Anal. Chem. 2004, 76, 2813-2819
Secondary Structure Analysis of Proteins Embedded in Spherical Polyelectrolyte Brushes by FT-IR Spectroscopy Alexander Wittemann and Matthias Ballauff*
Physikalische Chemie I, Universita¨t Bayreuth, Universitaetsstrasse 30, D-95440 Bayreuth, Germany
The adsorption of bovine serum albumin (BSA), bovine β-lactoglobulin, and bovine pancreatic ribonuclease A onto spherical polyelectrolyte brushes (SPB) is reported. The SPB consist of narrowly distributed poly(styrene) core particles (diameter ∼100 nm) onto which linear chains of anionic polyelectrolytes are grafted. The polyelectrolyte shell consists of either the weak polyelectrolyte poly(acrylic acid) or the strong polyacid poly(styrenesulfonate). The SPB particles are dispersed in H2O at room temperature. The secondary structure of the proteins was investigated by Fourier transform infrared spectroscopy in transmission mode before and during adsorption to these colloidal brushes. The r-helix and β-sheet content of the proteins was nearly fully retained in the adsorbed state for all systems. Only in the case of BSA interacting with poly(styrenesulfonic) brushes could a slight loss of r-helix structure be observed. As the interaction of SPB and proteins can be controlled by the ionic strength in the buffer, additional experiments were performed to release the adsorbed protein. The amount of released protein was quantified and was found to be strongly dependent on the kind of protein and brush used. The secondary structure of the released proteins could be analyzed as well. An almost full preservation of secondary structure was found. This demonstrates that SPB are well-suited to immobilize proteins. The SPB can be charged and decharged under retention of the secondary structure of the biomolecules. Immobilization of proteins, especially enzymes and antibodies on solid supports, has attracted much interest in biotechnology in recent years.1-3 Suitable supports for practical applications should comply with two major requirements: (i) they should widely maintain the biological function of the biomolecules while preventing possible leaching out during the reaction,2 and (ii) a large surface should be provided to bind an appropriate number of molecules. The latter requirement is given by suspensions of * To whom correspondence should be addressed. Matthias.Ballauff@ uni-bayreuth.de. Fax: +49 921 55 2780. (1) Proteins at Interfaces: Fundamentals and Applications, Horbett, T. A., Brash, J. L., Eds.; ACS Symposium Series 602; American Chemical Society: Washington, DC, 1995. (2) Hartmeier, W. Immobilisierte Biokatalyse: Eine Einfu ¨ hrung; SpringerVerlag: Duesseldorf, Germany, 1986. (3) Copeland, R. A. Enyzmes, A Practical Introduction to Structure, Mechanism, and Data Analysis; Wiley-VCH: New York, 2000. 10.1021/ac0354692 CCC: $27.50 Published on Web 04/14/2004
© 2004 American Chemical Society
colloidal particles whose diameters are in the order of a few hundred nanometers.4 Immobilization can be performed by adsorption of the biomolecules from solution. A large number of studies have already been done on this aspect.5-16 It was shown, however, that adsorption on flat surfaces may lead to a loss of activity in the case of immobilized enzymes.13,16 This could be fully understood by a flattening and deformation of adsorbed proteins, which could be monitored for planar macroscopic surfaces.17-23 Proteins can also be immobilized via polyelectrolyte multilayers that were built up on colloidal objects.24-29 Caruso et al. found the activity of enzymes embedded in such multilayers widely retained. 24,25,29 However, secondary structure analyses were limited to planar films of polyelectrolyte multilayers so far. In these cases, the secondary structure was found to be almost retained for the embedded proteins.30-33 (4) Kawaguchi, H. Prog. Polym. Sci. 2000, 25, 1171. (5) Norde, W.; Lyklema, J. J. Colloid Interface Sci. 1978, 66, 257. (6) Norde, W.; Lyklema, J. J. Colloid Interface Sci. 1978, 66, 266. (7) Norde, W.; Lyklema, J. J. Colloid Interface Sci. 1978, 66, 277. (8) Norde, W.; Lyklema, J. J. Colloid Interface Sci. 1978, 66, 285. (9) Norde, W.; Lyklema, J. J. Colloid Interface Sci. 1978, 66, 295. (10) Peula, J. M.; Callejas, J.; de las Nieves, F. J. In Surface Properties of Biomaterials; West, R., Batts, G., Eds.; Butterworth-Heinemann: Manchester, U.K.,1994. (11) Peula, J. M.; de las Nieves, F. J. Colloids Surf., A 1994, 90, 55. (12) Yoon, J.-Y.; Kim, J.-H.; Kim, W.-S J. Colloid Interface Sci. 1996, 177, 613. (13) Norde, W.; Giacomelli, C. A. Macromol. Symp. 1999, 145, 125. (14) Giacomelli, C. A.; Vermeer, A. W. P.; Norde, W. J. Colloid Interface Sci. 2000, 231, 283. (15) Yoon, J.-Y.; Kim, J.-H.; Kim, W.-S. Colloids Surf., A 1999, 153, 413. (16) Oh, J.-T.; Kim, J.-H. Enzyme Microb. Technol. 2000, 27, 356. (17) Vermonden, T.; Giacomelli, C. E.; Norde, W. Langmuir 2001, 17, 3734. (18) Wertz, C. F.; Santore, M. M. Langmuir 2001, 17, 3006. (19) Wertz, C. F.; Santore, M. M. Langmuir 2002, 18, 706. (20) Wertz, C. F.; Santore, M. M. Langmuir 2002, 18, 1190. (21) Czeslik, C.; Winter, R. Phys. Chem. Chem. Phys. 2001, 3, 235. (22) Jackler, G.; Steitz, R.; Czeslik, C. Langmuir 2002, 18, 6565. (23) Czeslik, C.; Royer, C.; Hazlett, T.; Mantulin, W. Biophys. J. 2003, 84, 2533. (24) Fragneto, G.; Su, T. J.; Lu, J. R.; Thomas, R. K.; Rennie, A. R. Phys. Chem. Chem. Phys. 2000, 2, 5214. (25) Caruso, F.; Trau, D.; Mo¨hwald, D.; Renneberg, R. Langmuir 2000, 16, 1485. (26) Caruso, F.; Schu ¨ ler, C. Langmuir 2000, 16, 9595. (27) Caruso, F.; Schu ¨ ler, C. Macromol. Rapid Commun. 2000, 21, 750. (28) Caruso, F.; Fiedler, H.; Haage, K. Colloids Surf., A 2000, 169, 287. (29) Lvov, Y.; Caruso, F. Anal. Chem. 2001, 73, 4212. (30) Jin, W.; Shi, X.; Caruso, F. J. Am. Chem. Soc. 2001, 123, 8121. (31) Schwinte´, P.; Voegel, J.-C.; Picart, C.; Haikel, Y.; Schaaf, P.; Szalontai, B. J. Phys. Chem. B 2001, 105, 11906. (32) Schwinte´, P.; Ball, V., Szalontai, B.; Haikel, Y.; Voegel, J.-C.; Schaaf, P. Biomacromolecules 2002, 3, 1135.
Analytical Chemistry, Vol. 76, No. 10, May 15, 2004 2813
Figure 1. Schematic representation of the SPB used in this study. The particles consist of a solid poly(styrene) core onto which long polyelectrolyte chains are grafted. Two systems have been used in this study: (i) particles bearing chains of the weak polyelectrolyte poly(acrylic acid) (annealed brush) and (ii) particles having chains of the strong polyelectrolyte poly(styrenesulfonic acid) (quenched brush). As shown in previous studies, the salt concentration cs inside the brush differs markedly from the salt concentration outside ca if ca is low.34,35 The thickness L of the brush layer depends strongly on ca as has been shown recently.34,35
Protein-polyelectrolyte complexes have also attracted much interest in the past decades.34-41 A stabilizing effect by the polyelectrolyte chains could be shown that enabled longer shelf stabilities and prevented aggregation of the proteins.35 Xia et al. showed that enzymes were active during complexation.37 However, these systems present no real carrier-bound immobilization. Previously we presented spherical polyelectrolyte brushes (SPB) as a new class of colloidal particles for the adsorptive immobilization of proteins.42-45 The SPB schematically shown in Figure 1 consist of a solid poly(styrene) core onto which a dense layer of linear polyelectrolyte chains is attached.46-49 If the ionic strength is low, the brush layer is strongly stretched by the osmotic pressure of the counterions.47,48 We observed that proteins such as bovine serum albumin (BSA) are strongly adsorbed onto SPB as long as the ionic strength is kept low.42 Proteins adsorbed in the so-called osmotic limit49 will be fixed to the SPB even if the suspension is subjected to an exhaustive ultrafiltration against (33) Szyk, L.; Schaaf, P.; Gergely, C.; Voegel, J.-C.; Tinland, B. Langmuir 2001, 17, 6248. (34) Mu ¨ller, M.; Rieser, Th.; Dubin, P. L.; Lunkwitz, K. Macromol. Rapid Commun. 2001, 22, 390. (35) Dikov, M. M.; Osipov, A. P.; Egorov, A. M.; Karulin, A. Y.; Mustafayev, M. I.; Kabanov, V. A. J. Solid-Phase Biochem. 1980, 5, 1. (36) Andersson, M. M.; Hatti-Kaul, R. J. Biotechnol. 1999, 72, 21. (37) Macromolecular Complexes in Chemistry and Biology; Dubin, P., Bock, J., Davies, R. M., Schulz, D. N., Thies, C., Eds.; Springer-Verlag: New York, 1994. (38) Xia, J.; Mattison, K.; Romano, V.; Dubin, P. L.; Muhoberac, B. B. Biopolymers 1997, 41, 359. (39) Mattison, K. W.; Wang, Y.; Grymonpre´, K.; Dubin, P. L. Macromol. Symp. 1999, 140, 53. (40) Hattori, T.; Hallberg, R.; Dubin, P. L. Langmuir 2000, 16, 9738. (41) Porcar, I.; Gareil, P.; Tribet, C. J. Phys. Chem. B 1998, 102, 7906. (42) Borrega, R.; Tribet, C.; Audebert, R. Macromolecules 1999, 32, 7798. (43) Wittemann, A.; Haupt, B.; Ballauff, M. Phys. Chem. Chem. Phys. 2003, 5, 1671. (44) Wittemann, A.; Haupt, B.; Merkle, R.; Ballauff, M. Macromol. Symp. 2003, 191, 81. (45) Neumann, Th.; Haupt, B.; Ballauff, M. Macromol. Biosci. 2004, 4, 13. (46) Haupt, B.; Neumann, Th.; Ballauff, M., in preparation. (47) Guo, X.; Weiss, A.; Ballauff, M. Macromolecules 1999, 32, 6043. (48) Guo, X.; Ballauff, M. Langmuir 2000, 16, 8719. (49) Guo, X.; Ballauff, M. Phys. Rev. E 2001, 64, 051406.
2814
Analytical Chemistry, Vol. 76, No. 10, May 15, 2004
buffer solution of the same ionic strength. If, on the other hand, the ionic strength is high, almost no adsorption takes place. This effect can be used to release the adsorbed protein by increasing the ionic strength.42 Therefore, SPB are a unique system to immobilize proteins on colloidal particles in a controlled way. The major driving force for the adsorption process was found in the interaction of the negative polyelectrolyte chains with positive patches on the protein surface. The adsorbed protein patches act as a multivalent counterion of the polyelectrolyte brush.42 In return, quite a large number of counterions of brush and protein will be released and thus increase the entropy of the system. We found previously that the enzymatic activity of glucoamylase is retained during adsorption onto SPB.44,45 In addition to this, CD spectroscopy showed that the secondary structure of BSA liberated from the SPB has not been disturbed by the adsorption process.50 However, a direct analysis of the adsorbed protein by CD spectroscopy could not be performed due to the strong absorption and scattering of UV light by the SPB. The present study uses infrared spectroscopy to overcome the problem of the strong scattering of visible light because this method can be used in optically turbid media as well.51 Analysis of the secondary structure of proteins by Fourier transform infrared (FT-IR) spectroscopy in transmission mode51-54 as well as by attenuated total reflection (ATR)55 has made considerable progress during the past decade. Most FT-IR studies were performed in D2O to overcome the strong water absorption at the amide I region of the proteins56 Recent work has shown that quantitative subtraction of the H2O spectrum from the transmission or reflection spectra is now technically possible.51-54 For measurements in H2O, a short path length of 6 µm is required to avoid total adsorption by water.54 As a consequence, relatively high sample concentrations (>5 mg mL-1) are necessary. The secondary structure is generally derived from the IR spectra by the amide I band.51-54,57 In some cases, the amide II band is also included. Several reports on the derivation of the secondary structure were given recently.51-54 One approach is based on pattern recognition. This technique is based on a calibration matrix of infrared spectra of proteins with known secondary structure from X-ray crystallography. The partial leastsquares method (PLS) 58,59 has become widely accepted over the classical least-squares method because the analysis needs only one matrix inversion and generates loading vectors that are linear combinations of the calibration spectra.51 Moreover, it analyses one conformation at a time. Dousseau and Pe´zolet pointed out that PLS provides best results if both the amide I and the amide II bands are used and the assumption is made that the secondary structure is composed of four types: ordered helix, disordered (50) Das, B.; Guo, X.; Ballauff, M. Prog. Colloid Polym. Sci. 2002, 121, 34. (51) Jackler, G.; Wittemann, A.; Ballauff, M.; Czeslik, C. Spectroscopy, submitted. (52) Dousseau, F.; Pe´zolet, M. Biochemistry 1990, 29, 8771. (53) Surewicz, W. K.; Mantsch, H.; Chapman, D. Biochemistry 1993, 32 (2), 389. (54) Rahmelow, K.; Hu ¨ bner, W. Anal. Biochem. 1996, 241, 5. (55) Fabian, H.; Schultz, C. P. Fourier Transform Infrared Sectroscopy in Peptide and Protein Analysis. In Encyclopedia of Analytical Chemistry; Meyers, R. A., Ed.; Wiley: Chichester, U.K., 2000; p 5779. (56) Oberg, K. A.; Fink, A. L. Anal. Biochem. 1998, 256, 92. (57) Kalnin, N. N.; Baikalov, I. A.; Venyaminov, S. Y. Biopolymers 1990, 30, 1273. (58) Methoden der biophysikalischen Chemie; Winter, R., Noll, F., Eds.; Teubner: Stuttgart, Germany, 1998. (59) Haaland, D. M.; Thomas, E. V. Anal. Chem. 1988, 60, 1193.
Table 1. Characteristic Parameters of the Proteins: M Molecular Mass, rst Stokes Radius, and pI Isoelectric Point protein
M (kDa)
rst (nm)
pI
BSA BLG (variants A and B) RNase A (bovine pancreas)
66.3 18.4 13.7
3.61 1.79 1.92
5.1 5.1; 5.2 9.6
helix, β-sheet, and undefined conformation including all remaining elements.51 Correlation between the X-ray data and the IR results gave an error of 4.8% for R-helix, 3.7% for β-sheet, and 5.1% for the undefined structure.51 Despite the enormous recent technical advance in infrared spectroscopy and the various numbers of studies of proteins in aqueous solution, only few detailed studies are available that look into the conformation of adsorbed proteins so far. Most of them were performed by ATR.30-33,60-63 The only detailed structure analysis of adsorbed protein obtained by transmission spectroscopy was given by Baron et al.64,65 They looked at the interaction of BSA and R-chymotrypsin with montmorillonite, an electronegative phyllosilicate. The conformation of BSA known as “soft” protein was strongly changed while the structure of the “hard” protein R-chymotrypsin did show smaller modifications. Here we present the first analysis of the secondary structure of adsorbed proteins to SPB by infrared spectroscopy in transmission mode. The secondary structure information was obtained by a commercial setup based on a similar PLS analysis as described by Dousseau and Pe´zolet.51 EXPERIMENTAL SECTION Materials. BSA, bovine β-lactoglobulin (BLG), and bovine pancreatic ribonuclease A (RNase A) were purchased from Sigma and used without further purification. Characteristic parameters of the proteins are given in Table 1. The synthesis of the spherical polyelectrolyte brushes was performed by photoemulsion polymerization along the lines given recently.33,34 Four core shell latexes have been used in the present study: latex KpS14 and KpS15 bearing chains of poly(acrylic acid) and latexes KpSS1 and KpSS2 bearing chains of poly(styrenesulfonic acid) (see Figure 1). All systems were purified by exhaustive ultrafiltration against pure water. The pH of the suspensions was adjusted to 6.1 by 10 mM N-morpholinoethanesulfonic acid (MES) buffer and to 9.3 by 10 mM 2-(cyclohexylamino)ethanesulfonic acid (CHES) buffer according to the method described previously.36 A 2 mM concentration of NaN3 was added to the buffer solutions to avoid microbial growth. The characterization of the particles was done as described in refs 33-35, and the thickness of the brush layer L in the buffer solution was determined by dynamic light scattering.34,35 The grafting density (60) Geladi, P.; Kowalski, B. R. Anal. Chim. Acta 1986, 185, 1. (61) Tretinnikov, O. N.; Tamada, Y. Langmuir 2001, 17, 7406. (62) Husband, F. A.; Garrood, M. J.; Mackie, A. R.; Burnett, G. R.; Wilde, P. J. J. Agric. Food Chem. 2001, 49, 859. (63) Noinville, S.; Revault, M.; Baron, M.-H. Biopolymers 2002, 67, 323. (64) Baron, M.-H.; Revault, M.; Servagent-Noinville, S.; Abadie, J.; Quiquampoix, H. J. Colloid Interface Sci. 1999, 214, 319. (65) Servagent-Noinville, S.; Revault, M.; Quiquampoix, H.; Baron, M.-H. J. Colloid Interface Sci. 2000, 221, 273.
Table 2. Characteristic Parameters of the Spherical Polyelectrolyte Brushes: Rc Radius of the Poly(styrene) Core and L Hydrodynamic Brush Layer Thickness in the Corresponding Buffer Media SPB
Rc (nm)
L (nm) (buffer)
KpS14 KpS15 KpSS1 KpSS2
50 50 56 56
54 (MES) 93 (MES) 63 (CHES) 95 (MES), 100 (CHES)
Figure 2. Schematic representation of the experiment. Solutions of BSA were prepared in buffer solutions (10 mM MES or 10 mM CHES). These solutions were mixed with solutions of the SPB in the corresponding buffer medium and equilibrated for 24 h under gentle stirring at 4 °C. The nonadsorbed protein fraction was removed by careful ultrafiltration against fresh buffer.42
of the polyelectrolyte chains on the poly(styrene) core is in the order 0.1 nm-2. The spatial dimensions of the systems are summarized in Table 2. Figure 2 shows the procedure employed here: Given amounts of protein were dissolved in aqueous 10 mM MES buffer or in the case of ribonuclease in 10 mM CHES buffer, added to the solutions of the SPB and stirred for 24 h at low temperature to avoid microbial growth. The fraction of nonadsorbed protein was removed by ultrafiltration as described recently.36 The amount of nonadsorbed protein was determined by extinction measurement of the eluate (extinction coefficient at 278 nm 44 300 M-1 cm-1 for BSA and 8910 M-1 cm-1 for RNase A), which in turn gives the amount of firmly adsorbed protein. The main fraction of the BSA-coated SPB was flushed with MES buffer containing 500 mM added NaCl in an ultrafiltration cell to release the proteins from the SPB. The proteins in the eluate were subjected to a second ultrafiltration against pure water to remove the added salt. For the conformation analysis, the freeze-dried protein was redissolved in buffer solution. The analysis of the secondary structure of the proteins was performed using a FT-IR system developed for protein analytics (Bruker Optik Confocheck) including a spectrometer (Tensor 27), a calcium fluoride flow through liquid cell (Aquaspec AS 1100 M) with 6.5-µm path length, and a highly sensitive photovoltaic Analytical Chemistry, Vol. 76, No. 10, May 15, 2004
2815
MCT detector with liquid nitrogen cooling. Resolution was set at 4 cm-1. At least two measurements of 64 scans at 20 kHz were performed and the Fourier transformed data was computed in the region of 400-4000 cm-1 and averaged. The temperature of the measurement cell was maintained at 25 °C via a cryostat (Haake DC 30-K20) which was controlled directly by the spectrometer software (OPUS). The freeze-dried proteins were dissolved in fresh buffer solution to final concentrations of 10-20 mg/mL whereas the protein-coated latex was diluted to a concentration of 0.5-2 wt % SPB particles. In the latter case, the protein concentration was in the range of 5-10 mg/mL. At concentrations of 1-2 wt % latex spheres, the pure SPB spectra were recorded. The flow-through liquid cell was filled via a 50-µL HPLC syringe. Sample and reference solutions have been filtrated through a 1-µm polyester membrane syringe filter (membraPure Membrex 25 Pet). In addition, the solution passed a 2-µm filter frit before it reached the measurement cell. This procedure made sure that the presence of dust particles or agglomerates of latex particles in the cell can be excluded. The spectrometer was sealed to minimize water vapor interferences. Two internal desiccant cartridges in the spectrometer help to keep the air dry. The beam path within the sample compartment was encapsulated but also purged with dried nitrogen to remove water vapor and carbon dioxide and to thermally insulate the sample cell and the spectrometer. In addition to this, an automatic water vapor correction was performed.51 The evaluation of the secondary structure was performed using the Quant 2 analysis in the spectrometer software OPUS. This analysis is based on a protein databank (Bruker Optik Bruker Protein Library) composed of the IR spectra of 30 reference proteins with known X-ray structure analysis.66 For the secondary structure determination, the R-helix and β-sheet content of the protein of unknown conformation is derived from all 30 protein spectra using a PLS58,59 algorithm. The Quant 2 analysis was optimized for the determination of R-helix and β-sheet contents. Cross correlation of the reference data in the spectrometer software with X-ray data gave an error in the R-helix content of 4% and accordingly 3% for β-sheet in the reported range of protein concentrations of 5-20 mg/mL. Recently Rahmelow and Hu¨bner reviewed secondary structure analysis derived from X-ray data by, for instance, the PLS algorithm and reported the quality of such an analysis as a function of the number of reference proteins used.53 RESULTS AND DISCUSSION Adsorption of BSA. In poly(acrylic acid) brushes, the charge distance along the polyelectrolyte chains depends strongly on the pH (annealed brush) while poly(styrenesulfonate) systems are fully dissociated in the whole pH range, and therefore, the charge parameter along the chains is fixed (quenched brush). In the following, the SPB were just mentioned as annealed or quenched brushes for the sake of simplicity. Previously the adsorption of BSA onto annealed brushes was reported in detail.42 The adsorption curve for BSA adsorbed onto KpS14 particles in Figure 3 represents a typical adsorption curve for the interaction of BSA molecules with annealed brushes. The (66) Tanford, C.; Hauenstein, J. D. J. Am. Chem. Soc. 1956, 78, 5287.
2816
Analytical Chemistry, Vol. 76, No. 10, May 15, 2004
Figure 3. Protein adsorption on the SPB. The amount of adsorbed protein τads per unit mass particles is plotted against the concentration of the protein left unadsorbed in solution csol. The squares refer to the adsorption of BSA to the annealed brush KpS14 in 10 mM MES buffer (pH 6.1), the triangles to RNase A adsorbed on quenched brush KpSS2 in 10 mM CHES buffer (pH 9.3), the circles to BLG adsorbed on quenched brush KpSS2, and the crosses to BLG adsorbed on the annealed brush KpS14 (both experiments carried out in 10 mM MES buffer at pH 6.1). The filled symbols represent the corresponding data points according to the samples used for the FT-IR studies. The data points are fitted by a previously described protein adsorption model, which was slightly modified.42
adsorption behavior can be described in terms of a recently published adsorption model that takes a multilayer formation of proteins in the brush layer into account.42 In the case of quenched brushes, the protein binding capacity for all investigated proteins was found to be larger than for annealed brushes. The latter point can be explained by a lower pH and lower osmotic activities in these brush layers42 and, therefore, a stronger protein fixation as macroion of the brush. In addition, hydrophobic interaction between the poly(styrenesulfonate) chains and hydrophobic parts of the protein should also been taken into account. BSA molecules interacted so strongly with quenched brushes that almost no free protein was left in solution. If, for instance, 80 mg of BSA was incubated with 100 mg of KpSS2 particles in 10 mM MES buffer, 94% of the protein was bound to the particles. Therefore, adsorption curves as a function of the nonadsorbed protein fraction could not be obtained in the osmotic limit of the brushes. This experimental finding was also stated by a desorption study: By flushing with a buffer solution of 500 mM added NaCl, only a fraction of 12% of the 688 mg of BSA bound/g of KpSS2 particles could be released while for the annealed brush KpS15 76% of the 948 mg of BSA bound/g of particles could be liberated. Adsorption of BLG. In the case of BLG, a lower adsorption degree was observed for both types of SPB (Figure 3): If, for instance, 80 mg of BLG was added to 100 mg of particles of quenched brush KpSS2, 50% of the protein was found to be adsorbed onto the microspheres after ultrafiltration. The adsorption degree for annealed brushes was, however, only in the region of a few percent. The affinity of the protein was found to be significantly lower compared to BSA, but BLG can still be bound in high amounts to quenched brushes (Table 5). Hence, it should be released from the KpSS2 particles in larger scale than BSA. In fact, a fraction of 92% of the adsorbed protein could be liberated by ultrafiltration against MES buffer containing 500 mM added NaCl.
Table 3. Secondary Structure Analysis of Different Proteins in Aqueous Buffer Solution by FT-IR Spectroscopya protein (PDB code)
R-helix (%)
β-sheet (%)
BSA (1 BM0) myoglobin (1WLA) hemoglobin (4HHB) RNase A (7RSA) lysozyme (1BB7) BLG (3BLG)
67 (67.7) 71 (73.9) 70 (67.1-68.8) 17 (17.7) 33 (34.1) 14 (12.4)
2 (0.0) 2 (0.0) 0 (0.0) 31 (31.5) 4 (6.2) 35 (36.4)
a For comparison the corresponding values from the PDB66 are given in parentheses.
Adsorption of RNase A. The interaction strength of RNase A with quenched brushes was found to be between BSA and BLG (Figure 3). The experiments were now performed in 10 mM CHES buffer at pH 9.3, which is close to the isoelectric point (IEP) of the protein of 9.6.67 Protein binding is normally strongest close to the IEP5,6,9 because protein-protein and protein-microsphere repulsion vanishes at the IEP. Therefore, a high protein binding of 72% could be found after incubating 77 mg of protein with 100 mg of SPB. This corresponds to 559 mg of RNase A/g of KpSS1 particles. The degree of adsorbed RNase A onto annealed brushes was in the range of 7% for a comparable experiment. Secondary Structure Analysis. Infrared spectroscopy enables accurate secondary structure analysis in aqueous solution if it complies with the following requirements: a highly sensitive spectrometer, low-temperature tolerances, low tolerances in the cuvette path length, and an appropriate water compensation routine.54 Despite the low difference between sample and reference spectra due to the strong water absorption, a clear protein spectra can be obtained if all these prerequisites are given. The spectrometer setup did comply with all these technical points, but for an experimental test of the accuracy of the data, we performed a secondary structure analysis for some selected proteins dissolved in buffer solution. Table 3 shows the experimentally found content of R-helix and β-sheet compared to values from X-ray crystallography.73 The values obtained by FT-IR spectroscopy were in excellent accordance to the data from PDB if protein concentrations in the range of 5-20 mg/mL were used. BSA on Annealed Brushes. Beside the prevention of total absorption by water, the small path length of the cell of 6.5 µm should also minimize light scattering of the infrared light by latex particles. Therefore, we recorded IR spectra of bare SPB dispersed in 10 mM MES buffer in a concentration range of 0.35-1.41 wt %. In comparison to the reference solution, water is missing at the location of the latex spheres. This causes a “negative” water band in the spectra as shown in Figure 4. Beside this dominant band, two signals of the microspheres at 1452 and 1493 cm-1 can be investigated without any overlay by other signals. As the band height was found to be proportional to these signals and the experiments were reproducible, we did not forsee any difficulties in performing measurements in the presence of charged stabilized microspheres in transmission mode. The next step was to measure protein-coated polyelectrolyte brushes (Figure 5). In a first experiment, we investigated three (67) The Protein Data Bank. PDB, http://www.rcsb.org/pdb/.
Figure 4. FT-IR spectra of the annealed brush KpS14 at different concentrations (boldface line 1.41 wt %, semiboldface line 0.70 wt %, thin line 0.35 wt %). The striking “negative” band between 1560 and 1720 cm-1 is caused by the difference in the water amount of sample and reference due to the volume of the microspheres. Two of the four bands caused by CdC stretching vibrations (band maximums at 1580 and 1600 cm-1) of the aromatic groups of the poly(styrene) core are overlaid by this “water” band. The other ones occur without any interference (band maximums at 1452 and 1493 cm-1), and their band height is scaling with the concentration of the microspheres (band heights: (1) 0.000 75, (2) 0.001 58, (3) 0.003 25).
Figure 5. Spectra of BSA-coated SPB KpS14 at various amounts of adsorbed protein (boldface line 1116, semiboldface 828, and thin line 577 mg of BSA/g of KpS14). The protein amide I and II bands are the two dominant signals in the spectra. As comparison a spectrum of pure SPB is shown (dashed line). All spectra are scaled to a concentration of 1 wt % SPB.
samples of an annealed brush onto which different amounts of BSA were adsorbed according to the adsorption curve in Figure 3. Due to the enormous protein binding of SPB,42 the amide I and II bands of the proteins were found to be the dominant signals in the spectra. After subtraction of the bare SPB spectra, the spectra of the adsorbed proteins could be received without any significant increase in the noise of the subtracted spectra (Figure 6). The band height and the integral of the amide I band, respectively, scales with the amount of the adsorbed protein. Figure 7 shows the spectra of BSA adsorbed on poly(styrene)/ poly(acrylic acid) brushes and after desorption from the same system in comparison to the protein before the adsorption studies. The visible deviations in the spectra as well as the corresponding values obtained for the R-helix and β-helix content (shown in Table 4) are small and in the limit of experimental error. Despite the strong interaction with the brushes, the protein retains widely its secondary structure during adsorption and it can be liberated again from the SPB without any structural change. This result is surprising on one hand because BSA is mentioned as “soft” protein, which easily undergoes conformational changes.65 On the Analytical Chemistry, Vol. 76, No. 10, May 15, 2004
2817
Table 5. Adsorption of BSA, BLG, and RNase A into Quenched Brushes: Secondary Structure Analysis of the Proteins in the Native, in the Adsorbed State, and after Desorption from a Poly(styrene)/ Poly(styrenesulfonate) Brusha
experiment
Figure 6. FT-IR spectra of adsorbed BSA in SPB KpS14 (boldface line 1116, semiboldface 828, and thin line 577 mg of BSA/g of KpS14). The spectra were obtained from the spectra shown in Figure 5 by subtraction of the pure SPB spectra. The area under the amide I band scales with the adsorbed protein amount (0.61, 0.41, 0.27).
BSA BSA adsorbed on KpSS2 BLG BLG adsorbed on KpSS2 BLG desorbed from KpSS1 RNase A RNase A adsorbed on KpSS1 RNase A adsorbed on KpSS1 RNase A adsorbed on KpSS2 RNase A adsorbed on KpSS2 RNase A adsorbed on KpSS2
adsorbed protein amt (mg/g of SPB) 688 666
559 572 485 726 869
R-helix (%), FT-IR
β-sheet (%), FT-IR
67 55 14 14 13 17 13 12 15 16 15
2 2 35 31 34 31 26 27 26 26 26
a The experiments with RNase A were performed in 10 mM CHES buffer (pH 9.3) and the others in 10 mM MES buffer (pH 6.1).
Figure 7. Comparison of BSA before (solid line, bottom), during adsorption to SPB KpS15 (center), and after desorption from the SPB (top). The spectra were drawn staggered for ease of distinction. The almost overlap of all three spectra and the difference spectra of the desorbed protein (dashed line, magnified 25 times) to the initial protein show that SPB allow immobilization of protein with nearly full preservation of the secondary structure.
Table 4. Adsorption of BSA onto Annealed Brushes in 10 m MES Buffer (pH 6.1): Secondary Structure Analysis of BSA in the Native, in the Adsorbed State at Various Amounts of Bound Protein, and on Two SPB Systems of Different Morphology (See Table 2) and after Desorption from the SPB
experiment BSA BSA adsorbed on KpS14 BSA adsorbed on KpS14 BSA adsorbed on KpS14 BSA adsorbed on KpS15 BSA adsorbed on KpS15 BSA desorbed from KpS15
adsorbed protein amt (mg/g of SPB)
R-helix (%), FT-IR
β-sheet (%), FT-IR
577 829 1116 950 948
67 60 60 64 64 61 67
2 5 5 2 1 2 2
other hand, it shows the unique properties of the SPB as protein carrier. BSA on Quenched Brushes. The interaction of proteins with quenched brushes, that is to say, systems where the brushes are made up by a strong polyelectrolyte, were found to be much stronger than in the case of annealed brushes as mentioned above. An analysis of BSA desorbed from quenched brushes was not performed because of the low amount of releasable protein. 2818 Analytical Chemistry, Vol. 76, No. 10, May 15, 2004
The R-helix content of BSA adsorbed to the quenched brush KpSS2 was found to be 55% (Table 5). This means a decrease of 12% of helix structure compared to the native protein whereas the β-sheet content did not change. The decrease in the R-helix content is outside the limit of error of (4%. Therefore, the difference is more strongly pronounced than in the case of annealed brushes where the decrease was between 3 and 7% and still in the limit of error. Also, the degree of releasable protein adsorbed on quenched brushes was significantly lower than for annealed brushes as mentioned above. Hence, both observations clearly indicate a stronger interaction of the protein with quenched brushes. BLG Adsorbed on Quenched Brushes. Table 5 shows also the results of the secondary structure analysis for the following experimental setup: BLG before adsorption, BLG adsorbed on KpSS2, and BLG released from KpSS2 particles. The R-helix content remained constant in all three cases while the β-sheet content was 5% lower during adsorption. In the limit of experimental error of (3% for β-sheets, one can state that almost no changes in secondary structure of BLG occur during and after interaction with quenched brushes. RNase A Adsorbed on Quenched Brushes. In further experiments, the conformation of RNase A adsorbed on the quenched brushes KpSS1 and KpSS2 was investigated. The amount of at least 485 mg of protein/g of latex particles was found to be sufficient for secondary structure whereas analyses with less than 400 mg of protein/g of microspheres did not lead to accurate secondary structure determinations. In this case, the errors in the subtraction of the signal of the microspheres were not negligible compared to the difference spectra of protein during and before adsorption in which the change in conformation is expressed. Therefore, analyses for BLG and RNase A adsorbed on annealed brushes were not performed due to the low adsorption degrees in these cases. Therefore, we restricted the study to the analysis of RNase A adsorbed on quenched brushes. RNase A was selected in this study for two reasons: First, its interaction strength with quenched brushes was found to be strong enough. Second, RNase A is
known as a “hard” protein, which has a rather rigid conformation in comparison to BSA.65 Table 5 shows a loss of 1-5% of R-helix and 5% of β-sheet structure for RNase A during adsorption. This small decrease is still within the limit of experimental error. However, there is a general tendency in all adsorption experiments of a slight loss in secondary structure during adsorption but only in the range of a few percent. On the other hand, the experiments of protein that was desorbed from the particles have shown that even this small change in secondary structure was reconstituted (Table 5). All experiments gave a clear indication that the secondary structure could be almost or widely preserved in the adsorbed state even for BSA as a “soft” protein. This is a prerequisite for immobilized enzymes to retain their biologic activity. It should be noted that preservation in the secondary structure does not automatically mean preservation in enzymatic activity. Recent experimental studies 44,45 showed that the enzymatic activity of glucoamylase adsorbed on either annealed or quenched brushes is retained. These two observations, namely, the full preservation of secondary structure and of enzymatic activity, indicate the potential of SPB as new systems for protein immobilization. In addition, the interaction of the proteins with the SPB can be adjusted by either the brush morphology (as shown here) or by the ionic strength (as shown recently42). If the interaction strength was lowered, the protein could be widely released from the SPB in a simple ultrafiltration by raising the ionic strength. CONCLUSION The present study has shown that infrared spectroscopy in transmission mode is a well-suited tool to observe the conforma-
tion of a protein even in the presence of latex particles. The method allows one to investigate the secondary structure of proteins adsorbed to microspheres if a sufficient amount of protein could be adsorbed. SPB were well-suited for such an analysis because of their high protein-binding capacity. The secondary structure of different proteins adsorbed on SPB of different morphologies was found to be nearly fully retained during adsorption. Together with the previously found preservation of activity of enzymes immobilized to such systems, the capability of SPB could be shown. Additionally, the interaction strength of the proteins with the SPB can be controlled by the ionic strength and the proteins can be released if the interaction strength during adsorption was kept moderate. SPB can be therefore charged and decharged with proteins. Applications of SPB as “nanoreactors” with immobilized enzymes are at hand.44,45 ACKNOWLEDGMENT Financial support by the Roche Diagnostics Co., the Bundesministerium fu¨r Forschung und Technologie, and the Deutsche Forschungsgemeinschaft is gratefully acknowledged. The authors thank M. Luft from Bruker Optik GmbH, 76275 Ettlingen (Germany) for his kind introduction in the spectrometer setup.
Received for review December 12, 2003. Accepted March 5, 2004. AC0354692
Analytical Chemistry, Vol. 76, No. 10, May 15, 2004
2819
