E XP E RI ME N TA L CE L L RE S E A RCH ( 00 ) –26 a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m w w w. e l s e v i e r. c o m / l o c a t e / y e x c r Research Article Probing effects of pH change on dynamic response of Claudin-2 mediated adhesion using single molecule force spectroscopy Tong Seng Lim a,b , Sri Ram Krishna Vedula c , Shi Hui c , P. Jaya Kausalya d , Walter Hunziker d , Chwee Teck Lim c,⁎ a Bioinformatics Institute, Agency for Science, Technology and Research (A⁎STAR), 30 Biopolis Street, #07-01 Matrix, Singapore 138671 NUS Graduate School for Integrative Sciences & Engineering (NGS), Centre for Life Sciences (CeLS), #05-01, 28 Medical Drive Singapore 117456 c Division of Bioengineering & Department of Mechanical Engineering, Engineering Drive 1, National University of Singapore, Singapore 117576 d Institute of Molecular and Cell Biology, Agency for Science, Technology and Research (A⁎STAR), 61 Biopolis Drive, Proteos, Singapore 138673 b ARTICLE INFORMATION ABSTRACT Article Chronology: Claudins belong to a large family of transmembrane proteins that localize at tight junctions Received April 2008 (TJs) where they play a central role in regulating paracellular transport of solutes and Revised version received nutrients across epithelial monolayers. Their ability to regulate the paracellular pathway is 27 May 2008 highly influenced by changes in extracellular pH. However, the effect of changes in pH on the Accepted 27 May 2008 strength and kinetics of claudin mediated adhesion is poorly understood. Using atomic force Available online June 2008 microscopy, we characterized the kinetic properties of homophilic trans-interactions between full length recombinant GST tagged Claudin-2 (Cldn2) under different pH Keywords: conditions. In measurements covering three orders of magnitude change in force loading Claudin rate of 102–104 pN/s, the Cldn2/Cldn2 force spectrum (i.e., unbinding force versus loading rate) Tight junction revealed a fast and a slow loading regime that characterized a steep inner activation barrier Cell–cell adhesion and a wide outer activation barrier throughout pH range of 4.5–8. Comparing to the neutral Molecular force spectroscopy condition (pH 6.9), differences in the inner energy barriers for the dissociation of Cldn2/Cldn2 Atomic force microscopy mediated interactions at acidic and alkaline environments were found to be b0.65 kBT, which pH is much lower than the outer dissociation energy barrier (N1.37 kBT). The relatively stable interaction of Cldn2/Cldn2 in neutral environment suggests that electrostatic interactions may contribute to the overall adhesion strength of Cldn2 interactions. Our results provide an insight into the changes in the inter-molecular forces and adhesion kinetics of Cldn2 mediated interactions in acidic, neutral and alkaline environments. © 2008 Elsevier Inc. All rights reserved. Introduction Tight junctions (TJs) are the apical most constituents of the intercellular adhesion complex in epithelial monolayers. Their primary function is to regulate the paracellular transport of ions, solutes and water across epithelia. In addition, they interact with a variety of signaling and trafficking molecules to regulate cell differentiation, proliferation and polarity [1,2]. The selective permeability of TJs is largely determined by a protein family called claudins (Cldns) [3–5]. Although the contribution of ⁎ Corresponding author. Division of Bioengineering and Department of Mechanical Engineering, National University of Singapore, Engineering Drive 1, Singapore 117576. E-mail address: ctlim@nus.edu.sg (C.T. Lim). 0014-4827/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2008.05.015 2644 E XP E RI ME N TA L CE LL RE S E A RCH ( 00 ) –26 Cldns to the charge selective permeability and ion homeostasis of epithelia is well established [6–16], details about the strength and adhesion kinetics of the interactions mediated by Cldns are being understood only recently [17]. Cldn2 was found to exhibit Ca2+-independent adhesion activities in cell aggregation assays [18]. Expression of Cldn2 has been shown to induce cation-selective channels in TJs of epithelial cells [6]. Also, increased expression of Cldn2 has been shown to decrease transepithelial electrical resistance (TER) [19,20] while increasing the density of small TJ pores [21]. Furthermore, the knockdown of endogenous Cldn2 expression in MDCK cells using siRNA resulted in decreased Na+ permeation and loss of cation selectivity [22]. In a more recent study, Cldn2 was shown to be critical for Vitamin D-dependent Ca2+ absorption between enterocytes [23]. Since it has been shown that the transport of several nutrients can be influenced by varying the extracellular pH [24–26], understanding the pHassociated changes of the adhesion kinetics mediated by Cldns will provide us a better perspective on how it regulates the paracellular transportation of solutes and intercellular adhesions. To address this question, we used single molecule force spectroscopy to investigate the molecular interactions between recombinant N-terminal glutathione S-transferase (GST) tagged full length human Cldn2 (GST-Cldn2) under different pH conditions (pH 4.5, 5.1, 6.9 and 8). Our results show that dissociation of homophilic Cldn2/ Cldn2 complexes follows a two-step energy barriers model within the pH range of 4.5–8 and loading rates of 102–104 pN/s. The energy landscape of the dissociations was found to be dynamically dependent on the changes in environmental pH. Comparison of adhesion kinetics further revealed that Cldn2/ Cldn2 is relatively more stable in neutral solution (pH 6.9) when compared to acidic or alkaline environments, implying that electrostatic interactions may contribute to the adhesion strength of Cldn2 mediated adhesions. Materials and methods Protein immobilization and cantilever functionalization Functionalization of AFM cantilevers was performed using methods described previously [27]. Soft silicon nitride tips (Vecco, Santa Barbara, CA) were UV irradiated for 15 and incubated in a mixture of 30% H2O2/70% H2SO4 for 30 min. After washing thoroughly in ddH2O, tips were dried and treated with a 4% solution of APTES (3-aminopropyltriethoxysilane, Sigma) in acetone for min. They were then rinsed thrice in acetone and incubated in a solution of BS3 (Bis (Sulfosuccinimidyl) suberate, mg/ml, Pierce) for 30 min, followed by the incubation of antiGST antibody (10 µg/ml, Invitrogen) for h. The reaction was quenched using M Tris buffer, followed by the incubation with recombinant full length GST-Cldn2 (10 µg/ml, Proteintech Group, Inc, USA) or GST-Cldn1 (10 µg/ml, Abnova, Taiwan) for h. Unbound recombinant proteins were washed off with PBS. Tips were blocked in 1% BSA before experiments [27]. Recombinant GST-Cldn2 or GST-Cldn1 was immobilized on glass cover slips using the same procedure as described above. To confirm that GST-Cldn2 was efficiently linked to the silanized tips, pri- mary mouse anti-Cldn2 antibody (Abnova, Taiwan) and Alexa 488-labeled goat anti-mouse secondary antibody (Molecular Probes, Invitrogen) were used to stain the GST-Cldn2-coupled tips. For control experiments, all steps were similar except that incubation of recombinant GST-Cldn2 proteins was omitted. For blocking experiments, tips and cover slips were incubated with antibody targeting the first extracellular loop of Cldn2 (10 µg/ml, Abnova, Taiwan) for 30 min. They were then washed to remove any unbound antibody before the experiments. For competition assays, interactions of GST-Cldn2/GST-Cldn2 were probed in the presence of GST-C2E1 (10 µg/ml, Abnova, Taiwan) (C2E1: first extracellular loop of Claudin-2, UUNP_065117UU, 29 a.a.–81 a.a.) or GST-C2E2 (10 µg/ml, Abnova, Taiwan) (C2E2: second extracellular loop of Claudin-2, UUNP_065117UU, 138 a. a.–163 a.a.) in PBS buffer. Molecular force spectroscopy Force curves were acquired on a MultiMode™ Picoforce™ AFM (Vecco, Santa Barbara, CA) coupled to an upright microscope at room temperature using a fluid cell. Cantilevers with a nominal spring constant of 0.01–0.03 N/m were used for obtaining force plots. Prior to obtaining force curves, the spring constant was determined using the thermal tune module. Target proteins (GST-Cldn2 or GST-Cldn1) immobilized on the glass cover slips were probed with cantilevers functionalized with recombinant proteins (GST-Cldn2 or GST-Cldn1) under different pH conditions (pH 4.5–8) in PBS buffer. Force plots were obtained at different reproach velocities (0.1–2 μm/s) and were analyzed for the magnitude of the rupture events and the apparent loading rate (defined as the slope of the retrace curve prior to the rupture event multiplied by the reproach velocity) using MATLAB version 7.1 (The MathWorks, Natick, MA.). Following Hanley et al. [28] and Panorchan et al. [29], rupture force measurements were partitioned by using binning windows of 50 pN/s for loading rates between 100 and 1000 pN/s and by binning windows of 500 pN/s for loading rates between 1000 and 10,000 pN/s. Each bin yields a mean force by Gaussian fitting. By plotting the mean force as a function of loading rate, the unstressed dissociation rate and reactive compliance characterizing Cldn2/Cldn2 interactions in different pH conditions were extracted (see Results section). Results Measurement of Cldn2/Cldn2 interaction forces Trans-interactions between full length human Cldn2 (Cldn2/ Cldn2) were characterized at the level of single molecule using atomic force microscopy (AFM) (Fig. 1) [28–30]. The interaction was established by bringing GST-Cldn2 functionalized cantilever in close contact to a glass cover slip coated with GSTCldn2 under different pH conditions (pH 4.5–8) (see Materials and methods). To confirm that GST-Cldn2 was efficiently linked to the silanized tips, primary mouse anti-Cldn2 antibody and Alexa 488-labeled goat anti-mouse secondary antibody were used to stain the GST-Cldn2-coupled tips. Confocal images showed that GST-Cldn2 was efficiently coupled to the AFM tips and cover slips (Fig. 2). E XP E RI ME N TA L CE L L RE S E A RCH ( 00 ) –26 2645 has previously been used to characterize binding interactions between other intercellular adhesion molecules, such as nectin/nectin [38,39], VE-cadherin/VE-cadherin [40], N-cadherin/ N-cadherin and E-cadherin/E-cadherin interactions [29,38,41]. In this model, the probability density function for the dissociation of a bound complex at force f is given by: Pð f Þ ¼ Fig. – Schematic of the atomic force microscopy (AFM) experimental setup. Recombinant GST-Cldn2 was linked to the AFM tip or immobilized on glass cover slip using the linker APTES-BS3-anti-GST (see Materials and methods for details). GST-Cldn2 immobilized on glass cover slip was probed using these functionalized tips under different pH condition in PBS buffer. The arrow indicates the direction of pull in the AFM experiment. GST: glutathione S-transferase; Cldn2: Claudin-2; APTES: 3-aminopropyltriethoxysilane; BS3: Bis (Sulfosuccinimidyl) suberate; anti-GST: antibody targeting GST; PBS: phosphate buffered saline. To measure the de-adhesion forces at single molecular resolution [30–32], a contact force of 200 pN and a contact time of ms were used. Under such conditions, low frequency of de-adhesion events (b25%) was achieved, which ensured a N86% probability of single bond rupture based on Poisson statistics [33]. Upon retraction of the cantilever, force as a function of pulling distance was recorded (Fig. 3a) [34]. For each reproach velocity, hundreds of force–distance curves (n N 500) were collected and analyzed to extract rupture force, F and loading rate, rf (Fig. 3b). Data obtained were subsequently pooled into histograms to analyze the frequency of adhesion events for different interactions (Table 1, Fig. 4). Results showed that adhesion frequency was significantly reduced in control experiments performed using AFM tips functionalized with only anti-GST antibody (Cldn2_Anti-GST, Table1, Fig. 4). On the other hand, the low frequency of interaction between Cldn1 and Cldn2 (Cldn1_Cldn2 or Cldn2_Cldn1, Table 1, Fig. 4) demonstrated that they not trans-interact which is consistent with previous findings [35]. In addition, it was found that experiments performed using antibody specifically targeting the first extracellular loop of Cldn2 (C2E1) significantly reduced the binding frequency. Competition assays revealed that only the first but not the second extracellular loop of Cldn2 can compete for the interactions between Cldn2/ Cldn2 (Cldn2_Cldn2_C2E1, Cldn2_Cldn2_C2E2, Table 1, Fig. 4). These results suggest that C2E1 itself is sufficient to promote the trans-interactions of Cldn2/Cldn2. Extraction of the kinetic parameters of Cldn2/Cldn2 interactions Biophysical parameters characterizing the kinetics of Cldn2/ Cldn2 interactions under different pH conditions were evaluated using the Bell–Evans model [36,37]. This model relates the bond rupture force to the loading rate applied to the bond. It  &  !'    koff k kB T x f x f À exp b exp b exp off kB T kB T xβ rf rf ð1Þ where rf is rate of force application (i.e., loading rate), kB is Boltzmann constant, T is the absolute temperature, k0off is the unstressed dissociation constant and x β is the reactive compliance. Moreover, as shown in Eq. (2), the average unbinding force of a complex, bfN, increases with rf [27,28,32,42,43],     k kB T k kB T kB T exp off Ei off ð2Þ bf N ¼ xβ rf xβ rf xβ Rl À1 t expðÀtÞdt is the exponential integral. Eq. (2) Here Ei ðzÞ ¼ z describes the dynamic properties of a system consisting of a single activation barrier. For each pH condition, by fitting the rupture force vs. loading rate data points using Eq. (2), the Fig. – Confocal images of silanized AFM cantilevers functionalized (a) with GST-Cldn2 and (b) without GST-Cldn2. Both images were taken after the cantilevers/tips were incubated with anti-Cldn2 primary antibody and Alexa 488-labeled secondary antibody (see Materials and methods section for details). Images were acquired under the same conditions (pixel dwell time, laser power and gain). Scale bar size: 50 μm. 2646 E XP E RI ME N TA L CE LL RE S E A RCH ( 00 ) –26 Fig. – Force–Displacement curves showing rupture of individual bonds mediated by Cldn2/Cldn2 interactions. (a) Typical force–distance curves obtained between tip functionalized with GST-Cldn2 and GST-Cldn2 immobilized on glass cover slips. Arrows indicate rupture of homophilic Cldn2/Cldn2 interactions. The curves show either no or single bond rupture event. Only curves showing a single clear rupture event were used for generating the histograms. (b) The slope of the curve just before rupture multiplied by the reproach velocity (Vr , expressed in nm/s) defines the loading rate (rf, expressed in pN/s). The height of the rupture event defines the magnitude of rupture force (F) (expressed in pN). GST: glutathione S-transferase. unstressed dissociation constant (k0off ) and the reactive compliance (xβ) of the Cldn2/Cldn2 interaction were extracted (Fig. 5). Within the range of loading rates probed (102–104 pN/s), the average unbinding force of Cldn2/Cldn2 complexes was found to be higher with increasing loading rate. Moreover, the Cldn2/ Cldn2 force spectrum (i.e., unbinding force versus loading rate) revealed a fast and a slow loading regime that characterized a steep inner activation barrier and a wide outer activation barrier throughout all pH conditions (pH 4.5, 5.1, 6.9 and 8.0) (Fig. 5). A Table – Experiments for studying homophilic Cldn2/Cldn2 interactions corresponding to histograms depicted in Fig. Interaction type Cldn2_Anti-GST Cldn2_Cldn1 Cldn1_Cldn2 Cldn2_Cldn2_Ab Cldn2_Cldn2_C2E1 Cldn2_Cldn2 Cldn2_Cldn2_C2E2 AFM tip a Glass substrate a Anti-GST GST-Cldn2 GST-Cldn2 GST-Cldn1 GST-Cldn1 GST-Cldn2 GST-Cldn2 + Antibody GST-Cldn2 + Antibody GST-Cldn2 + GST-C2E1 GST-Cldn2 + GST-C2E1 GST-Cldn2 GST-Cldn2 GST-Cldn2 + GST-C2E2 GST-Cldn2 + GST-C2E2 gradual increase in unbinding force was observed with increasing loading rate of up to ~103 pN/s. Beyond this point, a second loading regime exhibiting a faster increase in the unbinding force was observed. Table lists the kinetic parameters (unstressed dissociation off rate k0off and reactive compliance xβ ) of the two energy barriers that were derived from fitting the experimental data with Eq. (2) using non-linear least square method with trust-region algorithm [44]. The fitted curves using Bell–Evans model are overlaid on the experimental measurements (Fig. 5). Monte Carlo simulation Monte Carlo (MC) simulations of receptor–ligand bond rupture under constant loading rates were performed to further corroborate our experimental results with Bell–Evans model predictions using a previously described procedure [27,29]. One thousand rupture forces (Frup = (rf)(nΔt)) were calculated for which the probability of bond rupture Prup: " # koff KB T  xh rf nDt À1 ð3Þ Prup ¼ À exp À exp kB t xh rf a Anti-GST: antibody targeting GST; GST-Cldn1: recombinant GSTtagged Claudin-1 protein; GST-Cldn2: recombinant GST-tagged Claudin-2 protein; GST-C2E1: recombinant GST-tagged first extracellular loop of Claudin-2 protein. GST-C2E2: recombinant GSTtagged second extracellular loop of Claudin-2 protein; antibody: antibody targeting the first extracellular loop of Claudin-2 protein. (See Materials and methods for details about the immobilization of proteins onto AFM tip and glass substrate.) was greater than Pran, a random number between zero and one. Here nΔt is the time interval needed to break a bond and Δt is the time step (Δt = 10−6 s was used in the simulation). The values of unstressed dissociation rate (k0off = 6.39 s−1) and reactive compliance (xβ = 0.19 nm) obtained experimentally in neutral pH (pH 6.9) for Cldn2/Cldn2 between loading rate of 103 – 104 pN/s were used in the simulation. Results obtained E XP E RI ME N TA L CE L L RE S E A RCH ( 00 ) –26 2647 Fig. – Rupture force histograms obtained at a reproach velocity of µm/s for the different interaction types listed in Table 1. Cldn2: Claudin-2; C2E1: first extracellular loop of Claudin-2; C2E2: second extracellular loop of Claudin-2. Fig. – Molecular force spectroscopy of homophilic Cldn2/Cldn2 interactions probed under different pH conditions. The mean rupture force was plotted as a function of loading rate. There was a gradual increase in rupture force along with loading rates up to ~1000 pN/s. This was followed by the faster increase in the unbinding force for loading rates greater than 1000 pN/s. By fitting the experimental data from each loading rate regime to Eq. (2), the unstressed dissociation rate (k0off) and reactive compliance (xβ ) for Cldn2/Cldn2 interactions were extracted (see Table 2). The error bars are the standard errors of the measurements. 2648 E XP E RI ME N TA L CE LL RE S E A RCH ( 00 ) –26 Table – pH dependence of adhesion kinetics mediated by Claudin-2/Claudin-2 interactions pH Loading rate (pN/s) Rate of dissociation a k0off (s−1) Reactive compliance a xβ (nm) 4.5 102–103 103–104 102–103 103–104 102–103 103–104 102–103 103–104 1.79 × 10−2 23.66 4.55 × 10−3 8.19 4.42 × 10−3 6.39 0.61 12.58 0.68 0.02 0.73 0.06 0.94 0.19 0.24 0.03 5.1 6.9 8.0 a Reactive compliance xβ, and the unstressed bond dissociation rate k0off, were fitted from the loading rate curve (Fig. 5) using Eq. (2) using non-linear least square method with trust-region algorithm (Gilles et al., 2002). from the simulation agreed well with the experimental results (Fig. 6b). The loading rate (logarithmic scale) was assumed to be normally distributed within the simulation range of loading (103 – 104 pN/s) (x2 test, p b 0.05). This assumption is valid and will not cause significant deviations to the simulation as the cumulative distribution function of the loading rate agrees well between experimental data and MC simulation (Fig. 6a). The good agreement between computational and experimental distributions indicates that the rupture event is mainly caused by breaking of a single bond and not from the breaking of multiple bonds. Discussion Claudins are critical tetra-span proteins localizing at TJs which create the barrier and regulate electrical resistance and size and ionic charge selectivity during paracellular epithelial transport [45]. While the role of Cldns in creating charge selectivity in the paracellular pathway is well established and has been characterized in some detail [6–16], little is known about the strength of adhesion forces mediated by Cldns at TJs. Here, we used GST tagged full length human Cldn2 to understand kinetic properties and adhesion strength of homophilic Cldn2 interactions in more detail. The present work examines the interactions at the level of single molecules instead of describing global cellular adhesion behavior which has been measured previously using dual pipette [17] or cell aggregation [18] assays. It has been shown previously that the transportation of several nutrients as well as changes to TJ permeability (for example induced by copper) are influenced by variations of the extracellular pH [24–26]. To understand the dynamic response of individual Cldn2/Cldn2 interactions to the change in pH, we used AFM to probe the trans-interaction of Cldn2/ Fig. – Comparison of the experimental and theoretical rupture force distribution of Cldn2/Cldn2 interaction. (a) Empirical cumulative distribution function of loading rate for experimental data (dotted line) and Monte Carlo (MC) simulations (continuous solid line). (b) Experimental (black) and theoretical (white) histograms of rupture forces to break a single Cldn2/Cldn2 bond at loading rate between 103 and 104 pN/s. MC simulations, which were conducted using the Bell-Evans model unstressed off rate (k0off = 6.39 s−1) and reactive compliance (xβ = 0.19 nm) obtained from force spectroscopy experiments, were consistent with experimental data and further indicate that only one single type of bond was analyzed. E XP E RI ME N TA L CE L L RE S E A RCH ( 00 ) –26 2649 Fig. – Comparison of conceptual energy landscapes of dissociation pathway between homophilic Cldn2/Cldn2 interactions probed under different pH conditions. The dissociation of Cldn2/Cldn2 involves two energy activation barriers. They were constructed using the kinetic parameters obtained from the molecular force spectroscopy (Table 2, Fig. 4). Activation energy difference for inner and outer barriers was sketched in the figure, respectively. In general, dissociation pathways for Cldn2/Cldn2 interactions may take different reactive coordinates. Here, the geometric locations for their bound states in different pH conditions were plotted on the same reactive coordinates for the purpose of comparison. Cldn2 pairs at different pH conditions. In the range of pH 4.5–8, dissociations of Cldn2/Cldn2 complexes were found to follow a two-step energy activation barrier process within the probed loading rate of 102–104 pN/s (Figs. and 7). The dissociation rate of the bound complex in N barrier model is given by multiple Bell's model arranged in series [36,46,47]: kÀ1 ¼ N  X À ÁÃÀ1 k0i exp xβi F=kB T ð4Þ i¼1 where kB is the Boltzmann constant, T is the absolute temperature, and xβ i and k0i (i = 1, 2, ., N) are parameters corresponding to reactive compliance and unstressed dissociation rate for ith activation barrier along dissociation of bounded complex. These kinetic parameters can be further used to construct the geometry of the conceptual energy landscape for the dissociation pathway. To compare the topography of the energy landscapes of the dissociation of Cldn2/Cldn2 complexes at different pH conditions, the geometric locations of their bound states were plotted on the same reactive coordinates (Fig. 7). In general, dissociations of Cldn2/Cldn2 may involve different reactive coordi- nates. Compared to acidic (pH ~ 4.5) or alkaline (pH 8.0) conditions, the energy barriers for the dissociation of Cldn2/ Cldn2 are highest at pH 6.9, indicating that the Cldn2/Cldn2 bond is more stable in neutral solution. This may explain why the structural integrity of the tight junctions in epithelia is affected by low pH [48]. The dissociation rate constants were used to estimate the energy differences (ΔG) among transition state energies of Cldn2/Cldn2 complexes at different pH conditions: À Á DG ¼ ÀkB Tln k0A =k0B ð5Þ where k0A and k0B are dissociation rate constants of the Cldn2/ Cldn2 interaction at different pH conditions of A and B, respectively. The analysis reveals that the outer activation barriers of the Cldn2/Cldn2 complex at acidic (pH 4.5) and alkaline conditions (pH 8.0) are 1.37 and 4.9 kBT greater than at neutral condition (pH 6.9) (Fig. 7). Moreover, the energy difference of the inner barrier is small (b0.65 kBT), which implies that the difference in equilibrium dissociation constant between Cldn2/Cldn2 complexes at acidic or alkaline conditions arises from differences in the activation energies of the outer barrier. 2650 E XP E RI ME N TA L CE LL RE S E A RCH ( 00 ) –26 In this study, we have elucidated the dynamic response of Cldn2 mediated interactions in various acidic, neutral, and alkaline environments. The relative stability of the trans-interactions at neutral conditions suggests that electrostatic interactions may contribute to the overall adhesion strength. Decreased stability of Cldn2/Cldn2 interactions in acidic and alkaline environments could be attributed either to conformational changes in Cldn2 or changes in the charge distribution of critical amino acids mediating trans-interactions. It has previously been shown that the first extracellular loop of Cldn2 [6,7] confers charge selective paracellular permeability to epithelial monolayers. In the acidic environment (pH ~ 4.5), most of the amide groups in the side chain of positively charged residues such as tyrosine (NP_065117, a.a. 35 and 67), lysine (a.a. 31 and 48) and histidine (a.a. 57) in the first extracellular loop of Cldn2 are protonated (pK value N 4.5) whereas the carboxylic acids in the side chain of aspartic (a.a. 65 and 76) and glutamic acids (a.a. 53) are uncharged. Under such conditions, though the transinteractions could still occur via other means, such as the hydrophobic interactions of nonpolar residues, the electrostatic repulsion caused by the positive charged residues would destabilize the overall adhesion strength. When the pH is changed from neutral to 5.1, only the charge of a single histidine residue (a.a. 57) is affected, which could explain the small effect on the adhesion kinetics of Cldn2/Cldn2 interactions under this condition (Table 2, Fig. 7). Given that the charged residues of Cldns influence the paracellular ion selectivity in the TJs pore, it will be interesting to compare the kinetics of interactions by mutating the charged residues in future experiments. Understanding interactions mediated by Cldns is important not only because of the role that they play in regulating paracellular transport of solutes and intercellular adhesion but also because of their pathological role in acting as receptors for bacterial toxin [49,50] and co-receptors for the entry of hepatitis C virus [51]. [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] Acknowledgments This work was supported by the Biomedical Research Council (BMRC) from the Agency for Science, Technology & Research (A⁎STAR), Singapore. Their funding support is gratefully acknowledged. [17] REFERENCES [18] [1] K. Matter, M.S. Balda, Signalling to and from tight junctions, Nat. Rev. Mol. Cell. Biol. (2003) 225–236. [2] S. Tsukita, M. Furuse, M. Itoh, Multifunctional strands in tight junctions, Nat. Rev. Mol. Cell. Biol. (2001) 285–293. [3] M. Furuse, K. Fujita, T. Hiiragi, K. Fujimoto, S. Tsukita, Claudin-1 and -2: novel integral membrane proteins localizing at tight junctions with no sequence similarity to occludin, J. Cell. Biol. 141 (1998) 1539–1550. [4] K. Morita, M. Furuse, K. Fujimoto, S. Tsukita, Claudin multigene family encoding four-transmembrane domain protein components of tight junction strands, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 511–516. [5] Y.H. Loh, A. Christoffels, S. Brenner, W. Hunziker, B. Venkatesh, Extensive expansion of the claudin gene family in the [19] [20] [21] teleost fish, Fugu rubripes, Genome. Res. 14 (2004) 1248–1257. S. Amasheh, N. Meiri, A.H. Gitter, T. Schoneberg, J. Mankertz, J.D. Schulzke, M. Fromm, Claudin-2 expression induces cation-selective channels in tight junctions of epithelial cells, J. Cell. Sci. 115 (2002) 4969–4976. O.R. Colegio, C. Van Itallie, C. Rahner, J.M. Anderson, Claudin extracellular domains determine paracellular charge selectivity and resistance but not tight junction fibril architecture, Am. J. Physiol. Cell. Physiol. 284 (2003) C1346–C1354. O.R. Colegio, C.M. Van Itallie, H.J. McCrea, C. Rahner, J.M. Anderson, Claudins create charge-selective channels in the paracellular pathway between epithelial cells, Am. J. Physiol. Cell. Physiol. 283 (2002) C142–C147. C. Van Itallie, C. Rahner, J.M. Anderson, Regulated expression of claudin-4 decreases paracellular conductance through a selective decrease in sodium permeability, J. Clin. Invest. 107 (2001) 1319–1327. H. Wen, D.D. Watry, M.C. Marcondes, H.S. Fox, Selective decrease in paracellular conductance of tight junctions: role of the first extracellular domain of claudin-5, Mol. Cell. Biol. 24 (2004) 8408–8417. M.D. Alexandre, B.G. Jeansonne, R.H. Renegar, R. Tatum, Y.H. Chen, The first extracellular domain of claudin-7 affects paracellular Cl− permeability, Biochem. Biophys. Res. Commun. 357 (2007) 87–91. A.S. Yu, A.H. Enck, W.I. Lencer, E.E. Schneeberger, Claudin-8 expression in Madin-Darby canine kidney cells augments the paracellular barrier to cation permeation, J. Biol. Chem. 278 (2003) 17350–17359. C.M. Van Itallie, A.S. Fanning, J.M. Anderson, Reversal of charge selectivity in cation or anion-selective epithelial lines by expression of different claudins, Am. J. Physiol. Renal. Physiol. 285 (2003) F1078–F1084. J. Hou, D.L. Paul, D.A. Goodenough, Paracellin-1 and the modulation of ion selectivity of tight junctions, J. Cell. Sci. 118 (2005) 5109–5118. S. Angelow, R. El-Husseini, S.A. Kanzawa, A.S. Yu, Renal localization and function of the tight junction protein, claudin-19, Am. J. Physiol. Renal. Physiol. 293 (2007) F166–F177. M. Konrad, A. Schaller, D. Seelow, A.V. Pandey, S. Waldegger, A. Lesslauer, H. Vitzthum, Y. Suzuki, J.M. Luk, C. Becker, K.P. Schlingmann, M. Schmid, J. Rodriguez-Soriano, G. Ariceta, F. Cano, R. Enriquez, H. Juppner, S.A. Bakkaloglu, M.A. Hediger, S. Gallati, S.C. Neuhauss, P. Nurnberg, S. Weber, Mutations in the tight-junction gene claudin 19 (CLDN19) are associated with renal magnesium wasting, renal failure, and severe ocular involvement, Am. J. Hum. Genet. 79 (2006) 949–957. S.R.K. Vedula, T.S. Lim, P.J. Kausalya, B. Lane, G. Rajagopal, W. Hunziker, C.T. Lim, Quantifying forces mediated by integral tight junction proteins in cell-cell adhesion, Exp. Mech. (in press), doi:10.1007/s11340-007-9113-1. K. Kubota, M. Furuse, H. Sasaki, N. Sonoda, K. Fujita, A. Nagafuchi, S. Tsukita, Ca(2+)-independent cell-adhesion activity of claudins, a family of integral membrane proteins localized at tight junctions, Curr. Biol. (1999) 1035–1038. M. Furuse, K. Furuse, H. Sasaki, S. Tsukita, Conversion of zonulae occludentes from tight to leaky strand type by introducing claudin-2 into Madin-Darby canine kidney I cells, J. Cell. Biol. 153 (2001) 263–272. J.H. Lipschutz, S. Li, A. Arisco, D.F. Balkovetz, Extracellular signal-regulated kinases 1/2 control claudin-2 expression in Madin-Darby canine kidney strain I and II cells, J. Biol. Chem. 280 (2005) 3780–3788. C.M. Van Itallie, J. Holmes, A. Bridges, J.L. Gookin, M.R. Coccaro, W. Proctor, O.R. Colegio, J.M. Anderson, The density of small E XP E RI ME N TA L CE L L RE S E A RCH ( 00 ) –26 [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] tight junction pores varies among cell types and is increased by expression of claudin-2, J. Cell. Sci. 121 (2008) 298–305. J. Hou, A.S. Gomes, D.L. Paul, D.A. Goodenough, Study of claudin function by RNA interference, J. Biol. Chem. 281 (2006) 36117–36123. H. Fujita, K. Sugimoto, S. Inatomi, T. Maeda, M. Osanai, Y. Uchiyama, Y. Yamamoto, T. Wada, T. Kojima, H. Yokozaki, T. Yamashita, S. Kato, N. Sawada, H. Chiba, Tight junction proteins claudin-2 and -12 are critical for vitamin D-dependent Ca2+ absorption between enterocytes, Mol. Biol. Cell. 19 (2008) 1912–1921. D.T. Thwaites, G.T. McEwan, N.L. Simmons, The role of the proton electrochemical gradient in the transepithelial absorption of amino acids by human intestinal Caco-2 cell monolayers, J. Membr. Biol. 145 (1995) 245–256. F.H. Leibach, V. Ganapathy, Peptide transporters in the intestine and the kidney, Annu. Rev. Nutr. 16 (1996) 99–119. S. Ferruzza, M.L. Scarino, G. Rotilio, M.R. Ciriolo, P. Santaroni, A.O. Muda, Y. Sambuy, Copper treatment alters the permeability of tight junctions in cultured human intestinal Caco-2 cells, Am. J. Physiol. 277 (1999) G1138–G1148. W. Hanley, O. McCarty, S. Jadhav, Y. Tseng, D. Wirtz, K. Konstantopoulos, Single molecule characterization of P-selectin/ligand binding, J. Biol. Chem. 278 (2003) 10556–10561. W.D. Hanley, D. Wirtz, K. Konstantopoulos, Distinct kinetic and mechanical properties govern selectin–leukocyte interactions, J. Cell. Sci. 117 (2004) 2503–2511. P. Panorchan, M.S. Thompson, K.J. Davis, Y. Tseng, K. Konstantopoulos, D. Wirtz, Single-molecule analysis of cadherin-mediated cell–cell adhesion, J. Cell. Sci. 119 (2006) 66–74. M. Benoit, D. Gabriel, G. Gerisch, H.E. Gaub, Discrete interactions in cell adhesion measured by single-molecule force spectroscopy, Nat. Cell. Biol. (2000) 313–317. C.M. Franz, A. Taubenberger, P.H. Puech, D.J. Muller, Studying integrin-mediated cell adhesion at the single-molecule level using AFM force spectroscopy, Sci. STKE 2007 (2007) pl5. D.F. Tees, R.E. Waugh, D.A. Hammer, A microcantilever device to assess the effect of force on the lifetime of selectin– carbohydrate bonds, Biophys. J. 80 (2001) 668–682. S.E. Chesla, P. Selvaraj, C. Zhu, Measuring two-dimensional receptor–ligand binding kinetics by micropipette, Biophys. J. 75 (1998) 1553–1572. S.R. Vedula, T.S. Lim, P.J. Kausalya, W. Hunziker, G. Rajagopal, C.T. Lim, Biophysical approaches for studying the integrity and function of tight junctions, Mol. Cell. Biomech. (2005) 105–123. M. Furuse, H. Sasaki, S. Tsukita, Manner of interaction of heterogeneous claudin species within and between tight junction strands, J. Cell. Biol. 147 (1999) 891–903. G.I. Bell, Models for the specific adhesion of cells to cells, Science 200 (1978) 618–627. E. Evans, K. Ritchie, Dynamic strength of molecular adhesion bonds, Biophys. J. 72 (1997) 1541–1555. Y. Tsukasaki, K. Kitamura, K. Shimizu, A.H. Iwane, Y. Takai, T. Yanagida, Role of multiple bonds between the single [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] 2651 cell adhesion molecules, nectin and cadherin, revealed by high sensitive force measurements, J. Mol. Biol. 367 (2007) 996–1006. S.R. Vedula, T.S. Lim, S. Hui, P.J. Kausalya, E.B. Lane, G. Rajagopal, W. Hunziker, C.T. Lim, Molecular force spectroscopy of homophilic nectin-1 interactions, Biochem. Biophys. Res. Commun. 362 (2007) 886–892. W. Baumgartner, P. Hinterdorfer, W. Ness, A. Raab, D. Vestweber, H. Schindler, D. Drenckhahn, Cadherin interaction probed by atomic force microscopy, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 4005–4010. E. Perret, A. Leung, H. Feracci, E. Evans, Trans-bonded pairs of E-cadherin exhibit a remarkable hierarchy of mechanical strengths, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 16472–16477. C. Gergely, J. Voegel, P. Schaaf, B. Senger, M. Maaloum, J.K. Horber, J. Hemmerle, Unbinding process of adsorbed proteins under external stress studied by atomic force microscopy spectroscopy, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 10802–10807. X. Zhang, E. Wojcikiewicz, V.T. Moy, Force spectroscopy of the leukocyte function-associated antigen-1/intercellular adhesion molecule-1 interaction, Biophys. J. 83 (2002) 2270–2279. L. Gilles, C.R. Vogel, B.L. Ellerbroek, Multigrid preconditioned conjugate-gradient method for large-scale wave-front reconstruction, J. Opt. Soc. Am. A. Opt. Image. Sci. Vis. 19 (2002) 1817–1822. C.M. Van Itallie, J.M. Anderson, Claudins and epithelial paracellular transport, Annu. Rev. Physiol. 68 (2006) 403–429. R. Merkel, P. Nassoy, A. Leung, K. Ritchie, E. Evans, Energy landscapes of receptor–ligand bonds explored with dynamic force spectroscopy, Nature 397 (1999) 50–53. E. Evans, Probing the relation between force–lifetime–and chemistry in single molecular bonds, Annu. Rev. Biophys. Biomol. Struct. 30 (2001) 105–128. K.T. Ferreira, B.S. Hill, The effect of low external pH on properties of the paracellular pathway and junctional structure in isolated frog skin, J. Physiol. 332 (1982) 59–67. K. Fujita, J. Katahira, Y. Horiguchi, N. Sonoda, M. Furuse, S. Tsukita, Clostridium perfringens enterotoxin binds to the second extracellular loop of claudin-3, a tight junction integral membrane protein, FEBS Lett. 476 (2000) 258–261. B. Litkouhi, J. Kwong, C.M. Lo, J.G. Smedley III, B.A. McClane, M. Aponte, Z. Gao, J.L. Sarno, J. Hinners, W.R. Welch, R.S. Berkowitz, S.C. Mok, E.I. Garner, Claudin-4 overexpression in epithelial ovarian cancer is associated with hypomethylation and is a potential target for modulation of tight junction barrier function using a C-terminal fragment of Clostridium perfringens enterotoxin, Neoplasia (New York, N.Y.) (2007) 304–314. M.J. Evans, T. von Hahn, D.M. Tscherne, A.J. Syder, M. Panis, B. Wolk, T. Hatziioannou, J.A. McKeating, P.D. Bieniasz, C.M. Rice, Claudin-1 is a hepatitis C virus co-receptor required for a late step in entry, Nature 446 (2007) 801–805. . reserved. Keywords: Claudin Tight junction Cell–cell adhesion Molecular force spectroscopy Atomic force microscopy pH Introduction Tight junctions (TJs) are the apical most constituents of the intercellular adhesion. dissociation of bounded complex. These kinetic parameters can be further used to construct the geometry of the conceptual energy landscape for the dissociation pathway. To compare the topography of. the topography of the energy landscapes of the dissociation of Cldn2/Cldn2 complexes at different pH conditions, the geometric locations of their bound states were plotted on the same reactive coordinates