Mode of Heavy Meromyosin Adsorption and Motor Function

Langmuir 2007, 23, 11147-11156

11147

Mode of Heavy Meromyosin Adsorption and Motor Function Correlated with Surface Hydrophobicity and Charge Nuria Albet-Torres,† John O’Mahony,† Christy Charlton,† Martina Balaz,† Patricia Lisboa,‡ Teodor Aastrup,§ Alf Månsson,*,† and Ian A. Nicholls*,† Department of Chemistry and Biomedical Sciences, UniVersity of Kalmar, SE-391 82 Kalmar, Sweden, European Commission - Joint Research Centre, Institute for Health and Consumer Protection, I-21020 Ispra (VA), Italy, and Attana AB, Bjo¨rnna¨sV. 21, SE-113 47 Stockholm, Sweden ReceiVed March 25, 2007. In Final Form: June 21, 2007 The in vitro motility assay is valuable for fundamental studies of actomyosin function and has recently been combined with nanostructuring techniques for the development of nanotechnological applications. However, the limited understanding of the interaction mechanisms between myosin motor fragments (heavy meromyosin, HMM) and artificial surfaces hampers the development as well as the interpretation of fundamental studies. Here we elucidate the HMM-surface interaction mechanisms for a range of negatively charged surfaces (silanized glass and SiO2), which is relevant both to nanotechnology and fundamental studies. The results show that the HMM-propelled actin filament sliding speed (after a single injection of HMM, 120 µg/mL) increased with the contact angle of the surfaces (in the range of 20-80°). However, quartz crystal microbalance (QCM) studies suggested a reduction in the adsorption of HMM (with coupled water) under these conditions. This result and actin filament binding data, together with previous measurements of the HMM density (Sundberg, M.; Balaz, M.; Bunk, R.; Rosengren-Holmberg, J. P.; Montelius, L.; Nicholls, I. A.; Omling, P.; Tågerud, S.; Månsson, A. Langmuir 2006, 22, 7302-7312. Balaz, M.; Sundberg, M.; Persson, M.; Kvassman, J.; Månsson, A. Biochemistry 2007, 46, 7233-7251), are consistent with (1) an HMM monolayer and (2) different HMM configurations at different contact angles of the surface. More specifically, the QCM and in vitro motility assay data are consistent with a model where the molecules are adsorbed either via their flexible C-terminal tail part (HMMC) or via their positively charged N-terminal motor domain (HMMN) without other surface contact points. Measurements of ζ potentials suggest that an increased contact angle is correlated with a reduced negative charge of the surfaces. As a consequence, the HMMC configuration would be the dominant configuration at high contact angles but would be supplemented with electrostatically adsorbed HMM molecules (HMMN configuration) at low contact angles. This would explain the higher initial HMM adsorption (from probability arguments) under the latter conditions. Furthermore, because the HMMN mode would have no actin binding it would also account for the lower sliding velocity at low contact angles. The results are compared to previous studies of the microtubule-kinesin system and are also discussed in relation to fundamental studies of actomyosin and nanotechnological developments and applications.

Introduction Muscle contraction is caused by ATP-driven cyclic interactions between globular units of myosin II molecules and actin filaments.1,2 This biomolecular function can be reconstructed in vitro using an assay procedure (the in vitro motility assay), which permits the direct observation of actin filament motility.3,4 The assay entails the adsorption of myosin motors, or preferably heavy meromyosin (HMM) motor fragments, to a suitably functionalized surface. Fluorophore-labeled actin filaments are translated over these surfaces by the myosin motors in the presence of ATP. Despite the widespread use of the assay, relatively few attempts5-7 to investigate the nature of the critical HMM-surface interactions have been undertaken. In one recent study,7 HMM * Corresponding authors. (A.M.) E-mail: [email protected]. Tel: +46480-446243. Fax: +46-480-446244. (I.A.N.) E-mail: [email protected]. Tel: +46-480-446258. Fax: +46-480-446244. † University of Kalmar. ‡ European Commission - Joint Research Centre. § Attana AB. (1) Huxley, A. F. Prog. Biophys. Biophys. Chem. 1957, 7, 255-318. (2) Geeves, M. A.; Fedorov, R.; Manstein, D. J. Cell. Mol. Life Sci. 2005, 62, 1462-1477. (3) Kron, S. J.; Spudich, J. A. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 62726276. (4) Kron, S. J.; Toyoshima, Y. Y.; Uyeda, T. Q.; Spudich, J. A. Methods Enzymol. 1991, 196, 399-416. (5) Toyoshima, Y. Y. AdV. Exp. Med. Biol. 1993, 332, 259-265. (6) Balaz, M.; Sundberg, M.; Persson, M.; Kvassman, J.; Månsson, A. Biochemistry 2007, 46, 7233-7251.

was adsorbed to underivatized silicon dioxide or glass surfaces or to such surfaces derivatized by trimethylchlorosilane (TMCS). The sliding velocities were low or zero on the hydrophilic surfaces but were consistently high on the more hydrophobic TMCS and indicated a possible correlation between the sliding velocity and contact angle of the HMM-adsorbing surface. A more detailed account of the relationship between contact angle and sliding velocity, together with an elucidation of the underlying mechanisms, would be of importance to a number of areas, for example, for fundamental biophysical studies of actomyosin function7,8 and for efforts to achieve spatial control of motor function using nanopatterned and micropatterned surfaces.9,10 (See also ref 11.) Such control of function has been proposed to hold considerable potential for applications in nanotechnology.12-15 (7) Sundberg, M.; Balaz, M.; Bunk, R.; Rosengren-Holmberg, J. P.; Montelius, L.; Nicholls, I. A.; Omling, P.; Tagerud, S.; Mansson, A. Langmuir 2006, 22, 7302-7312. (8) Månsson, A.; Nicholls, I. A.; Omling, P.; Tågerud, S.; Montelius, L. Nanotechnology Enhanced Functional Assays of Actomyosin Motility - Potentials and Challenges. In Controlled Nanoscale Motion; Nobel Symposium 131; Linke, H., Ma˚nssom, A., Eds.; Lecture Notes in Physics 711; Springer-Verlag: Berlin, 2007; pp 385-406. (9) Clemmens, J.; Hess, H.; Lipscomb, R.; Hanein, Y.; Bohringer, K. F.; Matzke, C. M.; Bachand, G. D.; Bunker, B. C.; Vogel, V. Langmuir 2003, 19, 1096710974. (10) Sundberg, M.; Bunk, R.; Albet-Torres, N.; Kvennefors, A.; Persson, F.; Montelius, L.; Nicholls, I. A.; Ghatnekar-Nilsson, S.; Omling, P.; Tagerud, S.; Mansson, A. Langmuir 2006, 22, 7286-7295. (11) Fischer, T.; Hess, H. J. Mater. Chem. 2007, 17, 943-951.

10.1021/la7008682 CCC: $37.00 © 2007 American Chemical Society Published on Web 08/15/2007

11148 Langmuir, Vol. 23, No. 22, 2007

Previous fluorescence spectroscopic studies7 suggested that there was higher HMM density on hydrophilic than on hydrophobic surfaces. This observation implied that the poor actomyosin motility on the more hydrophilic surfaces is attributed to a different mode of HMM adsorption. Further results from studies using total internal reflection fluorescence (TIRF) spectroscopy7 suggested considerably lower binding of actin to HMM on a hydrophilic substrate than in the case of a hydrophobic substrate. However, the interpretation of the fluorescence data was subject to some uncertainty because there was evidence from fluorescence microscopy7 that the fluorescence intensity of the labeled actin filaments was quenched on SiO2. Here we provide the first detailed study correlating HMMinduced sliding velocity in vitro with several physical-chemical properties of surfaces including the contact angle, surface charge, and rms roughness. Glass surfaces, either nonderivatized or derivatized with one of six different trialkylchlorosilanes, were used to test the hypothesis7 that there is a correlation between contact angle and actin filament sliding velocity on negatively charged surfaces. The span of contact angles was extended to include angles lower than in our previous work. Furthermore, atomic force microscopy was employed as a tool for the screening of surfaces to allow us to take any effects of significant roughness into account. In addition, ζ-potential measurements were used directly to establish differences in the degree of negative charge of SiO2 and silanized surfaces and correlate these changes with the contact angles. The quartz crystal microbalance (QCM) technique16-18 has been used in only one previous study of the actomyosin system, then investigating adsorption to poly(tert-butyl methacrylate),19 using similar HMM concentrations as in this study but higher actin concentrations. This technique assesses mass deposition and is not subject to the same types of uncertainties in interpretation as are inherent to the TIRF method. Here we used QCM studies to examine HMM adsorption and subsequent actin binding to SiO2- and TMCS-derivatized surfaces. Collectively, the data suggest the refinement of a previously proposed model7 (Figure 1) for HMM adsorption to negatively charged surfaces with different contact angles. The data also extend previous work19 showing the utility of QCM as a tool for the study of several aspects of the in vitro motility assay and actomyosin function in general. We compare our data with results from previous studies of the kinesin-microtubule motor system and discuss the implication of the results for fundamental studies of actomyosin function. We also discuss the implications for spatial control and the possible development of on/off switches for contractility in nanotechnological applications. Furthermore, on the basis of the more refined correlation between contact angle and sliding velocity we propose that the sliding velocity of actin filaments may be an important tool for the characterization of the surface chemistry of nanoscale surface domains. Presently, there are relatively few such methods. (12) Mansson, A.; Sundberg, M.; Bunk, R.; Balaz, M.; Nicholls, I. A.; Omling, P.; Tegenfeldt, J. O.; Tagerud, S.; Montelius, L. IEEE Trans. AdV. Packag. 2005, 28, 547-555. (13) Bakewell, D. J. G.; Nicolau, D. V. Aust. J. Chem. 2007, 60, 314-332. (14) Hess, H.; Clemmens, J.; Qin, D.; Howard, J.; Vogel, V. Nano Lett. 2001, 1, 235-239. (15) van den Heuvel, M. G.; de Graaff, M. P.; Dekker, C. Science 2006, 312, 910-914. (16) Marx, K. A. Biomacromolecules 2003, 4, 1099-1120. (17) O’Sullivan, C. K.; Guilbault, G. G. Biosens. Bioelectron. 1999, 14, 663670. (18) Janshoff, A.; Galla, H. J.; Steinem, C. Angew. Chem., Int. Ed. 2000, 39, 4004-4032. (19) Hanson, K. L.; Viidyanathan, V.; Nicolau, D. V. Proc. SPIE 2006, 6036, 1-8.

Albet-Torres et al.

Figure 1. (A) Model for HMM adsorption onto SiO2 (left; strongly negatively charged) and the TMCS-derivatized surface (right; hydrophobic and low negative charge). On the TMCS-derivatized surface, HMM is assumed to adsorb only via the C-terminal end of the tail segment (configuration 1). In addition to this configuration, two other configurations (2 and 3), involving adsorption via the myosin head, were believed to be possible on this surface chemistry mainly as a result of electrostatic interactions between the myosin head and the SiO2 surface. As illustrated by the short attached actin filament segments to the right, only configuration 1 is assumed to be available for the binding of actin filaments. (B) Schematic illustration of the overall charge distribution within the HMM molecule (cf. refs 37-39 and 44 and search of Swiss Prot-TrEMBL, entry name MYH4 (myosin heavy chain 2b)). Note the negatively charged S2 domain, negatively charged neck domain with essential and regulatory light chains, and positively charged head domain, particularly in the actin-binding region (e.g., surface loops between residues 626 and 647 and between residues 567 and 578).37

A preliminary account of this work has been presented previously.20 Materials and Methods Materials. Trimethylchlorosilane, triethylchlorosilane (TECS), tripropylchlorosilane (TPCS), tributylchlorosilane (TBCS), phenyldimethylchlorosilane (PDMCS), and diphenylmethylchlorosilane (DPMCS) were purchased from Sigma-Aldrich (Germany). Concentrated sulfuric acid was obtained from Merck (KGaA Darmstadt, Germany), and 30% hydrogen peroxide was purchased from Fluka (Seelze, Germany). Chemicals for in vitro motility assays were of analytical grade and, unless otherwise stated, purchased from SigmaAldrich (St Louis, MO), Fluka (Germany), or Merck (Germany). Tetramethyl rhodamineisothiocyanate phalloidin (RhPh) for fluorescence labeling of actin filaments was purchased from InvitrogenMolecular Probes (Eugene, OR). Deionized water was used as a solvent for QCM analysis. QCM analysis was performed on both Attana 80 and Attana 100 QCM instruments (Attana AB, Stockholm, Sweden). Each instrument contained a 5 µL flow cell. Silicon dioxidecoated QCM chips (10 MHz resonance) were used in the analysis (also purchased from Attana). For motility measurements, 24 × 60 mm2 cover slips from Schott AG (Mainz, Germany) were used. They were made of borosilicate glass from Schott AG (Mainz, Germany) with a SiO2 content corresponding to 64% of the total mass (Schott D263M). Preparation of QCM Chips for Analysis. Attana silicon dioxidecoated resonator chips were dipped in 10 mL of freshly prepared piranha solution (7:3 concd H2SO4/30% H2O2) for 5 min and rinsed with deionized water. (Caution! Piranha solution is a highly corrosiVe acidic solution that can react Violently with organic materials. Do not store in a closed container, and take appropriate safety precautions.) QCM chips were placed in water prior to analysis to ensure good wetting prior to measurements. Silanized QCM chips were prepared by following the procedure outlined below for the production of silanized glass surfaces. Preparation of Silanized Glass Surfaces for Motility Assays. The glass slides were first cleaned by immersion in a bath of piranha solution at 80 °C for 5 min. Subsequent washing steps were immersion in deionized water (30 s) twice, followed by dry acetone, methanol, (20) Månsson, A.; Sundberg, M.; Balaz, M.; Albet-Torres, N.; Vikhorev, P.; O’Mahony, J.; Charlton, C.; Kvennefors, A.; Ghatnekar-Nilsson, S. G.; Bunk, R.; Tågerud, S.; Nicholls, I. A.; Omling, P.; Heidari, B.; Montelius, L. J. Muscle Res. Cell Motil. 2006, 27, 482.

HeaVy Meromyosin-Surface Binding and distilled chloroform. The slide was then immersed in a 5% solution of the silane in chloroform for 30 s. Slides were then rinsed with freshly distilled chloroform and were dried under a stream of dry N2. In all cases, blank slides were included. These were cleaned as above but without the silanization step. After cleaning and/or silanization, the surfaces were stored under ambient conditions (or, in some cases, under deionized water) for up to 16 h before contact angle measurements were made. Contact Angle Measurements. Following the preparation of the glass slides or QCM chips, contact angle measurements were generally performed to examine the properties of the surfaces.7 Droplets of deionized water were pumped onto the surface, and advancing contact angles (triplicate determinations) were recorded with a Nikon Coolpix 4500 (Nikon Corp., Tokyo, Japan) camera and processed using a Matlab algorithm (MathWorks, Natick, MA). After the droplets had expanded sufficiently, the pump was reversed, and receding contact angles were also recorded (triplicate determinations). Measurements of ζ Potentials. The measurements were carried out using an electrokinetic analyzer (EKA, model EKS100, Anton Paar KG, Graz, Austria) by the streaming potential method. The values required for the calculation of the ζ potential, such as pH, conductivity, and temperature, were determined using external electrodes and a temperature sensor. VisioLab for EKA evaluation software used the Fairbrother-Mastin21 approach. Because the sample is in contact with an insert made of PMMA, the measured ζ potential is also derived from PMMA. The potential of the sample can thus be calculated according to ζsurface ) 2ζmeasured - ζPMMA The ζ-potential measurements were performed with 450 mL of a 1 mM KCl (Fluka) solution between approximately pH 2 and 10. The employed automatic titration program measured the ζ potential in relation to pH using an automatic titration unit equipped with a 1 mL syringe. For measuring relevant quantities (e.g., pH, conductivity, and temperature), a pressure ramp program was applied where the pressure did not exceed 600 mbar. The duration for each pressure ramp, which consists of two single pressure ramps, one in each flow direction, was 120 s in order to reach electrochemical equilibrium. This step was repeated twice to provide a statistical basis for the comparison of results (i.e., every measuring point was the average of four single points). Protein Preparations and In Vitro Motility Assays. Heavy meromyosin (HMM) was obtained by chymotryptic cleavage of myosin isolated from rabbit fast skeletal muscle.4,22 Actin was obtained from rabbit muscle,22,23 and the actin filaments were labeled with tetramethylrhodamineisothiocyanate-phalloidin22 (Rh-Ph). The motility assay was performed essentially as described previously.7 A wash solution (buffer A; 25 mM imidazole hydrochloride at pH 7.4, 4 mM MgCl2, 1 mM EGTA, 25 mM KCl; ionic strength 70 mM) was used for several rinsing steps and for diluting HMM to its final incubation concentration. Furthermore, we employed two different ATP-containing solutions as follows: The a40 assay solution (ionic strength 40 mM) contained 3-(N-morpholino)-propanesulfonic acid (10 mM, pH 7.4), Na2ATP (1 mM), MgCl2 (2 mM), KCl (25 mM), EGTA (0.1 mM), DTT (10 mM), and an antibleach mixture (20 units/mL glucose oxidase, 3 mg/mL glucose, and 920 units/mL catalase). The aMC130 assay solution (ionic strength 130 mM) contained the same substances as the a40 solution, but 115 mM instead of 25 mM KCl, and also contained 0.64% methylcellulose. Flow cells, constructed from one silanized and one nonsilanized cover glass (after measurement of the contact angle), were incubated as follows: heavy meromyosin (HMM, 120 µg/mL, 2 min) followed by blocking actin (unlabeled actin filaments used to block nonfunctional HMM heads, 1 min of incubation), a40 assay solution, (21) Elimelech, M.; Chen, W. H.; Waypa, J. J. Desalination 1994, 95, 269286. (22) Klinth, J.; Arner, A.; Mansson, A. J. Muscle Res. Cell Motil. 2003, 24, 15-32. (23) Pardee, J. D.; Spudich, J. A. Methods Cell Biol. 1982, 24, 271-289.

Langmuir, Vol. 23, No. 22, 2007 11149 buffer A, 2 nM Rh-Ph-labeled actin (30 s), buffer A again, and finally, aMC130 solution for the observation of actin filament sliding. As suggested by one control experiment using nitrocellulose- and TMCS-derivatized surfaces, the omission of a preincubation step with bovine serum albumin (BSA)7 had no effect on motility quality (sliding velocity and fraction of motile filaments) in the in vitro motility assay. In contrast, another control experiment suggested that the motility quality on these surfaces was considerably impaired by the omission of preincubation with unlabeled blocking actin. The fluorescently labeled actin filaments were imaged using an inverted epi-fluorescence microscope (Nikon Eclipse TE300), and data were recorded with a CCD camera as described previously.7 Motility assays were performed at 28-31 °C with a maximum variation of (0.9 °C within a given experiment. QCM Experiments. Degassed buffer A was pumped through the QCM flow cell (volume 5 µL) at a rate of 0.045 mL/min to obtain a stable baseline level before the first HMM injection. Continuous flow at this rate was then maintained during the entire experiment with brief interruptions during injections of protein or ATP-containing buffer. Each HMM injection corresponded to 12 flow cell volumes at 120 µg/mL (flow rate 0.045 mL/min). To ensure complete coverage of the chip surface, HMM injections were repeated for up to 10 000 s or until no further apparent change in sensor response was observed. A similar procedure was repeated with the injection of actin filaments to ensure saturation of the binding sites on surface-attached HMM molecules. Finally, ATP-containing buffer was injected to release bound actin filaments. The changes in QCM sensor frequency, ∆f, can be related to the adsorbed mass, ∆ms, according to the Sauerbrey relation: ∆ms )

- ANFq ∆f f02

(1)

Here A is the area of the electrode (0.160 cm2), N is the frequency constant (1670 × 102 Hz cm), Fq is the density of quartz (2.65 g/mL), and f0 is the fundamental frequency of the crystal (10 MHz). AFM Experiments. To ascertain the topographical similarity of the glass and the different silanized surfaces, an atomic force microscope (Edu-Scope, nQuip AB, Lund, Sweden) study was performed. The AFM images were obtained in contact mode, and at least three different points were imaged on each surface (glass, TMCS, TECS, TPCS, TBCS, DPMCS, and PDMCS). Some analyses were performed using WSxM 4.0 Develop 8.5 software (Nanotec Electronica, Spain). The surface roughness was established as a root-mean-square (rms) value based on the different height frequencies of the analyzed image. The AFM images (256 pixels × 256 pixels) used for the analysis of surface roughness were selected from larger areas of 5 × 5 µm2. Posterior analysis of roughness was done after focusing on an area of 256 × 256 nm2 and avoiding large surface features (Supporting information for further details). Such analysis was performed on 5-7 surfaces of each type from two different batches. Modeling of QCM Data. The QCM data were simulated using a simple kinetic model described in Supporting information. The systems of differential equations that governed this model were solved by numerical integration using the fourth-order RungeKutta-Fehlberg method implemented in the Simnon software (v. 1.3; SSPA, Gothenburg, Sweden). Data Analysis and Statistics. Motility assay data (sliding velocity, fraction of motile filaments) was analyzed using a computer program developed in the MATLAB environment.24 Curve fittings and graphic representations (mean ( standard error of the mean; SEM) were performed using GraphPad Prism software (GraphPad Software, San Diego, CA). Statistical hypothesis testing was performed using the two-tailed t test with the Welch correction for unequal variances if required. In the case of multiple comparisons, these were performed using one-way analysis of variance (ANOVA) followed by Dunnett’s post hoc test. (24) Mansson, A.; Tagerud, S. Anal. Biochem. 2003, 314, 281-293.

11150 Langmuir, Vol. 23, No. 22, 2007

Albet-Torres et al.

Figure 2. (A) Advancing contact angles (open bars), receding contact angles (gray bars), and rms roughness (black bars; right vertical axis) for piranha-cleaned glass and different silanized surfaces. Contact angle data are based on five different surface batches unless otherwise stated in parentheses. Roughness data are based on measurements using five to seven surfaces from two different batches. The error bars reflect the standard error of the mean (SEM). There are statistically significant differences between glass and other surfaces {*** (p < 0.001), ** (p < 0.01), * (p < 0.05)} and between TMCS and other surfaces {### (p < 0.001), ## (p < 0.01), # (p < 0.05)}. (B) ζ potential (at pH 7) plotted against contact angle. The data refer to piranha-cleaned pure glass (three batches) and silanized cover slips: TMCS (three batches), TBCS (two batches), and one surface each for TPCS, PDMCS, and DPMCS as indicated by color coding. (C) Plots of ζ potential vs pH for several of the surfaces in B (using the same color scheme). Representative data for TMCS and glass surfaces.

Results Characterization of Surfaces. Contact angle measurements were performed on surfaces from each of the studied batches of silanized and nonderivatized cover slips. The results are summarized in Figure 2. Pure glass exhibited the lowest contact angle, followed by surfaces derivatized with TBCS, TPCS, and TECS. These surfaces (including glass) exhibited an advancing contact angle that was significantly different from that on TMCS. (See results from ANOVA in Figure 2.) Furthermore, these surfaces exhibited the largest variability between experiments and also the largest contact angle hysteresis (difference between advancing and receding contact angles). A third group of surfaces, derivatized with TMCS, PDMCS, and DPMCS, exhibited the highest contact angles together with the lowest variability between experiments and a low contact angle hysteresis. The advancing contact angles for these surfaces were not significantly different but were different from that of glass (p < 0.001). The advancing contact angle of the TMCS surfaces (75.1 ( 1.3°; n ) 10 surface batches) was similar to that previously reported for TMCS (7681°)7,25 and standard nitrocellulose surfaces (76°).25 Separate control experiments showed that storing the silanized surfaces for 14 days produced only minor changes in contact angle (