Dynamic Force Spectroscopy of the Specific Interaction between the

Feb 2, 2007 - Michael Odorico , Jean-Marie Teulon , Yannick Delcuze , Shu-wen Chen ... S.C. Baos , D.B. Phillips , L. Wildling , T.J. McMaster , M. Be...
0 downloads 0 Views 192KB Size
Download PDF
2668

Langmuir 2007, 23, 2668-2673

Dynamic Force Spectroscopy of the Specific Interaction between the PDZ Domain and Its Recognition Peptides Tei Maki,† Satoru Kidoaki,*,‡ Kengo Usui,† Harukazu Suzuki,§ Masayoshi Ito,§ Fuyu Ito,† Yoshihide Hayashizaki,†,§ and Takehisa Matsuda*,⊥ Core Research for EVolutional Science and Technology (CREST), Japan Science and Technology Agency, 4-1-8 Honcho, Saitama 332-0012, Japan, DiVision of Biomolecular Chemistry, Institute for Materials Science and Engineering, Kyushu UniVersity, Fukuoka 812-8581, Japan, Laboratory for Genome Exploration Research Group, RIKEN Genomic Sciences Center, RIKEN Yokohama Institute, 1-7-22 Suehiro-cho, Kanagawa 230-0045, Japan, and Genome Biotechnology Laboratory, Kanazawa Institute of Technology, Ishikawa 924-0838, Japan ReceiVed September 15, 2006. In Final Form: December 18, 2006 To characterize the molecular basis of specific interactions of PDZ proteins, dynamic force spectroscopy (DFS) for the PDZ protein Tax-interacting protein-1 (TIP-1) and its recognition peptide (PDZ-pep) derived from β-catenin was performed using an atomic force microscope (AFM), together with measurement of thermodynamic and kinetic parameters using surface plasmon resonance (SPR). The unbinding force of this pair was measured under different conditions of AFM tip-retraction velocity. The relationship between the unbinding force and the logarithmic forceloading rate, that is, the dynamic force spectrum, exhibited two different rate regimes, for each of which the forces increased linearly with the force-loading rate. On the basis of the theoretical treatment of the Bell-Evans model, the positions of two different activation barriers in the reaction coordinate and dissociation rate constants in each barrier were evaluated from slopes and x-intercepts of the two linear regimes (first barrier: 0.04 nm and 1.10 × 10 s-1; second barrier: 0.21 nm and 2.77 × 10-2 s-1, respectively). Although two-step unbinding kinetics between TIP-1 and PDZ-pep was suggested from the DFS analysis, SPR results showed single-step dissociation kinetics with a rate constant of 2.89 × 10-1 s-1. Different shapes of the free energy profile of the unbinding process were deduced from each result of DFS and SPR. The reason for such topographic differences in the energy landscape is discussed in relation to the differences in the pathways of forced unbinding and spontaneous dissociation.

Introduction Specific interactions between ligands and receptors, which play crucial and extensive roles in various biological aspects depending on molecular recognitions such as immunological responses, signaling, gene expression, molecular assembly, or cell adhesion, have remained a central theme in molecular biology.1,2 In most cases, ligand-receptor bindings involve multiple noncovalent interaction sites and different types of interaction forces (i.e., van der Waals, hydrogen bonding, ionic, and hydrophobic) within the binding pocket, which determine the intrinsic characteristics of specific binding in combination. Analyses of the equilibrium constant and the association and dissociation rate constants have been employed to characterize the physicochemical properties of ligand-receptor interactions, although these values are generally convoluted and coarse-grained information of such complex contents of interaction character. Acquisition of more detailed physicochemical information related to ligand-receptor interaction forces is required for clarifying * Corresponding author. † Japan Science and Technology Agency. ‡ Kyushu University. § RIKEN Yokohama Institute. ⊥ Kanazawa Institute of Technology. (1) Springer, T. A. Adhesion receptors of the immune systems. Nature 1990, 346, 425-434. (2) Frauenfelder, H.; Sligar, S. G.; Wolynes, P. G. The energy landscapes and motions of proteins. Science 1991, 13, 1598-1603. (3) Florin, E. L.; Moy, V. T.; Gaub, H. E. Adhesion forces between individual ligand-receptor pairs. Science 1994, 264, 415-417. (4) Lee, G. U.; Chrisey, L. A.; Colton, R. J. Direct measurement of the forces between complementary strands of DNA. Science 1994, 266, 771-773. (5) Fritz, J.; Katopodis, A. G.; Kolbinger, F.; Anselmetti, D. Force-mediated kinetics of single P-selectin/ligand complexes observed by atomic force microscopy. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 12283-12288.

the molecular basis of the biochemical interactions as well as the thermodynamic and kinetic parameters. For this task, direct intermolecular force measurement techniques including atomic force microscopy (AFM),3-6 micropipette suction,7-9 optical tweezers,10 surface force apparatus,11,12 and magnetic torsion devices13 have been developed and applied for the characterization of the dynamic response of individual ligand-receptor complexes to external mechanical forces. In this decade, dynamic force spectroscopy (DFS) based on the Bell-Evans theoretical framework9,14-16 has become a powerful analytical strategy for exploring the energy landscape (6) Hinterdorfer, P.; Baumgartner, W.; Gruber, H. J.; Schilcher, K.; Schindler, H. Detection and localization of individual antibody-antigen recognition events by atomic force microscopy. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 3477-3481. (7) Allen, S.; Davies, J.; Davies, M. C.; Dawkes, A. C.; Roberts, C. J., Tendler, S. J.; Williams, P. M. The influence of epitope availability on atomic-force microscope studies of antigen-antibody interactions. Biochem. J. 1999, 341, 173-178. (8) Shao, J. Y.; Hochmuth, R. M. Mechanical anchoring strength of L-selectin, β2 integrins, and CD45 to neutrophil cytoskeleton and membrane. Biophys. J. 1999, 77, 587-596. (9) Evans, E.; Leung, A.; Hammer, D.; Simon, S. Chemically distinct transition states govern rapid dissociation of single L-selectin bonds under force, Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 3784-3789. (10) Kuo, S. C.; Sheetz, M. P. Force of single kinesin molecules measured with optical tweezers. Science 1993, 260, 232-234. (11) Israelachvili, J. N. Thin film studies using multiple-beam interferometry. J. Colloid Interface Sci. 1973, 44, 259-272. (12) Helm, C. A.; Knoll, W.; Israelachvili, J. N. Measurement of ligandreceptor interactions. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 8169-8173. (13) Wang, N.; Butler, J. P.; Ingber, D. E. Mechanotransduction across the cell surface and through the cytoskeleton. Science 1993, 260, 1124-1127. (14) Bell, G. I. Models for the specific adhesion of cells to cells. Science 1978, 200, 618-627. (15) Evans, E.; Ritchie, K. Dynamic strength of molecular adhesion bonds. Biophys. J. 1997, 72, 1541-1555. (16) Evans, E. Probing the relation between force-lifetime and chemistry in single molecular bonds. Annu. ReV. Biophys. Biomol. Struct. 2001, 30, 105-128.

10.1021/la0627011 CCC: $37.00 © 2007 American Chemical Society Published on Web 02/02/2007

Specific Interaction between TIP-1 and PDZ-pep

of ligand-receptor unbinding. In precedent studies, DFS has been used to determine the dynamic strength and energy landscape of biotin/avidin,17,18 integrin R4β1/vascular cell adhesion molecule-1 (VCAM-1),19 leukocyte function-associated antigen-1 (LFA-1)/intercellular adhesion molecule-1 (ICAM-1),20 glycoprotein Ib-IX/von Willebrand factor,21 P-selectin/ligand,5 Mucin1/ antibody,22 and so forth. In the present study, we applied DFS to characterize the molecular basis of the PDZ protein interactions. The PDZ (PSD95/DLG/ZO-1 homology) domains are small protein-protein recognition modules that bind well-defined C-terminal residues in a sequence-specific manner.23 Several PDZ domain-containing proteins localize to sites of cell-cell contacts and are involved in the assembly of diverse signaling complexes.24 More and more PDZ proteins have been shown to shuttle between these signaling complexes and the nucleus, thereby exerting transcription regulatory functions.25-27 The unbinding force of the PDZ protein Tax-interacting protein-1 (TIP-1) and its recognition peptide (PDZ-pep) derived from β-catenin was measured under different conditions of AFM tip retracting velocity, which provides the relationship between the unbinding force and the logarithmic force-loading rate, that is, the dynamic force spectrum. The dynamic force spectrum was analyzed on the basis of the Bell-Evans model. The positions of the activation barriers in the reaction coordinate and dissociation rate constants in each barrier were evaluated from the slopes and x-intercepts of the linear regimes in the dynamic force spectrum, which suggested two-step unbinding kinetics between TIP-1 and PDZ-pep. Also measuring the thermodynamic and kinetic parameters with surface plasmon resonance (SPR), the energy profile of the unbinding process deduced from DFS and SPR were mutually compared, and the reasons for their differences are discussed. Materials and Methods Protein and Peptide Preparation. The PDZ domain murine TIP-1 was fused with the glutathione-S-transferase (GST) and hexahistidine (His)6 tags at the N-terminal and C-terminal ends, respectively. This fusion protein (GST-TIP-1-His) was expressed using an expression (17) Yuan, C.; Chen, A.; Kolb, P.; Moy, V. T. Energy landscape of streptavidinbiotin complexes measured using atomic force microscopy. Biochemistry 2000, 39, 10219-10223. (18) Paris, de R.; Strunz, T.; Oroszlan, K.; Guntherodt, H.-J.; Hegner, M. Force spectroscopy and dynamics of the biotin-avidin bond studied by scanning force microscopy. Single Mol. 2000, 1, 285-290. (19) Zhang, X.; Craig, S. E.; Kirby, H.; Humphries, M. J.; Moy, V. T. Molecular basis for the dynamic strength of the integrin R4β1/VCAM-1 interaction. Biophys. J. 2004, 87, 3470-3478. (20) Zhang, X.; Wojcikiewicz, E.; Moy, V. T. Force spectroscopy of the leukocyte function-associated antigen-1/intercellular adhesion molecule-1 interaction. Biophys. J. 2002, 83, 2270-2279. (21) Aya, M.; Kolomeisky, A. B.; Romo, G. M.; Cruz, M. A.; Lopez, J. A. Dynamic force spectroscopy of glycoprotein Ib-IX and von Willebrand factor. Biophys. J. 2005, 88, 4391-4401. (22) Sulchek, T.; Friddle, R. W.; Landy, K.; Lau, E. Y.; Albrecht, H.; Ratto. T. V.; DeNardo, S. J.; Colvin, M. E.; Noy, A. Dynamic force spectroscopy of parallel individual Mucin1-antibody bonds. Proc. Natl. Acad, Sci, U.S.A. 2005, 102, 16638-16643. (23) Hung, A. Y.; Sheng, M. PDZ domains: structural modules for protein complex assembly. J. Biol. Chem. 2002, 277, 5699-5702. (24) Harris, B. Z.; Lim, W. A. Mechanism and role of PDZ domains in signaling complex assembly. J. Cell Sci. 2001, 144, 3219-3231. (25) Hsueh, Y. P.; Wang, T. F.; Yang, F. C.; Sheng, M. Nuclear translocation and transcription regulation by the membrane-associated guanylate kinase CASK/LIN-2. Nature 2000, 404, 298-302. (26) Riefler, G. M.; Firestein, B. L. Binding of neuronal nitric-oxide synthase (nNOS) to carboxyl-terminal-binding protein (CtBP) changes the localization of CtBP from the nucleus to the cytosol. A novel function for targeting by the PDZ domain of nNOS. J. Biol. Chem. 2001, 276, 48262-48268. (27) Traweger, A.; Fuchs, R.; Krizbai, I. A.; Weigner, T. M.; Bauer, H. C.; Bauer, H. The tight junction protein ZO-2 localizes to the nucleus and interacts with the heterogeneous nuclear ribonucleoprotein scaffold attachment factor-B. J. Biol. Chem. 2003, 278, 2692-2700.

Langmuir, Vol. 23, No. 5, 2007 2669 Scheme 1. Schematic Representation of the Experimental Setup for the Unbinding Force Measurement

vector controlled with a Tac-promoter in the E. coli strain BL21CodonPlus (DE3)-RIL (Stratagene, La Jolla, CA) and purified using two-step affinity purification followed by a Ni2+-chelate affinity column (HisTrap HP; Amersham Biosciences Corp.) and a glutathione-agarose column (GSTrap FF; Amersham Biosciences Corp.). The GST-removed protein (TIP-1-His) was prepared using the in-column cleavage with a PreScission Protease (Amersham Biosciences Corp., Piscataway, NJ) GST-TIP1-His bound to GSTrap FF column. These fusion proteins were further purified by sizeexclusion chromatography on a HiLoad 16/60 Superdex 200 pg (Amersham Biosciences Corp.) with 10 mM HEPES, pH 7.4, and 150 mM NaCl. The PDZ domain recognition peptide (CQLAWFDTDL; italicized part ) PDZ-domain recognition sequence, designated as wild type, WT) and its mutants (CQLAWFDTDY, mC1; CQLAWFDADL, mC3; CQLAWFDADY, mC13; italicized part ) mutated residues), which were synthesized using Fmoc methods, were purchased from Invitrogen Corp. (Carlsbad, CA). They consist of the 10 amino acid sequences derived from the nine C-terminal residues of the β-catenin and an artificial cysteine residue at the N-terminal end of the peptide. Mutations were introduced to the first and third residues from the C-terminus of the peptides, which are the major determinants for the specific interaction with PDZ domains.28 Protein and Peptide Fixation onto Glass Substrates and AFM Probes. The following commercially available materials of special reagent grade were used for surface modification of glass substrates and AFM probes and for sample fixation: (3-aminopropyl)triethoxysilane (Shin-Etsu Chemical Co., Ltd., Tokyo, Japan), (3mercaptopropyl)trimethoxysilane (Shin-Etsu Chemical Co., Ltd.), and poly(ethylene glycol)-R-maleimide-ω-NHS ester (NHS-PEGMAL, Mw ) 3400; Nektar Therapeutics, San Carlos, CA). For the covalent fixation of TIP-1 and PDZ-pep onto glass substrates and AFM probe tips, these were modified with bifunctional a heterocrosslinker, NHS-PEG-MAL (see scheme 1), then reacted with the protein and peptide as follows: (1) Glass substrates (0.12-0.17 mm in thickness, 15 mm diameter; Matsunami Glass Ind., Ltd., Osaka, Japan) and AFM probes (OMCLTR400PSA-1, Olympus Optical Co., Ltd., Tokyo, Japan; holding a Si3N4 triangular cantilever with a sharpened pyramidal tip) were immersed in 80 °C piranha solution (concn H2SO4/30% H2O2 ) 7:3) for 1 h. (2) After rinsing with acetone and toluene, the glass substrates and AFM probes were immersed in 5% (v/v) toluene solutions of (3-mercaptopropyl)trimethoxysilane and (3-aminopropyl)triethoxysilane, respectively, and shaken for 18 h under an argon atmosphere at room temperature. (3) After sequential rinsing with toluene, acetone, and ethanol, the silanized glass substrates and AFM probes were immersed in deionized (DI) water; then the former was sonicated for 10 min and (28) Songyang, Z; Fanning, A. S.; Fu, C.; Xu, J.; Marfatia, S. M.; Chishti, A. H.; Crompton, A.; Chan, A. C.; Anderson, J. M.; Cantley, L. C. Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 1997, 275, 73-77.

2670 Langmuir, Vol. 23, No. 5, 2007 the latter was left at rest for 3 h to remove produced polysiloxane contaminants. (4) After drying at 115 °C for 10 min in air, both were immersed in a 2 mM aqueous solution of NHS-PEG-MAL for 90 min at room temperature. (5) After rinsing thoroughly with DI water, the NHS ends of the crosslinker whose MAL ends were immobilized onto thiol-silanized glass substrates and reacted with TIP-1 [2 mg‚mL-1 in phosphate-buffered saline (PBS, pH 7.4)] for 90 min (see Scheme 1). Alternatively, the MAL ends of the crosslinker whose NHS ends were fixed onto the amino-silanized AFM probes were reacted with PDZ-pep (2 mg‚mL-1 in PBS) for 90 min. (6) The TIP-1-fixed glass substrate and PDZ-pep-fixed AFM probes were rinsed with PBS thoroughly to remove the nonspecifically adsorbed and unreacted protein and peptides. (7) The glass substrates were immersed in 1.0 M Tris-HCl buffer (pH 7.5) for 30 min at room temperature to hydrolyze the unreacted activated ester group. Dynamic Force Spectroscopy. The unbinding force between TIP-1 and the PDZ-pep’s (WT, mC1, mC3, and mC13) was measured using an AFM (MFP-3D; Asylum Research, Santa Barbara, CA) in PBS at 25 ( 1 °C at different tip-retraction velocities (Vret ) 2012000 nm‚s-1). The tip approach velocity (Vapp) was set at 100 nm s-1 in the case of Vret values of 20, 50, 200, 500, 1500, 3000, 5000, and 12 000 nm‚s-1. Such an asymmetric velocity setting in the approach/retraction cycles was implemented using the MFP-3D software version IgorPro 5.03/MFP-3D Xop build 23 up1. The spring constants of the cantilever were determined according to the thermal noise/resonance method (21.93 ( 2.53 pN‚nm-1); the relative trigger mode was used to keep the contact force between the tip and the substrate almost constant (300-400 pN). In each Vret condition, around 10-30 force-versus-distance curves (f-d curves) were recorded at one position in repeated cycles of the approach/retraction, which were performed over 100 different positions of the substrate. From the several thousand f-d curves collected, histograms of the unbinding force were generated using a macro program of Excel software (Microsoft Corp.) that we originally constructed. The unbinding force histograms were fitted with multiple Gaussian functions using the least-squares method. The first peaks were assigned as a one-paired unbinding force. Surface Plasmon Resonance. SPR analyses were performed using a Biacore 3000 system (Biacore AB, Uppsala, Sweden) at 25 °C. The N-terminal biotinylated PDZ-pep (Thermo Electron GmbH, Ulm, Germany) was coupled to a streptavidin-coated sensor chip (Sensor Chip SA; Biacore AB). All measurements were carried out in HBS-EP buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM EDTA, and 0.005% Tween20) as the running buffer. For the association step, TIP-1-His solutions of five different concentrationss0.05, 0.1, 0.2, 0.5 and 1.0 µMswere flowed at 20 µL‚min-1 for 240 s. The dissociation step was performed by passing the running buffer at same flow rate used for the association step for 240 s. The sensorgram using each protein concentration was observed through triplicate analyses. Association (ka) and dissociation (kd) constants between TIP-1 and PDZ-pep’s were obtained by fitting a 1:1 (Langmuir) binding model modified under the mass transport limited condition using computer software (BIAevaluation software version 4.1; Biacore AB).

Results Unbinding Force Measurement under Different TipRetraction Velocities. To characterize the dynamic strength of the unbinding force between TIP-1 and PDZ-pep, the f-d curves between them were measured under eight different conditions of tip-retraction velocities (Vret ) 20, 50, 200, 500, 1500, 3000, 5000, and 12 000 nm‚s-1). Figure 1 shows representative f-d curves observed between TIP-1 and PDZ-pep WT. As the retraction velocity was raised from 20 to 12 000 nm‚s-1, the height of the force peak jump increased gradually from approximately 30 to 200 pN, and the Z-position of the peak moved toward the extension of the PEG-crosslinker. Particularly under the condition of higher Vret (3000, 5000, and 12 000 nm‚s-1), the approaching and retracting traces of the f-d curves in the

Maki et al.

Figure 1. Representative f-d curves measured under eight different conditions of AFM tip-retraction velocity: 20, 50, 200, 500, 1500, 3000, 5000, and 12 000 nm‚s-1. The black and red curves respectively show approaching and retracting traces.

tip-sample noncontact region did not mutually correspond because of the effect of increased drag force. Prior to the determination of unbinding force values from the force peak jump in the obtained f-d curves, the appearance frequencies of the peaks within the total measurements (10002500 curves) were compared with those observed in the control systems, including TIP-1-absent, PDZ-pep-absent, or both absent setups. As shown in Table 1, force peaks derived from nonspecific interactions in the control systems (PEG-crosslinker vs peptides, TIP-1 vs PEG-crosslinkers, or PEG-crosslinkers vs PEGcrosslinkers) were observed to be approximately 2-7% of the total measurement, whereas the appearance frequency of the force peaks observed for the sample setups including PDZ-pep mC1, mC3, or mC13 were comparable to the control levels, thereby indicating that the unbinding force between TIP-1 and the mutant peptides was too weak to be detected. In contrast, in the systems of TIP-1 and PDZ-pep WT, appearance frequencies of force peak were markedly larger than the control levels and were approximately 15-30%, suggesting that the force peaks are assignable to the unbinding event of the specific complex between them. On the basis of the above confirmations, the unbinding force was determined only for the systems of TIP-1 and PDZ-pep WT as follows. Figure 2 shows histograms of the unbinding force between TIP-1 and PDZ-pep WT in each condition of retraction velocity. The most frequent values were observed in the low-force region, and broad distribution appeared in the high-force region, which gradually shifted to higher force regions as the retraction velocity increased. From nonlinear least-mean-square multiple Gaussian fitting, each histogram was found to contain several Gaussian peaks in which the peak values increased as an almost integral multiplication of the first peak value. On the basis of this analysis, the first peak values were assigned as a one-paired unbinding force between TIP-1 and PDZ-pep WT, and were determined to be 41 ( 15 pN (Vret ) 20 nm‚s-1; the following parentheses also denote Vret), 54 ( 19 pN (50 nm‚s-1), 76 ( 17 pN (200 nm‚s-1), 71 ( 19 pN (500 nm‚s-1), 115 ( 31 pN (1500 nm‚s-1), 154 ( 40 pN (3000 nm‚s-1), 167 ( 65 pN (5000 nm‚s-1), and 217 ( 69 pN (120 000 nm‚s-1).

Specific Interaction between TIP-1 and PDZ-pep

Langmuir, Vol. 23, No. 5, 2007 2671

Table 1. Appearance Frequency of Detectable Force Peaks Measured under Eight Different Conditions of AFM Tip-Retraction Velocity for Sample Systems (pairs of TIP-1 and PDZ-pep’s of WT, mC3, mC1 and mC13) and for Control Systems (without one side or both side analytes)a appearance frequency of force peaks (%) samples Vret (nm s-1) 20 50 200 500 1500 3000 5000 12000 a

TIP-1 PDZ-pep

controls

(+) (+)

(-) (+)

WT

mC3

mC1

mC13

WT

mC3

mC1

mC13

14.9 21.2 22.6 17.9 21.6 26.3 23.0 31.2

3.0 5.0 4.5 7.5 3.5 2.5 4.0 6.0

1.1 3.5 4.7 5.9 8.8 6.0 5.1 3.5

1.9 2.7 2.4 6.3 4.7 2.5 2.9 2.9

2.0 1.0 2.0 1.0 7.0 3.0 3.0 1.0

4.0 5.0 1.0 3.0 1.0 3.0 3.0 3.0

1.0 1.0 1.0 3.0 5.0 1.0 4.0 3.0

2.0 2.0 3.0 3.0 1.0 2.0 5.0 3.0

(+) (-)

(-) (-)

3.0 1.0 1.0 3.0 2.0 3.0 4.0 2.0

0 0 0 1.0 1.0 1.0 1.0 2.0

Frequency was calculated from approximately 3000-5000 curves.

Figure 3. Dynamic force spectra for specific interactions between TIP-1 and PDZ-pep WT.

Figure 2. Distributions of the unbinding force measured under eight different conditions of AFM tip-retraction velocity. Measured numbers of f-d curves (N) and determined force values of the first unbinding force peak (f*) are inserted in each graph.

Dynamic Force Spectrum. Quantitative dependence of the unbinding force on the mechanical dissociation rate (i.e., the dynamic force spectrum) was plotted between 90 and 140 000 pN‚s-1 (Figure 3). The loading rate was calculated as the product of the tip-retraction velocity (Vret) and the slope of the f-d curves at the unbinding peak (V). Within this loading rate range, the unbinding force exhibited two regimes of an initial gradual increase between 90 and 3800 pN‚s-1, followed by a more rapid increase between 3800 and 140 000 pN‚s-1 with increasing loading rates. Slopes (fβ) and x-intercepts (r0f ) in the latter and former regions were respectively determined as fβ1 ) 100.08

pN, r0f1 ) 1102.5 pN‚s-1 and fβ2 ) 19.49 pN, r0f2 ) 0.541 pN‚s-1 (subscript numbers 1 and 2 denote their respective regions). SPR Analysis. The TIP-1 versus PDZ-pep interactions were analyzed using SPR from the view of the mass scale of the molecular ensemble to compare with the results from the single molecular scale measurement of DFS. Figure 4 shows the realtime change in the resonance unit (RU), reflecting the amount of bound peptides to TIP-1’s fixed onto the sensor substrate. The largest RU in the equilibrium stage was observed for WT, followed by mC3, mC1, and mC13. The RU observed for WT was about 15 times larger than that for mC3. The mC1 and mC13 exhibited little increase in RU in the equilibrium stage, indicating that these two peptides showed only slight interactions with TIP-1. The kinetic and thermodynamic parameters obtained from SPR analyses are tabulated in Table 2. These results indicate that mutation of the first residue from the C-terminus of PDZ-pep was found to reduce the interaction strength with the PDZ domain most critically, followed by that of the third residue. In the case where only the third residue was mutated, RU changes that resulted from mass scale behaviors were detectable, even though the

2672 Langmuir, Vol. 23, No. 5, 2007

Maki et al.

state S(t) is expressed as

dS(t) ) - k(f)S(t) dt

(3)

and is solvable as

[ { ( ) }]

S(t) ) exp -k0

fβ f(t) exp -1 rf fβ

(4)

where rf is the force loading rate (f(t) ) rft). Equation 3 expresses the rate of escape from the bound state under applied load f. The equilibrium condition of bond unbinding for the probability density of the unbinding process F,

Figure 4. SPR sensorgrams observed for the binding and unbinding processes of TIP-1 (analyte) on PDZ-pep’s (ligand). Analyte concentration: 1.0 µM. Table 2. Rate Constants (association: ka, dissociation: kd) and Equilibrium Constants (association: KA, dissociation: KD) Analyzed from SPR Sensorgrams Observed for the Interaction of TIP-1 with PDZ-pep WT, mC1, and mC3a PDZ-pep

ka (1/M s)

kd (1/s)

KA (1/M)

KD (M)

WT mC1 mC3 mC13

6.44 × 105 0.24 × 101 3.88 × 103 N.D.

2.89 × 10-1 2.90 × 10-3 4.61 × 10-3 N.D.

2.25 × 106 6.14 × 103 8.27 × 105 N.D.

4.46 × 10-7 1.63 × 10-4 1.23 × 10-6 N.D.

a For the system of TIP-1 and mC13, those values were not determined (N.D.).

unbinding force values measured under the single molecular level were undetectable.

Discussion In this study, to investigate the molecular basis for the interaction between the PDZ domain and its recognition peptide, the dynamic force spectrum for the interaction was characterized by AFM, and the thermodynamic and kinetic parameters were determined using SPR. These analyses can elucidate the rough geometry of the energy landscape for the interaction. Below, we discuss the assignment of a deduced value into the conceptual energy landscape, compare the information from DFS and SPR, and consider the reasons for their mutual differences. The dynamic force spectrum shown in Figure 3 was analyzed on the basis of the Bell-Evans model.14-16 This model assumes that the work done by an external pulling force, fx (f ) external force, x ) pulling distance), is imposed on the bound complex, thereby distorting the energy landscape of the interaction, and lowering the activation barriers. Consequently, in the presence of load f, the dissociation rate constant k(f) is perturbed from the original dissociation constant k0, which can be given as

k(f) ) k0 exp

() f fβ

(1)

Here, fβ is the thermal fluctuation force and can be written as

fβ )

kBT xβ

(2)

where kB is the Boltzmann constant, T is the absolute temperature, and xβ is the width of the potential barrier along the direction of the external pulling force. The likelihood of being in the bound

d2S(t) dF )0 )dt dt2

(5)

gives the simple relationship between the unbinding force peak f* and rf:

()

rf f* ) ln 0 fβ rf

(6)

Therein, r0f is the loading rate of the thermal fluctuation force. Equation 6 shows that the unbinding force peaks have a linear relationship with the logarithmic force-loading rate; its slope corresponds to fβ, which can be used to determine xβ based on eq 2. In addition, the x-intercept r0f is, in principle, equal to k0fβ, which can provide k0. Figure 3 shows that the dynamic force spectrum for the PDZ interaction exhibited two linear regimes, and both xβ and k0 were evaluated as 0.04 nm and 1.10 × 10 s-1 from the fast-rate regime, and 0.21 nm and 2.77 × 10-2 s-1 from the slow-rate regime. Squares of the correlation coefficient R2 for the fitting were 0.97 and 0.96, respectively. Similar types of such two-step dissociation kinetics have been reported in many cases of the ligand-receptor interactions, for example, positions of the first and second activation barriers for avidin/biotin (0.20 and 0.53 nm17), streptavidin/biotin (0.05 and 0.49 nm17), and integrin R4β1/VCAM-1 (0.1 and 0.59 nm19), respectively, which has also been confirmed theoretically.29 A characteristic part of the geometry of the energy landscape for the PDZ interaction can be deduced from these values: barrier positions xβ, and energy heights ∆E of two activation barriers. The latter values are written as

∆Ed ) -kBT(ln kd - A)

(7)

where ∆Ed and kd are energy barriers and dissociation rate constants, respectively. In addition, A is a constant derived from frequency factors in the dissociation process. The geometry of the energy landscape is deduced as shown in Figure 5a, arbitrarily presuming two energy levels of the bound complex as being the ground and a constant A. Although the absolute energy levels of each barrier from the level of the bound state are uncertain because of the uncertainty of constant A, the relative energy differences between them (6.0 kBT) and the absolute positions of each barrier (0.04 nm for the first peak and 0.21 nm for the second one) were determined. On the other hand, the rate constants and equilibrium constants of the dissociation and association processes, which were (29) Strunz, T.; Oroszlan, K.; Schumakovitch, I.; Guntherodt, H.-J.; Hegner, M. Model energy landscapes and the force-induced dissociation of ligand-receptor bonds. Biophys. J. 2000, 79, 1206-1212.

Specific Interaction between TIP-1 and PDZ-pep

Langmuir, Vol. 23, No. 5, 2007 2673

in Figure 5b, arbitrarily presuming two energy levels of the bound complex with PDZ-pep WT as the ground and a constant B. Additionally, for simplicity, frequency factors of the dissociation process were assumed to be approximately the same for both DFS and SPR measurement, thus the line of constant A in Figure 5b was drawn at the same level as that in Figure 5a. As might be apparent from Figure 5b, bound states of mC1 and mC3 are respectively destabilized by 5.88 kBT and 1.0 kBT compared to WT, and the activated complexes are more unstable than WT by 10.52 kBT and 5.14 kBT, respectively. To compare the topography of the energy landscape deduced by DFS with that deduced by SPR, panels a and b of Figure 5 were superimposed for PDZ-pep WT as shown in Figure 5c. The single barrier deduced by SPR was found to be positioned approximately between the double barriers deduced by DFS. Such differences in the topographic characteristics are attributable to the difference in the dissociation pathways between those two different measurements. The DFS analyses were performed under a forced unbinding process, whereas SPR analyses were done for spontaneous dissociation processes. In the present experiment, PDZ-pep’s were pulled from N-terminals crosslinked with a spacer polymer; consequently, the forced unbinding can be induced predominantly along the direction of the PDZ domain groove (groove size: 1.0 nm × 0.8 nm from the steric structure of the PDZ domain; PDB code 1DQI). In contrast to the SPR measurement, PDZ-pep’s freely dissociate without such limitation in the reaction pathway. The system can find the most probable pathway with the lowest activation barrier. This might explain why the height of the activation barriers characterized by spontaneous pathways in SPR measurements was lower than that of the second activation barrier seen in the forced pathway in DFS analysis. Figure 5. Conceptual energy landscapes deduced from DFS analysis (a) and SPR results (b). (c) Superimposed representation of panel a and the curves for WT in panel b. A and B are constants respectively derived from frequency factors in the dissociation and association processes.

determined using SPR and shown in Table 2, also provide characteristic features in the energy landscape of PDZ interactions on WT, mC1, and mC3. The activation energy barrier in the association process ∆Ea and the equilibrium energy differences between the bound state and the dissociated free state ∆Efrc are given as

∆Ea ) -kBT(ln ka - B)

(8)

∆Efrc ) -kBT ln KD

(9)

Conclusion The present study investigated the molecular basis of the specific interactions of PDZ proteins from the view of the energy landscape of the ligand-receptor complex. Two different methods of characterization of the specific interactionssDFS and SPRs were employed to deduce the rough topographies of the energy landscape. The reason for the topographic difference in the deduced energy landscapes between them was considered to be the result of differences in the reaction pathways in the forced unbinding (DFS) and spontaneous dissociation (SPR) processes. These findings demonstrate conclusively that, in the determination of the binding strength of ligand-receptor interaction, pathwaydependent aspects that are intrinsic to the characterization method should be considered for the definition of its strength.

and

where ka is the association rate constant, B is a constant derived from frequency factors in the association process, and KD is the dissociation equilibrium constant. Similar to the line shown in Figure 5a, the geometry of the energy landscape is depicted as

Acknowledgment. This work was supported by the Core Research for Evolutional Science and Technology (CREST) Program of the Japan Science and Technology Corporation (JST), and by the Research Grant for the RIKEN Genome Exploration Research Project from MEXT of Japan to Y.H. LA0627011