We image macroscopic transient pores in mechanically stretched giant vesicles. Holes open above a critical radius r c1 , grow up to a radius r c2 , and close. We interpret the upper limit r c2 by a relaxation of the membrane tension as the holes expand. The closing of the holes is caused by a further relaxation of the surface tension when the internal liquid leaks out. A dynamic model fits our data for the growth and closure of pores.Opening a hole in a biological membrane has been a challenge for drug delivery and gene therapy. Chemical strategies, based on the addition of a suitable agent (1, 2) [detergent proteins such as talin (3)] and physical means (4) [electroporation (5), osmotic shock (6), temperature jump (4), and adhesion on porous (7) or decorated substrates (8) (A. L. Bernard, M. A. GuedeauBoudeville, O.S., S. Palacin, J. M. diMiglio, and L. Jullien, unpublished data), have been developed to increase membrane permeability.Our approach is to stretch the vesicle membrane: even a weak adhesion or an intense light provides a tension , which relaxes by the formation of transient macroscopic pores. The standard difficulty of visualizing directly at video rate the fast dynamics of pore openings and closings is overcome here by the use of a viscous solvent. The role of solvent viscosity is surprising: under tension, vesicles start to burst, very much like viscous bubbles (9), and this is controlled by lipid viscosity. However, as we shall see, there is a leak-out of the internal liquid, which relaxes the tension and induces the closure of the pore. In a solution of low viscosity (like water), leak-out is fast: the pores close before reaching a visible size. If we make the solvent more viscous, leak-out is slowed down: the pores reach sizes up to 10 m. The immersion of vesicles in a viscous environment allows visualization of transient pores in a membrane stretched by either intense illumination or weak adhesion on a solid substrate.For clarity, we begin by presenting in the next section a simple theoretical picture of the pore's dynamics. After that, we describe the experiments.Model of the Opening and Closing of Pores. Closed membranes, such as red blood cells or vesicles, are under zero surface tension in their equilibrium unswollen phase (10). They undulate by thermal fluctuations. The resistance to deformations of the membrane is mainly caused by curvature energy. The surface is crumpled, as shown in Fig. 5a. An excess area ⌬A ϭ A Ϫ A p Ͼ 0 (where A is the total area and A p the projected area) is required to maintain the zero surface-tension state (11-13). When a vesicle adheres to a substrate or is sucked in a glass micropipette, the surface cannot adjust to its optimal value, and a surface tension of is created. The fluctuations of the membrane are strongly reduced, and the shape of the vesicle becomes spherical (Fig. 5b). The micropipette technique allows one to measure the relation between the tension (over four decades, 10 Ϫ3 -10 mN͞m) and the increase of the projected area A p (14). Two r...

We present a detailed analysis of the generation of second-harmonic radiation from biological membranes labeled with a styryl dye. In particular, we consider the high-numerical-aperture limit appropriate to highresolution microscopy in which an excitation beam is tightly focused from the side onto a membrane surface. In this limit the active surface area that contributes to second-harmonic generation (SHG) depends only on the tightness of the beam focus and the SHG radiation is confined by phase matching into two well-defined off-axis lobes. We derive expressions for the SHG radiation power, angular distribution, and polarization dependence in the cases of ideal or nonideal molecular alignment in the membrane and uniaxiality of the molecular hyperpolarizability. We define an SHG cross section similar to that used in two-photon-excited fluorescence (TPEF) to permit direct comparison of the two imaging modalities. Finally, we corroborate our results with experiments based on the excitation of a styryl dye in giant unilamellar vesicles with a mode-locked Ti:sapphire laser.

By focusing a pulsed laser beam into a sample, harmonic up-conversion can be generated as well as multi-photon excited fluorescence. Whereas multi-photon excited fluorescence microscopy is well established, the use of multi-harmonic generation for three-dimensional image contrast is very recent. Both techniques can provide similar resolution and, for adequate radiating source density, comparable signal levels, allowing them to be combined in a single versatile instrument. However, harmonic generation differs fundamentally from fluorescence generation in that it is coherent and produces radiation patterns that are highly sensitive to phase. As such, multi-harmonic generation microscopy provides a unique window into molecular spatial organization that is inaccessible to fluorescence.