Table 1.
Characteristics of Kohinoor, Padron, Dronpa, rsEGFP, Skylan-S and Skylan-NS.
The values were taken from literature, ref. [15–18]. Brightness is represented as a percentage with respect to that of Dronpa. N/A: no data available. *: Brightness of Skylan-S and Skylan-NS is a relative value to Dronpa calculated from values given in ref. [15] and [18], respectively.
Fig 1.
NL-SIM acquisition methods comparison between fluorophores with positive contrast photoswitching characterics and fluorophores with negative contrast photoswitching characterics.
Three different trigger timings of synchronization for NL-SIM experiments are present. Note that all triggers are on the low-to-high edge. The region marked in yellow indicates the timing of the camera readout. S.I.: structured illumination. WF I.: wide-field illumination. A) NL-SIM acquisition with Kohinoor. Kohinoor is deactivated with wide-field illumination at 405 nm and is activated and excited with structured illumination at 488 nm. B) NL-SIM acquisition with negatively switching RSFPs. A pre-activation with structured illumination at 405 nm is required. Then a structured illumination at 488 nm with a phase shift of π is used to acquire NL-SIM raw data. This method requires very precisely adjusted patterns switching between 405 nm and 488 nm illuminations over a large field of view. C) Further scheme for negative switching RSFPs. Pre-activation with uniform illumination at 405 nm followed by structured illumination at 488 nm to deactivate most of the fluorophores. A π phase shifted structured illumination at 488 nm with π is used to read out the signal from the remaining fluorophores. This scheme avoids precise achromatic matching between two patterns.
Fig 2.
Illustration of switching behavior of Kohinoor and simulation of fluorescent emission distribution in a two-beam NL-SIM scheme.
a) Our state diagram of Kohinoor used in the simulations. Excitation and photo-activation at 488 nm, deactivation at 405 nm. Soff, S0, S1 and kR denote the non-fluorescent dark state, the fluorescent ground state, the fluorescent excited state and the fluorescence emission rate, respectively. b) Simulated dependence of the relative lateral fluorescent emission in a steady state situation without photo-deactivation (blue solid line) and with deactivation at under strong 405 nm illumination (red solid line) at a relative transition rate of 0.1KR in the maximum. The blue dashed lines of (1) to (4) correspond to relative deactivation rates of 0.02kR, 0.04kR, 0.06kR and 0.08kR respectively. c) Absolute component strength in dependence of relative deactivation rate. d) Component strength normalized to the total intensity. The blue dashed lines (1) to (4) correspond to the dashed lines in panel b.
Fig 3.
a) Fourier transform of sample emission generated under uniform illumination. Sample information in Fourier space lying inside the region of support of the OTF, defining a cut-off frequency of Kx,max, is detectable. b) Fourier transform of linear sample emission under sinusoidal illumination (LSIM). In addition to the zero frequency as in a, two ±1st object components at ±Kb are generated. Hence high-frequency sample information is downshifted into the support region of the detection OTF. Accordingly, the effective OTF (lower part) is enlarged. c) Fourier transform of sample emission under nonlinear structured illumination conditions. The nonlinear sample response introduces higher illumination harmonics in Fourier space to all of which the sample information is attached.
Fig 4.
Fast nonlinear structured illumination microscope setup.
A 488 nm laser was coupled into a single-mode polarization-maintaining fiber for activation and excitation of the fluorophores. A polarizer ensured linear polarization of the light before the SLM. The fast ferroelectric liquid crystal SLM was limited to displaying binary patterns. An excitation filter was used for cleaning up fluorescence generated by the SLM. A quarter wave plate (QWP) together with the azimuthal polarizer ensures azimuthal polarization of the light for all grating orientations to guarantee high contrast of the grating pattern in the sample. The zero diffraction order of the 488 nm laser was blocked by the backside of a very small mirror (M1) used to reflect the 405 nm laser into the system near the Fourier plane of the SLM for uniform illumination. The ±1st diffraction orders of the 488 nm laser passed through a passive filter at the Fourier plane of the SLM. Both 405 nm and 488 nm lasers were sent to the sample plane through a 63× oil immersion microscope objective (NA 1.46, alpha Plan-Apochromat, Korr TIRF, Zeiss, Germany). Fluorescence was imaged and collected through an emission filter by the sCMOS camera (Orca Flash 4, Hamamatsu, Japan).
Fig 5.
Repeated photoswitching test on Kohinoor-actin expressed in a fixed HeLa cell.
a) A fixed HeLa cell expressing Kohinoor-actin was switched on and excited by a 488 nm laser. A 405 nm laser was used to switch the fluorophores off. Scale bar: 5 μm. Photoswitching of Kohinoor over time with the light intensity of 8.4 W/cm2 (488 nm) + 75.6 W/cm2 (405 nm), b) 14.0 W/cm2 (488 nm) + 75.6 W/cm2 (405 nm), c) 28.0 W/cm2 (488 nm) + 75.6 W/cm2 (405 nm), d) respectively. e) The maximum fluorescence intensity of the 20th raw image in b), c) and d). f) The mean fluorescence intensity in the last raw image of each excitation section with 488 nm irradiation over six switching cycles. ST1: 14 W/cm2 (488 nm) + 27 W/cm2 (405 nm) and ST2: 20 W/cm2 (488 nm) + 32 W/cm2 (405 nm), respectively. g) Photoswitching of Kohinoor at 20.0 W/cm2 (488 nm) + 11.0 W/cm2 (405 nm). The mean intensity denotes fluorescence emission intensity. The units are consistent in all plots.
Fig 6.
a) The value of the brightest pixel of each image in all 30 acquired raw images was calculated. 30 dark images were acquired for calculating the background value. b) Brightness histogram of six raw images in six grating directions.
Fig 7.
NL-SIM imaging of a fixed HeLa cell expressing Kohinoor-actin.
a) NL-SIM image in a large field of view. Magnified view of the yellow-boxed region corresponds to a conventional wide-field microscopy image b), LSIM image c), and NL-SIM image with one extra component d). e) Normalized intensity profiles between the white triangles in b)-d). Magnified view of the blue-boxed region represent conventional wide-field microscopy image f), LSIM image g), and NL-SIM image h). Scale bar in magnified images is 500 nm.