cm2 (D ) 0.71 × 10-14 cm2/s) for method A. Similar values of D for both methods indicate similar binding kinetics, but the thermodynamic equilibrium coverage is different. On the basis of similar kinetics, the rate-limiting step for immobilization is the diffusion of the RNase A into the brush layer. The low diffusion coefficient for RNase A measured here is similar to coefficients determined for diffusion of BSA, which has a molecular weight of 66.2 kDa compared to 13.7 kDa for RNase A, into polymer brushes.10 Surface protein adsorption thermodynamics occasionally follow Langmuir isotherms. The thermodynamics for binding of RNase A to the activated PAA brush was investigated as a function of the concentration of RNase A. The polymer brush thickness was kept consistent at 50 nm for this study. A Langmuir isotherm, which is represented by eq 5, was used to fit our data in Figure 7:

Γ)

Figure 7. Adsorption isotherms for method B (unfilled circles) and method A (filled squares). This study was done with a 50 nm thick poly(acrylic acid) (PAA) brush with 24 h of immobilization.

(5)

where C is the RNase A concentration in solution (mg/mL) and Kd is the dissociation constant (mg/mL). Method B again shows higher surface density at all concentrations. The surface density of RNase A begins to level off above 2 mg/mL for both methods. The equilibrium coverage from method A is about half the coverage obtained from method B. This is probably due to competing hydrolysis of the NHS ester. The adsorption isotherms are shown in Figure 7.

The Langmuir isotherm fit estimated a value of 8.7 µg/cm2 for Γm and a dissociation constant (Kd) of 0.11 mg/mL (8.0 × 10-6 M) for method B and a Γm of 5.7 µg/cm2 and Kd of 0.31 mg/mL (2.3 × 10-5 M) for method A. The dissociation constant found for method A takes into account the competitive hydrolysis of the PAA-NHS ester, which may account for the higher dissociation constant. For prior experiments, the concentration of RNase A in solution was 1 mg/mL, but as indicated by the isotherms a solution of 2 mg/mL or higher is optimum. The nearly 3 times higher dissociation constant and 1.5 times lower equilibrium coverage is indicative of a less efficient binding for method A compared to method B. Enzyme Activity. Proteins bind by method B through the coordination of histidine residues to metal ions. The enzyme activity may be disrupted if the histidine units lie in the active site. In Figure 8, the measured relative activity and activity is plotted as a function of surface density. The trend of relative activity is similar for RNase A bound by either method. The activity of both immobilized RNase A goes through a maxima as a function of surface density. This trend for RNase A is similar to other enzymes which show that increased surface density may decrease the activity of immobilized enzymes.34-36 The maximum for method A occurs near 2.0 µg/ cm2 which correlates to approximately five monolayers of RNase A. However, method B has maxima near 3.5 µg/cm2, which correlates to approximately nine monolayers of immobilized enzyme. This difference in SD at maximum activity can be attributed to the coordination of Cu2+ to the active site making (34) Chen, H.; Hsieh, Y. L. Biotechnol. Bioeng. 2005, 90, 405. (35) Haupt, B.; Neumann, T.; Wittemann, A.; Ballauff, M. Biomacromolecules 2005, 6, 948. (36) Neumann, T.; Haupt, B.; Ballauff, M. Macromol. Biosci. 2004, 4, 13.

Polymeric Brushes for Immobilizing RNase A

Figure 9. Temperature dependence of relative activity of RNase A bound by method B (unfilled circles) and method A (filled squares) for brush lengths of 50 nm.

the bound enzyme less active for the same SD. The initial increase in relative activity with SD is probably due to a higher concentration of RNase A. At an intermediate surface density of RNase A, the ability for RNA to diffuse to the active site of RNase A will decrease leading to the observed decrease in relative activity. The activity of the immobilized enzyme increases with increasing SD, but levels at 2.0 and 3.5 µg/cm2 for method A and B, respectively. The leveling off corresponds to the maximum seen in relative activity. This trend is attributed to crowding of RNase A, which will prevent the RNA from accessing the active site. We showed earlier that the time constant for diffusion of RNase A is of the order of hours. The diffusion of RNA will be similar to the diffusion of RNase A into the film. For RNA to fully diffuse into the film, the RNA would need to be in contact with the brushes for over 16 h. The activity measurements in Figure 8 were allowed to proceed for only 1 h. However, it is encouraging to note that a relative activity as high as 80% can be achieved from the immobilized RNase A even at these short binding times. We did investigate longer digestion periods and observed a corresponding increase in the activity. For the surface densities less than this steric hindrance limit, the relative activity of RNase A bound by method B is less than that bound by method A. Temperature Dependence of Immobilized RNase A. The activity of the immobilized enzyme as a function of temperature was investigated using a brush thicknesses of 50 nm. The temperature dependence of the immobilized enzyme can help determine if the binding of the enzyme has a significant effect on its conformation. If the enzyme maintains its temperature dependence as seen with the native enzyme, then it can be postulated that the conformation is not drastically different from its native state. From Figure 9, it is apparent that the method of binding plays a drastic role on the temperature dependence of activity. Method A has a maximum at 70 °C for both the absolute activity and relative activity, which is near the literature value of 65 °C.37 Interestingly, the activity of enzyme bound by method B was independent of temperature. This leads us to postulate that method B has a significant effect on the conformation of the enzyme. To the best of our knowledge this is the first observation of temperature dependence of activity of PAA-bound enzyme as a function of binding chemistry. The residues in the active site of RNase A are His12, Thr45, Lys66, Asn67, His119, Phe120, and Asp121. Examining the structure of RNase A, we find that the lysine residues are located on the surface of the protein and are not an integral part of the active site of the enzyme, whereas the histidine residues are part (37) Lin, M. C. J. Biol. Chem. 1970, 245, 6726.

Langmuir, Vol. 24, No. 3, 2008 919

Figure 10. TGA data on beads immobilized with RNase A. (a) Unmodified glass beads showed 0.1% weight loss. (b) Bead-PAA showed 0.5% weight loss. (c) RNase A immobilized by method A had 1.2% loss and (d) immobilized by method B had 2.5% loss. RNase A immobilized by method B has double the weight loss of conventional immobilization.

of the active site, and it has been found in earlier studies that disruption or replacement of these residues leads to dramatic reduction in the activity of the enzyme.38 RNase A Immobilized to Glass Beads. The polymeric brushes with bound RNase A can be readily adapted in plasmid DNA purification protocols, if we can further demonstrate the viability of this approach on glass beads or magnetic beads. Here we show the immobilization of RNase A on glass microbeads with diameter ranging from 3 to 10 µm. Each step of immobilization was monitored using XPS and TGA to confirm and quantify the formation of polymer brushes and the immobilization of RNase A. The beads were monitored by XPS (data not shown) by placing approximately 10 mg onto carbon tape and taking the spectra of the beads. Each step of immobilizing RNase A with both methods was qualitatively confirmed with XPS. With TGA, we were able to quantify the amount of organic material bound to the beads, which includes the brushes and RNase A. We expect a one-step weight loss near 300 °C, which correlates with the decomposition of organics. In Figure 10, we see that polymer brush bound to the beads decomposes between 200 and 400 °C with a midpoint of 300 °C. The weight loss due to PAA was 0.5%. RNase A immobilized on PAA by method A shows a similar transition near 300 °C with a weight loss of 1.2%, and method B has a weight loss of 2.5%. This indicates that method B binds approximately double the enzyme compared to method A which is consistent with our observations on flat silicon substrates. Glass bead immobilized RNase A was allowed to digest RNA, and the activity was estimated for both binding methods The specific activity of the immobilized enzyme was determined by taking the activity found and normalizing it with the amount of bound RNase A. The specific activity of RNase A bound by method A was 18 mg-1 and that of method B was 3.8 mg-1. As the PAA brush length on the beads are the same, the difference in specific activity is attributed to the mode of binding. The temperature dependence on glass beads also followed similar trends as those observed on flat substrates.

Conclusions RNase A, which is used in downstream processing of plasmid DNA for removal of RNA, was successfully immobilized onto a Si-g-PAA brush. In this study, we have systematically examined the effect of binding chemistry and brush length on the surface density of bound enzyme, the kinetics and thermodynamics of binding, and the resulting temperature dependence (38) Park, C.; Schultz, L. W.; Raines, R. T. Biochemistry (N.Y.) 2001, 40, 4949.

920 Langmuir, Vol. 24, No. 3, 2008

of activity of the enzyme. The main conclusions from this work can be summarized as follows: (1) The metal-ion complexation exhibited double the surface density compared to the covalent binding. The increased surface density of RNase A was attributed to a higher equilibrium coverage and lower dissociation constant for the metal-ion complexation method. (2) A maximum relative activity of 0.7-0.8 and activity was reached between five and nine monolayers of the immobilized enzyme. Hence, further immobilization beyond nine monolayers does not offer increased activity in the brush configuration. (3) Enzymes bound by both methods showed similar maximum relative activity, but the maxima with metal-ion complexation method occurred at approximately 1.75 times the surface density achieved by covalent immobilization. The relative activity of method B was lower than method A except for 3.5 µg/cm2, which corresponds to the maximum for method B. This shift in relative activity indicates that the metal-ion complexation involves the histidine residue in the active site. (4) When examining the temperature dependence of activity of the immobilized enzyme, the covalent immobilization showed

Cullen et al.

behavior similar to that of the native enzyme; however, the RNase A immobilized using complexation showed no temperature dependence indicating a significant change in the conformation of the active site. From these studies polymeric brushes can be effectively used to immobilize over 30 times more proteins than monolayers which may lead to decrease in detection limits of protein microarrays. The brush chemistry could also be easily extended to magnetic beads for easy removal. However, the chemistry of immobilization and constitution of the active site is critical to retaining the activity of the enzymes. The ability to use random copolymer brushes should provide an additional handle in optimizing the activity and quantity of bound enzyme. Acknowledgment. This work was partially supported by the UW Graduate School funds UW-NSF Nanoscale Science and Engineering Center (DMR-0425880). We also acknowledge the use of Materials Science Center Facilities at University of WisconsinsMadison. LA702510Z