Nikon Instruments Europe

Transcription

1 Nikon Instruments Europe Recommendations for N-SIM sample preparation and image reconstruction Dear customer, We hope you find the following guidelines useful in order to get the best performance out of the NSIM system. General considerations for sample preparation and acquisition: 1- The biological specimen should contain well defined structures, to certain extent: reconstruction calculations require sufficient information to identify and re-position high spatial frequency information, so the higher the structural density in the FOV, the better. 2- The sample should be as bright as possible: SIM is friendly to standard fluorophores, but please consider fluorophore quantum yield. Higher quantum yield translates in brighter samples given the same amount of sample labeling. As an example, use Alexa488 instead of FITC. Generally, Alexa dyes or Cyanines are a good choice. Dim samples may require: a. Long acquisition times, which increases the risk of vibration-induced artifacts. b. Higher EM gain levels, which could introduce more noise (noise = artifacts). Remember, SIM requires 9 (for 2D) or 15 (for 3D) images for each time-point/z-slice. This can lead to excessive photo-bleaching. 3- The specimen should be flat, and coverslips should be made of high quality 1.5 optical glass: a. We recommend the user of spacers in order to improve sample flatness. Please, keep in mind that lack of flatness will affect the illumination modulation contrast across the FOV. These spacers are manufactured, for example, by Sigma Aldrich (Grace Biolabs, SecureSeal imaging spacer, 0.12 mm depth, item GBL654006), or by CoreWell. These spacers are self-adhesive. 1

2 b. Coverglass thickness is critical for SIM. Only with the right glass thickness may the illumination pattern be correctly projected onto the sample, directly impacting illumination contrast quality. High Performance No. 1.5H glass coverslips (thickness tolerance = 170 µm ± 5 µm, compared to ± 15 µm for standard coverslips) are thus required. Recommended manufacturers are Schott and Marienfeld. We advise to test every batch. 4- The refractive index of the mounting medium should be the same as the immersion medium. a. We strongly recommend NF type 2 oil (1.515 refractive index, 23 C) from Nikon to obtain the best performance. Type A oil may also be used. b. As mounting medium for samples to be imaged with 100x oil objective, we suggest Prolong Gold, TDE or glycerol (no Vectashield). Pro-long Gold takes hr to cure. In case of using 60x water immersion objective (or 100x for live cell imaging), any water based medium (PBS, for example) is suitable. Agarose embedding may be used for imaging with both objectives, and is especially indicated for bead sample preparation. See annex. 5- In case of live cell imaging, make sure that your acquisition rate is faster than the dynamics to be observed, i.e. your time sampling is adequate to investigate the event of interest. The slowest acquisition is generally best in terms of quality (generally meaning higher photon yield and, thus, better signal to noise). a. 2D requires 9 images per time point (minimum of 0.6 sec to acquire the 9 frames with 3 msec exposure time). b. 3D requires 15 images per time point (minimum of 1 sec to acquire 15 frames). 2

3 6- For dead samples, make sure the sample is properly fixed (example, 4% PFA) and mounted. 7- Keep the coverglass and the objective lens always clean. This is absolutely critical for optimal SIM operation. Nikon strongly recommends petroleum ether as best organic solvent. NF oil maybe sticky, please clean the objective lens thoroughly. Use 100% pure spun cellulose fiber optical paper to wipe the objective lens. 8- The objective correction collar should be adjusted under the same conditions that the sample will be imaged in (same mounting medium, coverslip thickness, temperature ). Nikon s correction collar enables fine tuning of the optical components in the objective to compensate slight differences in coverglass thickness and/or refraction index. If the collar is not correctly adjusted, the Point Spread Function (PSF) of any diffraction limited object will be distorted, and reconstruction of the resulting SIM images will be faulty. PSF should always be as symmetrical as possible. a. Ideally, collar adjustments should be made using diffraction limited spots found in the sample and in the focal plane that you will be imaging. b. If you need to use diffraction limited beads for collar adjustment, prepare bead slides with mounting medium of same refractive index as that of the sample mounting medium. Try and find beads that are in a similar focal plane as your sample. c. Evaluate your PSF. PSFs should be symmetrical. Full width of half maxima measurements (FWHM) for wide-field PSFs, using 488 laser light, should yield 260 nm lateral and 600 nm axial resolution. 3

4 9- Optimize your signal to noise ratio: short exposure time, and a maximum intensity value that does not exceed ~12,000 counts for 16-bit mode and ~4,000 for 14-bit mode for the camera. Recommended acquisition camera settings: a. Format: No Binning b. Conversion Gain: 1.0x c. Readout Mode: i. EM Gain 1MHz, 16bit (recommended if acquisition speed is not an issue; better for samples with large intensity ranges) ii. EM Gain 10MHz, 14bit (recommended for speed). d. EM Gain Multiplier: Keep as low as possible to reduce noise (normal range 200 to 300). e. Maximum pixel value: The reconstruction process involves additive functions. To avoid oversaturation and artifacts in the final reconstructed image, use the histogram window to help adjust settings to achieve the following intensity values: i. For EM Gain 1MHz, 16bit = Maximum intensity is ~12,000 ii. For EM Gain 10MHz, 14bit = Maximum intensity is ~4, Keep system in a stable environment. Make sure the room hosting the system is isolated from environmental vibrations (check anti-vibration table) and, critically, from temperature fluctuations. Optimal system operation is 23 C. General considerations for SIM image reconstruction: The ability of the microscope to transfer contrast in the specimen to the intermediate image plane, as a function of frequency, is called the Modulation Transfer Function (MTF, or Optical Transfer Function) of the microscope. The fact that not all sample frequencies can pass through the back focal plane of the objective results in the limited resolution of the acquired images. This is reflected by the numerical aperture (NA) of the objective. The largest the NA, the more frequencies can pass the back focal plane of the objective, and the largest the obtained resolution can be. In SIM, by shifting and rotating the illumination pattern, we can separate and reposition the corresponding MTFs, enabling the extension of Fourier space (and effectively doubling the resolution of conventional wide-field images). The process of separation and repositioning of the different information components resulting from SIM imaging is done by a mathematical algorithm that requires some input in order to tune image reconstruction. In N-SIM, this process is highly automated, and few parameters need to be adjusted in order to obtain a super-resolved image. The parameters that need to be adjusted are: For 2D-SIM, TIRF-SIM a. Structured Illumination Contrast (how good the illumination modulation is at the sample focal plane) b. Apodization Filter (how many frequencies you allow to participate in the reconstruction. The lower this filter is set, the more frequencies are incorporated, at the cost of increasing the artifacts produced by the reconstruction process). For 3D-SIM (1 layer or Slice 3D ) a. Structured Illumination Contrast (same as before). 4

5 b. Apodization Filter (same as before). c. Width of 3D Filter (sectioning power of the algorithm in the axial direction). How to tune the reconstruction settings: 1. To restrain high frequency noise (to avoid reconstruction artifacts): Structured Illumination Contrast: increase Apodization filter: increase 2. To emphasize high frequency information (gain in resolution at the cost of having more artifacts): Structured Illumination Contrast: decrease Apodization filter: decrease 3. To increase the sectioning effect in Z: Structured Illumination Contrast: increase Width of 3D filter: increase 4. To make the axial image more continuous : Width of 3D filter: decrease Typical settings for 3D-SIM reconstruction: Typical 3D SIM settings Very thin and high S/N 3D samples Thicker 3D samples Structured Illumination contrast Auto Apodization filter Width of 3D filter Typical TIRF SIM settings Structured Illumination contrast Apodization filter