Molecular Recognition and Adhesion of Individual DNA Strands Studied by Dynamic Force Microscopy

The development of versatile scanning probe methods such as atomic force microscopy (AFM) makes it today possible to study bio-adhesion on a single molecule level. In this paper, we present AFM-force-spectroscopy experiments on complementary DNA strands. From such experiments, intrinsic thermodynamical properties (energy landscape) of these weak non covalent bonds can be determined. Introduction It has long been known that only molecules with an excess of energy over the average energy of the population can participate in chemical reactions. Accordingly, reactions between ligands and receptors follow pathways (in a virtual energy landscape) that involve the formation of some type of high-energy transition states whose accessibility along a reaction coordinate ultimately controls the rate of the reaction. Until recently, chemists and biologists could only act on molecules if these were present in large quantities. Consequently, scientists could only access macroscopic thermodynamical quantities, e.g. the free energy of complex formation and/or dissociation. Today, instruments offering a high spatial resolution and a sensitivity down to the picoor femto-Newton range allow one to study the adhesion of molecular bonds [1-13]. In particular, a novel type of force spectroscopy, dynamic force microscopy (DFS), has been developed. In a DFS experiment, the dependence of the rupture force on the loading rate is investigated using an atomic force microscope (AFM), a bio-membrane force probe (BFP), or eventually an optical tweezers setup. For a typical DFS experiment using an AFM, a ligand is immobilized on a sharp tip attached to a micro-fabricated cantilever and the receptor is immobilized on a surface. When approaching the surface of the tip a bond may form between ligand and receptor. The bond is then loaded with an increasing force when retracting the surface from the tip. From these measurements, the energy landscape of a single bond can be mapped. This paper is organized as follows: Part one introduces theoretical models that describe a chemical reaction when an external force is used to rupture a complex. Then, DFS experiments on complementary DNA strands are presented and illustrate the main ideas developed in part one. 76 Single Molecules RESEARCH PAPER Single Mol. 2 (2001) 2 Theoretical Background In this section, some thermodynamical models describing the rupture of a single bond will be briefly presented. More details can be found elsewhere [15-18]. Bell first stated that the bond lifetime τ of an energy barrier reads: τ τ ( ) exp / F E xF k T B = − ( ) [ ] 0 0 ∆ (1), where T is the temperature, E0 represents the bond energy (the height of the barrier), F is the external applied force per bond, kB is the Boltzmann constant, ∆x is the distance (projected along the direction of the applied force) between the ground state and the energy barrier (with energy E0), and τ0 is a pre-factor. Eq. (1) states that (i) a bond will rupture after a certain amount of time thanks to thermal fluctuations (ii) application of an external force dramatically changes the time it takes to overcome the energy barrier. Note finally that (1) can be re-written as: k F k F F off off ( ) exp = ( ) 0 (2), where koff is the thermal off-rate of the barrier, and F 0 is a force-scale factor (F=kBT/∆x). An important point is that the most probable force F* needed to overcome an energy barrier should a priori depend on the loading rate, i.e. the velocity in a typical DFS experiment (typical values for velocities are in the range between 10 nm/s and 5000 nm/s). Indeed, when the loading rate decreases, F* should decreases because of thermal fluctuations. In fact, a simple relation holds between F* and the loading rate r (r=kν, where k is the stiffness of the DFS force sensor and ν is the retraction speed): F F r F koff * ln = ( ) 0 0 (3), By plotting F* as a function of ln(r), one should therefore find different linear regimes, each of them corresponding to a specific region (a specific energy barrier) of the energy landscape. According to Evans [17], the kinetics runs as follows: application of an external force (i) selects a specific path (a reaction coordinate) in the energy landscape (ii) suppresses outer barriers (Eq. 1) and reveal inner barriers which start to govern the process. For instance, recent BFP and AFM experiments have revealed an intermediate state for the streptavidin (or avidin)-biotin complex [10, 13]. However, since each energy barrier defines a time-scale (a range of loading rate that has to be compatible with the time-scale of the experiment) only a specific part of the energy landscape can be mapped in a typical DFS experiment [18 19]. Experimental DFS measurements were performed using a commercial AFM instrument (Nanoscope III, Digital Instruments, Santa Barbara). The spring constants of all cantilevers (ranging from 12 to 17 pN/nm) were calibrated by the thermal fluctuation method [20] with an absolute uncertainty of 20 %. For the temperature measurements presented below, the temperature was controlled using a home built cell where the buffer solution that immersed both the probe surface and the AFM cantilever was in contact with a Peltier element (Melcor, Trenton, NJ), driven with a constant current source. Measurements at different points of the cell showed deviations of less than 2 °C. All chemicals were purchased from Fluka unless noted elsewhere. The preparation and immobilization of all oligonucleotides follow the protocol described in ref. 11. Fig. 1. A typical probability distribution for the rupture force (about 500 approach/retract cycles, retract velocity 100 nm/s) [11]. For this experiment, an oligomer a (see text) was attached to the tip of the AFM-cantilever and its complement b was immobilized on the surface (complements were pulling apart at their opposite 5’-ends). Gray rectangles (a against b), black rectangles (a against a). To minimize unspecific interactions and multiple unbinding events, 30-nm-long PEG linkers were attached to the 5’ends. Note that the scale-force F can be in principle determined from the width of the distribution.