logo

Partenaires

CNRS
ENS Departement de Physique ENS
UPMC
CdF
Univ. Evry

Rechercher

Sur ce site

Sur le Web du CNRS

Accueil du site

Single molecule observation of restriction enzyme interaction with DNA by fluorescence microscopy

par Maxime Dahan - 23 octobre 2007

Restriction enzymes display important tools in microbiology since they cut double stranded DNA at specific sequences 4-7 bp long. A remarkable, yet not fully understood feature is that they can find target sites among long, non-specific sequences at very high speed without consuming energy1,2. To explain this so-called "facilitated diffusion", two different mechanisms for enzyme translocation along the DNA have been proposed : "Sliding", in which the enzyme is bound non-specifically to the DNA and performs a restricted 1D diffusion along it while probing for the target site.3 "Jumping" involves detachment of the enzyme from the DNA, followed by a brief 3D diffusional translocation and rebinding to the DNA at a remote site from the starting sequence.4,5 We investigate "facilitated diffusion" of the restriction enzyme EcoRV6 by using a fluorescently labeled restriction enzyme which can be imaged by Total Internal Reflection Fluorescence Microscopy (TIRFM) while interacting with a stretched DNA on the surface. In contrast to ensemble kinetic experiments, our approach has the advantage that we can directly visualize the sliding separated from DNA cutting since the employed DNA has no target site for the restriction enzyme. For this purpose, we collaborate with a group of biochemists (A. Pingoud and W. Wende, Giessen, Germany) who produce modified EcoRV enzymes that can be specifially labeled without interference with enzyme activity.7 We use a DNA construct modified at both ends with multiple biotins which does not contain a recognition site for EcoRV (GATATC). This can be attached to the streptavidin-coated bottom of a flow cell, and is stretched by the application of a flow, resulting in a DNA which is bound to the surface only at its ends. After localization of DNA molecules by a DNA staining dye, labeled enzyme solution is inserted into the cell. The fluorescence light emitted from enzymes is filtered and collected by a highly sensitive EMCCD camera, achieving spatial and time resolution of about 40 nm and 20 ms, respectively. From single enzyme trajectories, the diffusion is analyzed by calculating the mean square displacement (MSD) of the enzyme over the time. For a diffusion process, this should display a linear dependence on time in the direction of the stretching (x). In transversal direction, however, the motion is confined to the DNA fluctuation without net displacement, resulting in a constant MSD.

Figure 1 : a)Scheme of experimental set-up : DNA molecules (dark green) are stretched on the surface and labelled enzymes inserted into the flow cell (red circles). The bottom of the flow cell is placed on a microscope objective which is used to excite and image the fluorescent enzyme interacting with the stretched DNA via TIRFM. b) Fluorescence image of the stretched DNA, visualized by hundreds of interacting EcoRV enzymes whereas the DNA itself is not stained. The DNA is stretched to about 70% of its contour length ( 2µm). The bright spot in the upper left corner stems from enzymes on a DNA which is bound to the surface only by one end.

Different aspects of investigation :

To elucidate the mechanism of facilitated diffusion of the EcoRV system, we determine the dependence of the DNA interactions on various external conditions : • Dependence on buffer reagents A commonly neglected effect is the nature of the ionic groups present in solution. Thus, it has been shown that the restriction rate of the same DNA changes with different buffer reagents even if pH and ionic strength are comparable. This is the more important since "native" buffer conditions, i.e. the cell environment are difficult to reproduce in in vitro experiments. • Dependence on effective size of the enzyme The question of the nature of sliding is not limited to the typical base pair length screened during a single event. Usually, it has been assumed that during the 1D confinement of sliding, the enzyme screens the base pairs by following the shape of the major groove in a rotational motion. However, there exists so far no experimental evidence for this hypothesis. We tackle this problem by labeling the enzyme with fluorophores of different sizes. Thus, the enzyme is either labeled with a fluorophore, a fluorophore-labeled protein or a quantum dots. Since the resulting enzyme constructs display large differences in the hydrodynamic radius, the different diffusion rates of the constructs can give an answer about the sliding movement. • First passage problem The question of facilitated diffusion also concerns the specific recognition of the target sequence, i.e. is it recognized instantly or has the site to be screened several times. This can be studied by following a labeled enzyme motion along a stretched DNA incorporating one or more specific sequences at known positions. The permanent stalling of the enzyme motion at the correct position than indicates target recognition. To induce target recognition, but to prevent cleavage which would lead to enzyme release, Ca-ions in the buffer can be used. This can be further expanded by using a DNA with several sites at different DNA positions to determine if target localization efficiency depends on the site position.

Refinement of experimental approach by application of double optical tweezers

To overcome limitations of the present experimental approach, we built a set-up combining single molecule fluorescence detection with double optical tweezers. By attaching the DNA to two transparent beads in the µm-size range, the DNA can thus not only be removed from the surface, but also the DNA stretching can be controlled in a very precise manner. Furthermore, the stretching of the DNA can be simply controlled by white light imaging of the beads circumventing the necessity of further DNA staining. The disadvantage that because of the increased surface distance less sensitive detection in the epifluorescence mode has to be applied can be compensated by the use of enzyme labeled with quantum dots. If the DNA is modified at both ends with different binding groups (biotin and digoxigenin), and beads of different sizes which each bear a different binding protein (streptavidin and anti-digoxigenin), the orientation of the DNA can be determined in addition. This application is specifically useful when the interaction with a specific site dislocated from the DNA center is investigated.

Figure 2 : Scheme of combining optical trapping with fluorescence : The DNA strand is attached to two beads which are held by independent double optical tweezers, allowing for a controlled DNA stretching. The interacting enzymes are again imaged by fluorescence labelling.

Contacts : Pierre Desbiolles, Andreas Biebricher, Pierre-Louis Porte.

References :

1. O. G. Berg, R. B. Winter, and P. H. von Hippel, Biochemistry 20 (24), 6929 (1981). 2. PH von Hippel and OG Berg, J. Biol. Chem. 264 (2), 675 (1989). 3. Nobuo Shimamoto, J. Biol. Chem. 274 (22), 15293 (1999). 4. N. P. Stanford, M. D. Szczelkun, J. F. Marko, and S. E. Halford, Embo J 19 (23), 6546 (2000). 5. D. M. Gowers and S. E. Halford, Embo J 22 (6), 1410 (2003). 6. A. Pingoud and A. Jeltsch, Nucleic Acids Res 29 (18), 3705 (2001). 7. C. Schulze, A. Jeltsch, I. Franke, C. Urbanke, A. Pingoud, Embo J. 17, 6757 (1998).

Dans la même rubrique :