Langmuir 2005, 21, 11903-11906
11903
Proteolytic Cleavage Reveals Interaction Patterns between Silica Nanoparticles and Two Variants of Human Carbonic Anhydrase Martin Lundqvist,† Cecilia Andresen,† Sara Christensson,† Sara Johansson,† Martin Karlsson,‡ Klas Broo,§ and Bengt-Harald Jonsson*,† Divisions of Molecular Biotechnology and Biochemistry, IFM, Linko¨ ping University, SE-58183 Linko¨ ping, Sweden, and Department of Occupational and Environmental Medicine, Sahlgrenska Academy at Go¨ teborg University, Go¨ teborg, Sweden Received February 23, 2005. In Final Form: July 5, 2005 To characterize the sites on the protein surface that are involved in the adsorption to silica nanoparticles and the subsequent rearrangements of the protein/nanoparticle interaction, a novel approach has been used. After incubation of protein with silica nanoparticles for 2 or 16 h, the protein was cleaved with trypsin and the peptide fragments were analyzed with mass spectrometry. The nanoparticle surface area was in 16-fold excess over available protein surface to minimize the probability that the initial binding would be affected by other protein molecules. When the fragment patterns obtained in the presence and absence of silica nanoparticles were compared, we were able to characterize the protein fragments that interact with the surface. This approach has allowed us to identify the initial binding sites on the protein structure and the rearrangement of the binding sites that occur upon prolonged incubation with the surface.
Introduction Many proteins interact with and/or adsorb to almost any material that is placed in contact with them. This has implications in diverse fields, including biology, medicine, biotechnology, and food processing.1,2 The adsorption is often causing considerable problems when the material is used in medical applications or in devices that monitor various biological activities. However, there are also examples of applications where adsorption is a desired characteristic of a protein, for instance, in adsorption chromatography, enzyme-linked immunoassays, protein delivery systems, and biosensors. For any rational design of such applications, knowledge of the adsorption orientation is essential, because it strongly affects factors such as access to the active site in enzymes and the recognition of epitopes by antibodies. Several techniques such as ellipsometry, surface plasmon resonance (SPR), quartz crystal microbalance (QCM), scanning angel reflectometry (SAR), and optical waveguide spectroscopy (OWS) have been used with the aim to determine the preferred orientation of adsorbed proteins. These techniques give information regarding the total mass of the adsorbed protein and the thickness of the protein layer. This information can (for example) be used to calculate if the protein adsorbs in an end-on or side-on orientation.3 Detailed characterization of the protein architecture and layer thickness and the dependence on the protein shape and protein concentration have been performed by Ramsden and collaborators using OWS4,5 * To whom correspondence should be addressed. Tel: +46-13-288935. Fax: +46-13-122587. E-mail:
[email protected]. † Division of Molecular Biotechnology, Linko ¨ ping University. ‡ Division of Biochemistry, Linko ¨ ping University. § Go ¨ teborg University. (1) Hlady, V.; Buijs, J. Curr. Opin. Biotechnol. 1996, 7, 72. (2) Nakanishi, K.; Sakiyama, T.; Imamura, K. J. Biosci. Bioeng. 2001, 91, 233. (3) Schaaf, P.; Dejardin, P.; Johner, A.; Schmitt, A. Langmuir 1987, 3, 1128. (4) Guemouri, L.; Ogier, J.; Ramsden, J. J. J. Chem. Phys. 1998, 109, 3265.
and by Schaaf and co-workers using SAR.3,6 It is obvious that the results that are obtained from the use of the mentioned techniques are dependent upon the conformational changes of the adsorbed protein,7 exchange processes,8 lateral interactions,9 reorientation,10-13 and multilayer formation.11,14 Hydrogen exchanges at peptide NH,which can be used to monitor protein compactness and amino acid exposure to the solvent, have also been used to characterize adsorption orientation,15,16 and the results obtained indicate that the orientation of the adsorbed proteins is nonrandom. In yet another approach, site-directed fluorescently labeled variants of human carbonic anhydrase II were used to characterize the initial binding site of the protein when it adsorbs to silica nanoparticles.17 In this study, we used two variants of human carbonic anhydrase II (HCAIIpwt and HCAIIS56C) that have identical native structure (except for the change of an -OH group to a -SH group in the side chain of position 56) but differ in stability. The interactions between HCAII and HCAIIS56C and silica nanoparticles have previously been investigated using circular dichroism (CD) and fluores(5) Guemouri, L.; Ogier, J.; Zekhnini, Z.; Ramsden, J. J. J. Chem. Phys. 2000, 113, 8183. (6) Schaaf, P.; Dejardin, P.; Schmitt, A. Langmuir 1987, 3, 1131. (7) Karlsson, M.; Mårtensson, L. G.; Jonsson, B. H.; Carlsson, U. Langmuir 2000, 16, 8470. (8) Vermonden, T.; Giacomelli, C. E.; Norde, W. Langmuir 2001, 17, 3734. (9) Moulin, A. M.; O’Shea, S. J.; Badley, R. A.; Doyle, P.; Welland, M. E. Langmuir 1999, 15, 8776. (10) Wertz, C. F.; Santore, M. M. Langmuir 2002, 18, 1190. (11) Lee, J. E.; Saavedra, S. S. Langmuir 1996, 12, 4025. (12) Daly, S. M.; Przybycien, T. M.; Tilton, R. D. Abstr. Pap. Am. Chem. Soc. 2003, 226, U387. (13) Ramsden, J. J. Phys. Rev. Lett. 1993, 71, 295. (14) Ho¨o¨k, F.; Rodahl, M.; Kasemo, B.; Brzezinski, P. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 12271. (15) Aizawa, T.; Koganesawa, N.; Kamakura, A.; Masaki, K.; Matsuura, A.; Nagadome, H.; Terada, Y.; Kawano, K.; Nitta, K. FEBS Lett. 1998, 422, 175. (16) Larsericsdotter, H.; Oscarsson, S.; Buijs, J. J. Colloid Interface Sci. 2004, 276, 261. (17) Karlsson, M.; Carlsson, U. Biophys. J. 2005, 88, 3536.
10.1021/la050477u CCC: $30.25 © 2005 American Chemical Society Published on Web 11/03/2005
11904
Langmuir, Vol. 21, No. 25, 2005
Lundqvist et al.
cence.7 The cited study indicated that the difference in their stability is correlated to the rate of conformational changes that occur upon the interaction with nanoparticles. The present study is focused on identifying and characterizing the segments of the protein that interact at the nanoparticle interface shortly after mixing with the nanoparticles. We have chosen to have an approximately 16-fold excess of total silica surface over an available protein surface (calculated as the maximal projected surface of the protein ellipsoid onto a plan) to minimize effects from lateral interactions and multilayer formation. A novel approach to investigating these interactions involves the identification of interacting peptide segments by mass spectrometry. Therefore, we treated the proteinparticle samples with trypsin and then analyzed the proteolytic cleavage pattern using MALDI-TOF mass spectrometry. To evaluate if the structure of the proteins rearranges at the silica surface after initial binding, the samples were also incubated overnight with nanoparticles before the proteolytic cleavage and MALDI-TOF MS analysis. Analyses of the peptide fragments by use of MALDI-TOF MS provide information on both early interaction sites and later rearrangement of the protein structure at the nanoparticle surfaces. Materials and Methods Proteins. The pseudo-wild type (pwt) of HCAII, in which cysteine 206 has been changed to a serine, was produced and purified as described in Freskgård et al.18 HCAIIS56C, in which serine 56 of the HCAIIpwt construct has been mutated to a cysteine, was produced and purified as described in Karlsson et al.7 Silica Particles. The colloidal, negatively charged silica particles used in this study were kindly provided by EKAChemicals, Stenungsund, Sweden. Before use, the particles were extensively dialyzed against sample buffer (20 mM Trs and 20 mM NaCl at pH 8.4). The particles are stable in solutions at pH above 8 and in moderate salt concentrations.19 The stock solution contained 5.09E + 17 particles/mL with an average diameter of 9 nm. MALDI-TOF MS. The stock protein sample solutions had a concentration of 40 µM for HCAIIpwt and 20 µM for HCAIIS56C. In all cases, the buffer was 20 mM Tris and 20 mM NaCl with a pH of 8.4. In the samples to which silica particles were added, the protein/particles molar ratio was 2:1. To 70 µL of the protein stock solutions, 35 µL of a 4.8E-3 mg/mL trypsin solution was added (final trypsin concentration, 1.6 µg/mL). A total of 15 µL of the mixture was added to 15 µL of 1% acetic acid at specific time points to quench the proteolytic reaction. The particles were removed from the solution by centrifugation for 30 min at 13 000 rpm, and then 15 µL of the supernatant was transferred to a new tube containing 50 µL of 2-propanol (to reduce the interaction between hydrophobic peptide fragments and the sample tube walls) and again centrifuged for 30 min at 13 000 rpm. A total of 30 µL of the second supernatant was mixed with 35 µL of Milli Q water and analyzed with MALDI-TOF MS using R-cyano4-hydroxycinnamic acid as the matrix (the samples without particles were treated in the same way). All experiments were performed twice and on different occasions, and the results were essentially identical. CD. CD spectra were recorded using a CD6 spectrodichrograph (Jobin-Yvon Instruments SA, Longjumeau, France) with constant N2 flushing. The instrument was calibrated with an aqueous solution of d10-(+)-camphorsulfonic acid. The sample was kept at 30 °C throughout the experiment (0-11 days). The spectra in the near-UV region (240-320 nm) were recorded by scanning 0.2 mM HCAIIpwt in 100 mM Tris buffer (pH 8.4) with a 2:1 molar ratio of protein and 9 nm silica particles in a 0.05 cm quartz cell. The data were collected at 0.5 nm intervals with an (18) Freskgård, P. O.; Carlsson, U.; Mårtensson, L. G.; Jonsson, B. H. FEBS Lett. 1991, 289, 117. (19) www.ekachemicals.se.
Figure 1. Proteolytic cleavage patterns for the two variants of HCAII after 5 min of cleavage. The colored bars represent residues that are detected in fragments by the MALDI-TOF MS. (a and b) Patterns obtained for HCAIIpwt and HCAIIS56C, respectively. Blue, green, and red represent samples incubated without particles, samples incubated for a period of 2 h with particles, and samples incubated overnight with particles, respectively. The arrows indicate missing fragments that are discussed in the text. integration time of 2 s for the region between 240 and 300 and 2 nm intervals between 300 and 320 nm. Each spectrum represented an average of two consecutive scans, and before summation, the two separate scans were compared to detect possible alterations in the sample during the scan period. The protein spectra were corrected by subtracting the spectrum of a reference solution that lacked the protein but was otherwise identical. The ellipticity is reported as the mean residue molar ellipticity ([θ], in degrees cm2 dmol-1) according to eq 1
[θ] ) [θ]obsmrw/10lc where [θ]obs is the ellipticity (in degrees), mrw is the mean residue molecular weight (HCAIIpwt has a molecular weight of 29 300 and 259 amino acid residues), c is the protein concentration (in g/mL), and l is the optical path length of the cell (in centimeters).
Results and Discussion Two different variants of human carbonic anhydrase were proteolytically cleaved with trypsin 2 h after the addition of silica nanoparticles to locate areas of the protein that interact with the nanoparticles. Samples where protein and nanoparticles were incubated overnight before proteolytic cleavage were also analyzed to gain information on rearrangements of the protein leading to new interaction sites. Protein samples without silica nanoparticles were analyzed in the same way as reference samples. The peptides detected by MALDI-TOF MS after cleavage for the two variants of HCAII are shown in Figure 1. The observed patterns show distinct differences between bound and free protein for the two variants of HCAII. Analysis of these observed patterns is based on the following considerations. First, the trypsin cleavage was
Silica Nanoparticles and Two Variants of Human Carbonic Anhydrase
Figure 2. CD spectra for HCAIIpwt with a 2:1 ratio of protein and 9 nm silica nanoparticles. The thick black line represents the spectrum for HCAIIpwt without particles, and the thin gray lines represent spectra obtained following incubation times with the nanoparticles 1, 2, 3, 4, 6, 8, and 11 days.
only allowed to proceed for a short time (5 min) to maximize cleavage at positions that are exposed to solution and minimize cleavage that may result from rearrangements in the structure that are caused by the cleavage itself (secondary rearrangements). Second, the cleavage will give rise to peptides that are “free” in solution and to peptides that remain firmly attached to the nanoparticles. Third, the peptides that are “free” in solution are detected with MALDI-TOF MS. Fourth, the differences between peptides detected in samples with nanoparticles and samples without nanoparticles indicate fragments that interact strongly with the particles. The discussion below considers the implications of the differences in the interaction patterns obtained with the two different proteins. For the region between residues 181-225, no conclusions can be drawn regarding its involvement in binding because this fragment was not detected under any conditions for cleavage (see Figure 1). The reason for the difficulty to observe the fragment is probably the low solubility of this fragment leading to self-association into very stable high molecular aggregates, which has earlier been shown by Henderson et al.20
Langmuir, Vol. 21, No. 25, 2005 11905
HCAIIpwt. Earlier, ellipsometry,21 gel permeation,7 and NMR22 studies of HCAIIpwt have shown that almost all of this protein is bound to the particles (under the experimental condition with an excess of available surface area) and there is very little unbound protein at equilibrium. In addition, the CD measurements in the near-UV region (see Figure 2) obtained in the present study show that the binding of HCAIIpwt affects the tertiary structure in a timedependent manner; i.e., the effects on the native structure become more extensive with time. The structural rearrangements that are reported by the CD spectra have earlier been discussed in detail by us for several variants of human carbonic anhydrase (HCA).7,21,23 The results presented in Figure 1a (arrows 1 and 2) show that peptide fragments from the N and C terminals are missing in the sample that was incubated for 2 h with nanoparticles, indicating that these peptides are strongly bound to the nanoparticles. Thus, the regions around the N terminal and close to the C terminal (see Figure 3a) are important for initial binding of the protein to the nanoparticles. These results are in accordance with results from a study by Karlsson et al.17 on HCAII labeled with fluorescence probes in different positions. The cited study showed that the interaction had strong effects on the fluorescence from probes at residues 10 and 37 (indicated in Figure 3a) but not from probes at other parts of the structure. The N- and C-terminal peptides give these regions of the protein an excess of positive charges, indicating that the initial interaction with the strongly negatively charged silica nanoparticles is dominated by electrostatic forces. Analysis of HCAIIpwt that had been incubated with particles overnight shows that the peptide fragment 132168 was bound to the nanoparticles (arrow 3 in Figure 1a), while both the N- and C-terminal peptides no longer bound to the particles. Fragment 132-168 contains a helix, a loop, and β-strand 6 (going from left to right in Figure 3), which in a native conformation penetrate the center of the molecule (see Figure 3b). The fragment has five positively charged side chains, three negatively charged
Figure 3. Fragments that participate in the binding to the surface of the nanoparticles are highlighted as black spheres. The C and N termini participate in binding after 2 h. The fragment 132-168 participates in binding after rearrangements of the structure that have been allowed to proceed overnight. The two residues, His 10 and Thr 37, which M. Karlsson et al.17 have shown to participate in the initial phase of the binding, are also highlighted in gray. The figure was prepared using MolMol.24
11906
Langmuir, Vol. 21, No. 25, 2005
Lundqvist et al.
Figure 4. Peptide fragments in HCAIIS56C that bind to the silica nanoparticle surface, showing residues 1-39 in dark gray, residues 89-131 in black, and residues 148-170 in light gray, from different angles. The figure was prepared using MolMol.24
side chains, and five residues with hydrogen-bonding capacity. Thus, the strength of its binding to the silica particles is likely to depend upon its configuration and to increase following rearrangement. These findings suggest that prolonged incubation with nanoparticles causes structural rearrangements of the protein that expose otherwise hidden parts of it, which can then efficiently interact with the silica surface. These results are compatible with the CD observations that the nanoparticles cause more pronounced structural alterations with time (Figure 2). HCAIIS56C. The only difference between HCAIIpwt and HCAIIS56C is a rather conservative replacement of serine by cysteine at position 56, which leads to a destabilization of the conformational stability by approximately 2.2 kcal/ mol.7 Thus, a comparison of this variant with HCAIIpwt with respect to surface binding properties should reveal differences that are entirely due to the difference in stability. It has been shown that all HCAIIS56C adsorbs almost instantly to the silica nanoparticles (under the experimental condition with an excess of available surface area) and that the structural changes observed in its structure after the adsorption occur at a significantly higher rate than for HCAIIpwt.7 The results indicate that the proteolytic cleavage pattern is more complex for HCAIIS56C than for HCAIIpwt. The N termini of both HCAIIS56C and HCAIIpwt appear to be part of an initial interaction area with the nanoparticles. (20) Henderson, L. E.; Henriksson, D.; Nyman, P. O. J. Biol. Chem. 1976, 251, 5457. (21) Billsten, P.; Freskgard, P. O.; Carlsson, U.; Jonsson, B. H.; Elwing, H. Febs Lett. 1997, 402, 67. (22) Lundqvist, M.; Sethson, I.; Jonsson, B. H. Langmuir 2005, 21, 5974. (23) Lundqvist, M.; Sethson, I.; Jonsson, B. H. Langmuir 2004, 20, 10639. (24) Koradi, R.; Billeter, M.; Wuthrich, K. J. Mol. Graphics 1996, 14, 51.
However, for HCAIIS56C, fragment 89-131 (arrow 2 in Figure 1b) also binds to the surface in an early phase. As shown in Figure 4, fragment 89-131 contains β-strands 4 and 5 and the loop connecting them. The loop, residues 99-115, is part of a possible interaction area (the loop and the N-terminal region) between the protein and the particle (Figure 4). Therefore, it seems that HCAIIS56C rearranges rapidly in the early stages of binding to the nanoparticles, allowing the 99-115 loop to be part of a binding site. Interestingly, this loop contains a cluster of three consecutive lysine residues. The observation indicates that the C terminal of HCAIIS56C, unlike that of HCAIIpwt, does not participate in binding to the nanoparticles, indicating that parts of the C-terminal interacts in an early phase of the binding, and after 2 h of incubation, HCAIIS56C has rearranged to a point where the C terminal is no longer part of the interaction area. The results of proteolytic cleavage after prolonged incubation indicate that the structure of HCAIIS56C subsequently undergoes further rearrangement, allowing interactions with the central region, residues 148-170, similar to the overnight rearrangement observed for HCAIIpwt. Interestingly, the interaction with the N terminal of HCAIIS56C is maintained after prolonged incubation in contrast to the situation with HCAIIpwt. Thus, it seems that the low intrinsic stability of HCAIIS56C allows more extensive rearrangement of its structure and thus maximization of the binding area. Acknowledgment. This work was supported by grants from the Swedish National Science Research Council to B.-H. J. (K5104-5999), and K. B. is indebted to the Knut and Alice Wallenberg Foundation for financial support. LA050477U
