Article,
Aberration-accounting calibration for 3D single-molecule localization microscopy
Optics Letters, (2018)
DOI: 10.1364/OL.43.000174
Abstract
We propose a straightforward sample-based technique to calibrate the axial detection in 3D single-molecule localiza-tion microscopy. Using microspheres coated with fluorescent molecules, the calibration curves of point spread function-shaping or intensity-based measurements can be obtained over the imaging depth range. This experimental method takes into account the effect of the spherical aberration without requiring computational correction. We demonstrate its efficiency for astigmatic imaging in a 1.2 $\mu$m range above the coverslip. Single-molecule localization microscopy (SMLM) is now a well-established method used in biology for a wide range of applications, especially single particle tracking 1 in living samples and super-resolution structural observation using (f) PALM, (d)STORM, or PAINT 2-6. Although retrieving the lateral positions in the focus plane is quite straightforward, 3D detection requires complementary axial information that can be provided either by point spread function (PSF) shape measurement methods 7-11, in which the axial information is encoded in the shape of the spots, or by intensity-based methods (such as interferometric measurements 12,13 or supercritical angle fluorescence (SAF) detection 14,15), which rely on the dependence of the intensity on the depth. In most cases, such an axial detection scheme requires a calibration to know the relationship between the measured value and the depth, or at least an experimental verification to validate the consistency of the results obtained from a theoretical characteristic curve. Most of the time this is performed by using fluorescent beads or molecules deposited on a coverslip and scanning the objective with a piezoelectric stage to introduce defocus in the system. While inexpensive and simple to perform , this method exhibits several drawbacks arising from the refractive index mismatch between the sample and the glass coverslip. First, the distance over which the focus plane is moved is not equal to the displacement of the objective. In practice, this so-called focal shift effect produces a stretching of the apparent distances. Although theoretical formulas 16 and experimental protocols 17,18 of various complexities are available to determine the value of the correction factor for different depth ranges, these methods are not sufficient to provide readily usable calibration data suitable for SMLM experiments. Indeed, they do not account for the effect of the spherical aberration on the PSFs. Such an aberration alters the shape of the spots and, thus, induces a bias in the axial positions detected through PSF-shaping methods. Calibrations performed by scanning the objective do not allow one to record the PSFs corresponding to a realistic experimental situation where the focus plane is fixed. Several techniques have been proposed to circumvent this issue, notably numerical computation 19. While this technique does not require a cumbersome experimental procedure, it does not fully correct the effect of the spherical aberration. Deng and Shaevitz proposed a reliable experimental method using optical tweezers to axially move fluorescent beads relative to the object plane 20 at the cost of a major modification of the setup. Similarly, adaptive optics can be used to correct the spherical aberration 21, but this requires expensive devices and induces a loss of photons. We designed a fully experimental, sample-based calibration method to provide unbiased calibration results that can be used for 3D SMLM measurements. In this protocol, all the axial information needed is provided by the known geometry of the sample, and the acquisitions are performed in the nominal conditions, i.e., for a given objective at a given position and for a given sample refractive index and fluorophore emission wavelength. More specifically, we used 15.0 $\mu$m (1.5 $\mu$m) diameter latex microspheres coated with biotin (Kisker Biotech, PC-BX-15.0), on which we attached the fluorophores of interest,
