DNA under Hydrodynamic and Mechanical Stretching : Structure and Dynamics

In Chapter 2 of this thesis the movement of individual tethered λ-DNA molecules is studied in the absence and in the presence of oscillatory shear ow using time-dependent uorescence microscopy. In the absence of shear the root mean square uctuation of the center-of-mass position of the molecules, the radius of gyration, and the longest relaxation time of end-grafted DNA molecules are determined. The agreement of the measured radius of gyration in the absence of shear, Rg = (0.85± 0.05) µm, close to literature values for Rg of λ-phage DNA in a bulk solvent, with the predictions of the Zimm model for a free polymer in a good solvent strongly hints at the importance of hydrodynamic interactions between the monomers. From the autocorrelation functions of the movement of individual DNA molecules the longest relaxation time of end-grafted DNA molecules is determined to be τ = (0.79 0.03) s. In a narrow frequency range 1 HzFor DNA molecules under oscillatory shear ow the relaxation time for the movement perpendicular to the shear, averaged over all experimentally accessible values of shear rate and shear frequency, is found to deviate only slightly from the value in the absence of shear. For the movement of the DNA molecules in the direction of the shear, on the other hand, the relaxation times are found (a) to extrapolate to a value close to the relaxation time in the absence of shear, as expected, and (b) to slightly decrease with shear.Comparing the amplitude of the DNA movement in the driven direction with the predictions of a simple bead spring model di erent values are obtained for the distance of the grafting point of the DNA molecules from the wall in the low and high-frequency ranges, respectively. The discrepancy between the two numbers is attributed to a de ciency of the bead spring model. Furthermore it is pointed out that, in the case of oscillatory shear ow, and in contrast to the case of steady shear ow, even in the simple bead spring model the response of the tethered polymer depends not only on the Weissenberg number, but also explicitly on the dimensionless frequency ωτ, which may be identi ed, up to a constant with the Deborah number, well known in the rheology of non-Newtonian liquids.The observed amplitudes of tethered DNA molecules in oscillatory shear, normalized to the shear and to the observed non-constant values of zc, are predicted by the bead spring model to depend only on the dimensionless frequency. The normalized amplitudes observed in the experiments follow the predictions of the bead spring model at low frequencies, but exhibit a drastic enhancement over the predictions of the model, of more than a factor of two, at around ωτ = 2. In the frequency range ωτ > 10, on the other hand, the amplitudes are smaller than predicted by the model. Because the shear-induced stress corresponds to the entropic elasticity range at low forces in single molecule stretching experiments, the observed enhancement around ωτ ≈ 2 and also the reduction at ωτ > 10 are tentatively assigned to hydrodynamic monomer interactions, which are intrinsically nonlinear, even in the low shear range.The construction of a micro- uidic device containing patterned gold thin lms and initial uorescence microscopy experiments involving DNA molecules tethered to the gold islands, as described in Chapter 3, lay the ground for a study, along similar lines as in Chapter 2, of the hydrodynamic interactions between two individual grafted DNA molecules at well-de ned distance to each other, in the absence of shear, as well in a steady shear ow. The steps necessary for the construction of the device, including the preparation of the PDMS micro-channel and the gold islands, and grafting of DNA to the latter are complex and considerable technical di culties had to be overcome to achieve single occupation of neighboring gold islands by DNA molecules.An attempt to study the nature of the overstretched state of DNA molecules in highly oriented lms of DNA in the B form by x-ray di raction is described in Chapter 4. It is shown that, in contrast to single molecule stretching experiments, the majority of the DNA molecules remains in the B form up to breakage of the lm. At the same time a meridional re ection at qz 0:8 Å-1, corresponding to a periodicity of a = 7:8 Å , is observed, which may be explained by the existence of part of the molecules in a maximally stretched state, as predicted by earlier simulations. This di raction peak increases somewhat in intensity with applied stress, but is already present in the lms before stretching. This may be explained by assuming that in the wet-spinning process part of the molecules are already overstretched, whereas most of the molecules remain in the B form. Because the sti ness of maximally stretched DNA will be determined by the sti ness of the rigid backbone and therefore be much larger than that of B-DNA, it appears likely that DNA molecules maximally stretched during wet-spinning support most of the externally applied stress in the lms and breakage of these molecules marks breakage of the lm. Melting of the ds-DNA, which is currently a widely favored model for the behavior of DNA under large stresses, as an alternative to various proposed forms of S-DNA, should result in a disappearance of the x-ray re ections typical for the B form and, possibly, to a plateau in the force-extension curve of the lm similar to that observed in single molecule stretching experiments using force microscopy or optical tweezers. Both are not observed for the present lms. The stretching of molecules in the lms studied here is therefore very inhomogeneous and obviously fundamentally di erent from the case of single molecule stretching experiments