As a first approximation, the amount of photobleaching and damage is proportional to the excited-state population of the molecules, which is reflected in the amount of two-photon excited fluorescence. quantitatively measures the difference between the observed two-photon fluorescence and trapping force between a conventional optical tweezer (Gaussian trap) and polarization-shaped vortex traps. In addition to experimental measurements, we have also constructed theoretical models so that we can compare the results predicted by theory with our experimental values. As a test sample, we used fluorescent beads with diameters of 100nm, 500nm, 1μm, and 3μm. For the 100nm and 500nm beads we used a linearly polarized
beam, and for 1μm and 3μm beads we used a right circularly polarized
beam. shows the difference in the observed two-photon fluorescence from beads in Gaussian () and vortex traps (); for these experiments, trapping powers ranged from 40mW to 130mW (40mW for 3μm and 1μm beads; 65mW for 500nm beads; 130mw for 100nm beads). shows simulations that illustrate the highest-intensity region and thus the trapping position for sub 400nm beads in a Gaussian (top), linearly polarized
(left), and right circularly polarized
(right). In a highly focused system, non-centrosymmetric polarizations generate non-centosymmetric intensity patterns, such as the observed intensity lobes in & (bottom left profile). For large particles that trap at the core, we no longer rely on the intensity lobes for trapping. In this case, the intensity lobes would only induce greater photodamage on the trapped particle, so we used circular polarization, which is centrosymmetric and thereby gives an even intensity distribution in the ring of the focused vortex beam.
Figure 3 (a-h) Two-photon fluorescence images of trapped beads in Gaussian (a-d), right circularly polarized (e, f), and linearly polarized (g, h) vortex traps. The sizes of the beads are 3μm (a, e), 1μm (b, f), 0.5μm (c, g), and 0.1μm (more ...)
shows the difference in the observed two-photon fluorescence intensity from beads trapped in Gaussian and vortex traps; we chose beads of four different sizes (with diameters of 100nm, 500nm, 1μm, and 3μm) to study the dependence of photodamage on particle size. The dark blue bars are experimental measurements and the light blue bars are simulations. We displayed the difference in two-photon fluorescence both in total intensity (i.e. intensity from all parts of the illuminated bead) and in maximum intensity (i.e. intensity recorded from the brightest pixel). While total intensity is a good estimate of photodamage for particles much smaller than the trap because they are completely illuminated by the trap, it does not reflect photodamage to particles larger than the trapping spot (e.g. 1μm & 3μm beads) for which maximum intensity is a better indication. We note that the use of beads as a model to quantify photodamage only approximates the behavior of subcellular organelles in traps. Unlike beads, for example, molecules in organelles are free to rotate and diffuse into and out of the laser beam. Our experiments and simulations match well, except for the 1μm beads. This discrepancy is caused by the fact that experimentally our
beam contains a small percentage of higher order LG modes, which slightly broadens the outer edge of the
beam. The discrepancy was reduced significantly (inset) when we performed a mixed-mode simulation, which mixed the pure
beam with 10% of
and 2% of
Using beads as a model system, we have also quantified the maximum trapping force between a Gaussian and the two polarization-shaped vortex traps (). We measured the trapping force by equating it to the Stokes drag force, in which we gradually decreased the trapping power to the level that was just sufficient for the drag force to “knock” the particle out of the trap as the particle was translated at constant velocity. For the linearly polarized vortex trap, the trapping force was measured by translating the bead at a 45° angle with respect to the side lobes, because the intensity gradient is sharper across the width than along the length of the intensity ring. In general, the trapping force is less in the vortex trap because the laser energy is spatially more dispersed. An exception to this is when the particle is larger than the beam waist and axial trapping efficiency begins to influence the total trapping force 27
. show the difference in two-photon fluorescence between the Gaussian and linear
(100nm and 500nm beads) and Gaussian and right circularly polarized
(1μm and 3μm beads) after normalizing to the difference in trapping force, because the ability to manipulate a trapped particle is dependent on the degree of photodamage inflicted at a given amount of force exerted on the particle. From these experiments, the photodamage normalized to the force is always less using polarization-shaped vortex trap than Gaussian trap (). Because both the dependence of the trapping force on the steepness of the intensity gradient and two-photon photodamage are non-linear, polarization-shaped vortex traps should offer substantial advantages in the manipulation of fragile cellular structures. Assuming extra laser power is available, using a polarization-shaped vortex trap is always beneficial over a Gaussian trap.
Next, we studied the photobleaching of isolated mitochondria held by an optical tweezer and by a linearly polarized LG trap. Here we used the same laser power (power at object plane is 75mW) for both traps. Because the mitochondria were easily photobleached and were much less robust than beads, we were unable to hold onto the mitochondria for the long periods of time (tens of seconds) needed for force measurements. shows the result of this experiment, in which we compared the time course of bleaching of mitochondria (labeled with Mitotracker) trapped in an optical tweezer and in a vortex trap. To check for fluorescence from the mitochondria, we used a low level of visible light excitation (15μW after the objective at 488nm). The mitochondria trapped in the optical tweezer were rapidly bleached within the first few seconds, but the ones held by the vortex trap displayed a relatively stable fluorescent signal over this duration. The insets to the right of the plot show the fluorescence micrographs of the two representative mitochondria at 15 seconds (arrows) imaged under 488nm illumination. The ones held by the vortex trap exhibited strong fluorescence, but the ones trapped by the optical tweezer were no longer visible owing to photobleaching.
Figure 4 (a) A time plot that compares photobleaching of trapped mitochondria (stained with Mitotracker Green dye) in a conventional optical tweezer and in a linearly polarized vortex trap (laser power is 75mW at the object plane). To check for fluorescence, the (more ...)