Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Strategic and practical guidelines for successful structured illumination microscopy

Abstract

Linear 2D- or 3D-structured illumination microscopy (SIM or3D-SIM, respectively) enables multicolor volumetric imaging of fixed and live specimens with subdiffraction resolution in all spatial dimensions. However, the reliance of SIM on algorithmic post-processing renders it particularly sensitive to artifacts that may reduce resolution, compromise data and its interpretations, and drain resources in terms of money and time spent. Here we present a protocol that allows users to generate high-quality SIM data while accounting and correcting for common artifacts. The protocol details preparation of calibration bead slides designed for SIM-based experiments, the acquisition of calibration data, the documentation of typically encountered SIM artifacts and corrective measures that should be taken to reduce them. It also includes a conceptual overview and checklist for experimental design and calibration decisions, and is applicable to any commercially available or custom platform. This protocol, plus accompanying guidelines, allows researchers from students to imaging professionals to create an optimal SIM imaging environment regardless of specimen type or structure of interest. The calibration sample preparation and system calibration protocol can be executed within 1–2 d.

Your institute does not have access to this article

Relevant articles

Open Access articles citing this article.

Access options

Buy article

Get time limited or full article access on ReadCube.

$32.00

All prices are NET prices.

Figure 1: Biological showcase of commonly presented SIM reconstruction artifacts.
Figure 2: Principle of extending resolution with structured illumination (SI).
Figure 3: Identifying differences in modulation contrast.
Figure 4: Overview of the workflow for SIM experiments.
Figure 5: Preparation of calibration bead slide set.
Figure 6: Expected outcomes from calibration slide preparation.
Figure 7: Matching OTFs and experimental conditions in the sample for multicolor acquisitions.

References

  1. Heintzmann, R. & Cremer, C.G. Laterally modulated excitation microscopy: improvement of resolution by using a diffraction grating. Proc. SPIE 3568, 185–196 (1999).

    Article  Google Scholar 

  2. Gustafsson, M.G.L. Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy. J. Microsc. 198, 82–87 (2000).

    CAS  Article  Google Scholar 

  3. Gustafsson, M.G.L. et al. Three-dimensional resolution doubling in wide-field fluorescence microscopy by structured illumination. Biophys. J. 94, 4957–4970 (2008).

    CAS  Article  Google Scholar 

  4. Schermelleh, L. et al. Subdiffraction multicolor imaging of the nuclear periphery with 3D structured illumination microscopy. Science 320, 1332–1336 (2008).

    CAS  Article  Google Scholar 

  5. Sahl, S.J. et al. Comment on 'Extended-resolution structured illumination imaging of endocytic and cytoskeletal dynamics'. Science 352, 527–527 (2016).

    CAS  Article  Google Scholar 

  6. Li, D. & Betzig, E. Response to comment on 'Extended-resolution structured illumination imaging of endocytic and cytoskeletal dynamics'. Science 352, 527–527 (2016).

    CAS  PubMed  Google Scholar 

  7. Shao, L. & Rego, E.H. in Fluorescence Microscopy 213–225 (Elsevier, 2014).

  8. Allen, J.R., Ross, S.T. & Davidson, M.W. Structured illumination microscopy for superresolution. Chem. Phys. Chem. 15, 566–576 (2014).

    CAS  Article  Google Scholar 

  9. Rego, E.H. & Shao, L. Practical structured illumination microscopy. Methods Mol. Biol. 1251, 175–192 (2015).

    CAS  Article  Google Scholar 

  10. Fiolka, R. in Quantitative Imaging in Cell Biology 123, 295–313 (Elsevier, 2014).

    Article  Google Scholar 

  11. Komis, G. et al. Superresolution live imaging of plant cells using structured illumination microscopy. Nat. Protoc. 10, 1248–1263 (2015).

    CAS  Article  Google Scholar 

  12. Engel, U. in Quantitative Imaging in Cell Biology 123, 315–333 (Elsevier, 2014).

    Article  Google Scholar 

  13. Ball, G. et al. SIMcheck: a toolbox for successful super-resolution structured illumination microscopy. Sci. Rep. 5, 15915 (2015).

    CAS  Article  Google Scholar 

  14. Terui, Y. Image processing for structured illumination microscopy. in 1–3 (IEEE, 2015).

  15. Kraus, F. et al. Quantitative 3D structured illumination microscopy of nuclear structures. Nat. Protoc. http://dx.doi.org/10.1038/nprot.2017.20.

  16. Young, L.J., Ströhl, F. & Kaminski, C.F. A guide to structured illumination TIRF microscopy at high speed with multiple colors. J. Vis. Exp. http://dx.doi.org/10.3791/53988 (2016).

  17. Turnbull, L. et al. Super-resolution imaging of the cytokinetic Z ring in live bacteria using fast 3D-structured illumination microscopy (f3D-SIM). J. Vis. Exp. e51469–e51469 (2014).

  18. Křížek, P., Lukeš, T., Ovesný, M., Fliegel, K. & Hagen, G.M. SIMToolbox: a MATLAB toolbox for structured illumination fluorescence microscopy. Bioinformatics 32, 318–320 (2015).

    PubMed  Google Scholar 

  19. Müller, M., Mönkemöller, V., Hennig, S., Hübner, W. & Huser, T. Open-source image reconstruction of super-resolution structured illumination microscopy data in ImageJ. Nat. Commun. 7, 10980 (2016).

    Article  Google Scholar 

  20. Gustafsson, M.G.L. Nonlinear structured-illumination microscopy: wide-field fluorescence imaging with theoretically unlimited resolution. Proc. Natl. Acad. Sci. USA 102, 13081–13086 (2005).

    CAS  Article  Google Scholar 

  21. Rego, E.H. et al. Nonlinear structured-illumination microscopy with a photoswitchable protein reveals cellular structures at 50-nm resolution. Proc. Natl. Acad. Sci. USA 109, E135 (2012).

    CAS  Article  Google Scholar 

  22. Li, D. et al. Extended-resolution structured illumination imaging of endocytic and cytoskeletal dynamics. Science 349, aab3500 (2015).

    Article  Google Scholar 

  23. Eggeling, C. & Hell, S.W. in Far-Field Optical Nanoscopy 3–26 (Springer, 2014).

  24. Müller, T., Schumann, C. & Kraegeloh, A. STED microscopy and its applications: new insights into cellular processes on the nanoscale. Chem. Phys. Chem. 13, 1986–2000 (2012).

    Article  Google Scholar 

  25. Shao, L., Kner, P., Rego, E.H. & Gustafsson, M.G.L. Super-resolution 3D microscopy of live whole cells using structured illumination. Nat. Methods 8, 1044–1046 (2011).

    CAS  Article  Google Scholar 

  26. Betzig, E. et al. Imaging intracellular fluorescent proteins at nanometer resolution. Science 313, 1642–1645 (2006).

    CAS  Article  Google Scholar 

  27. Liu, Z., Lavis, L.D. & Betzig, E. Imaging live-cell dynamics and structure at the single-molecule level. Mol. Cell 58, 644–659 (2015).

    CAS  Article  Google Scholar 

  28. Wegel, E. et al. Imaging cellular structures in super-resolution with SIM, STED and localisation microscopy: a practical comparison. Sci. Rep. 6, 27290 (2016).

    CAS  Article  Google Scholar 

  29. Fiolka, R., Shao, L., Rego, H.E., Davidson, M.W. & Gustafsson, M.G.L. Time-lapse two-color 3D imaging of live cells with doubled resolution using structured illumination. Proc. Natl. Acad. Sci. USA 109, 5311 (2012).

    CAS  Article  Google Scholar 

  30. Rothbauer, U. et al. Targeting and tracing antigens in live cells with fluorescent nanobodies. Nat. Methods 3, 887–889 (2006).

    CAS  Article  Google Scholar 

  31. Grimm, J.B. et al. A general method to improve fluorophores for live-cell and single-molecule microscopy. Nat. Methods 12, 244–250 (2015).

    CAS  Article  Google Scholar 

  32. Olivier, N., Keller, D., Rajan, V.S., Gönczy, P. & Manley, S. Simple buffers for 3D STORM microscopy. Biomed. Opt. Express 4, 885–899 (2013).

    CAS  Article  Google Scholar 

  33. MacDonald, L., Baldini, G. & Storrie, B. Does super-resolution fluorescence microscopy obsolete previous microscopic approaches to protein co-localization? Methods Mol. Biol. 1270, 255–275 (2015).

    CAS  Article  Google Scholar 

  34. Cerase, A. et al. Spatial separation of Xist RNA and polycomb proteins revealed by superresolution microscopy. Proc. Natl. Acad. Sci. USA 111, 2235–2240 (2014).

    CAS  Article  Google Scholar 

  35. Schmied, J.J. et al. DNA origami-based standards for quantitative fluorescence microscopy. Nat. Protoc. 9, 1367–1391 (2014).

    CAS  Article  Google Scholar 

  36. Marno, K. et al. The evolution of structured illumination microscopy in studies of HIV. Methods 88, 20–27 (2015).

    CAS  Article  Google Scholar 

  37. Schaefer, L.H., Schuster, D. & Schaffer, J. Structured illumination microscopy: artefact analysis and reduction utilizing a parameter optimization approach. J. Microsc. 216, 165–174 (2004).

    CAS  Article  Google Scholar 

  38. Shroff, S., Fienup, J. & Williams, D. OTF compensation in structured illumination superresolution images. Proc. SPIE 7094 http://dx.doi.org/10.1117/12.791052 (2008).

  39. Débarre, D., Botcherby, E.J., Booth, M.J. & Wilson, T. Adaptive optics for structured illumination microscopy. Opt. Express 16, 9290–9305 (2008).

    Article  Google Scholar 

  40. Righolt, C.H. et al. Image filtering in structured illumination microscopy using the Lukosz bound. Opt. Express 21, 24431 (2013).

    Article  Google Scholar 

  41. Wicker, K., Mandula, O., Best, G., Fiolka, R. & Heintzmann, R. Phase optimisation for structured illumination microscopy. Opt. Express 21, 2032–2049 (2013).

    Article  Google Scholar 

  42. Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).

    CAS  Article  Google Scholar 

  43. Schneider, C.A., Rasband, W.S. & Eliceiri, K.W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675 (2012).

    CAS  Article  Google Scholar 

  44. O'Holleran, K. & Shaw, M. Optimized approaches for optical sectioning and resolution enhancement in 2D structured illumination microscopy. Biomed. Optics Exp. 5, 2580–2590 (2014).

    Article  Google Scholar 

  45. Demmerle, J., Wegel, E., Schermelleh, L. & Dobbie, I.M. Assessing resolution in super-resolution imaging. Methods 88, 3–10 (2015).

    CAS  Article  Google Scholar 

  46. Allan, C. et al. OMERO: flexible, model-driven data management for experimental biology. Nat. Methods 9, 245–253 (2012).

    CAS  Article  Google Scholar 

  47. Nieuwenhuizen, R.P.J. et al. Measuring image resolution in optical nanoscopy. Nat. Methods 10, 557–562 (2013).

    CAS  Article  Google Scholar 

Download references

Acknowledgements

We acknowledge J. Neumann and M. Grange for technical support, and L. Ferrand of GE Healthcare for helpful discussion. The 2014 OMX User Meeting made notable contributions to the publication recommendations. We thank all colleagues who contributed to many discussions to shape this protocol. We are further indebted to H. Leonhard and I. Davis for their long-standing support. This work was funded by the Wellcome Trust Strategic Awards 091911 and 107457 supporting advanced microscopy at Micron Oxford. The OMX 3D-SIM system in the Rockefeller University Bio-Imaging Resource Center was funded by award no. S10RR031855 from the National Center for Research Resources. J.D. is supported by the NIH-Oxford-Cambridge Scholars Program. G.B. is supported by an MRC Next Generation Optical Microscopy Award (MR/K015869/1). A.M. is supported by JSPS KAKENHI grant nos. JP16H01440 (“resonance bio”), JP15K14500 and JP26292169.

Author information

Authors and Affiliations

Authors

Contributions

C.I., J.D., A.J.N., M.M., E.M., Y.M. and L.S. collected data and created figures. A.M., G.B., M.M. and I.M.D. provided technical expertise and advice. All authors contributed to artifact documentation. J.D., C.I., A.J.N. and L.S. wrote the manuscript. A.J.N., Y.M. and L.S. conceived the project.

Corresponding author

Correspondence to Lothar Schermelleh.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Integrated supplementary information

Supplementary Figure 1 Reconstructed noise artifact as a function of modulation contrast in the SIM raw data

Datasets progress from having a very low signal-to-noise ratio (SNR) (far left column) up to an acceptable SNR (far right column) with initial acquisition conditions directly affecting quality of the resulting reconstruction. Data collected from wild type C127 mouse cells immunostained for tubulin (Alexa Fluor 488) are presented as: representative frames of raw datasets, with stripes shown in one angular direction (top row); modulation contrast to noise ratio (MCNR) in the raw data mapped onto the reconstructed dataset (MCN readout of SIMcheck; maximum projections; second row); reconstructed single-slice images showing the full dynamic range (third row); reconstructed single-slice images showing the dynamic range after cut-off below the mode value of the intensity histogram (fourth row). Reconstruction intensity histograms (fifth row) of reconstructed 32-bit data are show as log-scaled (grey) or linear-scaled (black) intensity values, with the x-axis representing pixel intensity in the given range. The values removed in the mode-based cut-off overlaid in red, and the ratio between minimum and maximum values in the histogram (MMR) is expressed above each plot. Lateral 3D Fourier plots (FTL, sixth row) of the reconstructed data are shown with reciprocal distance in μm plotted as circles over the FFT. The radial profiles (FTR, last row) display the corresponding radially averaged amplitudes with the red line indicating a reference amplitude value (22, arbitrary units) for comparison. Note that with increasing signal-to-noise the profiles become smoother and the area between the reference and the curve become larger. For each image, the dashed box indicates the inset region. Scale bars: 5 μm and 1 μm (inset). For each row, T indicates % transmission of the 488 nm laser, and exp. indicates the exposure time. Images were acquired on a GE OMX V3 Blaze instrument equipped with an Olympus 60x/1.42 PlanApo N objective and PCO edge 4.2 sCMOS cameras.

Supplementary Figure 2 Illustration of ‘hatching’ artifacts caused by missing angle information

Data collected from wild type C127 mouse cells stained with either (a) DAPI to label chromatin distribution or (b, c) immunostained for RNA polymerase II phosphorylated at serine 2 of the C-terminal domain (RNAPII S2P). These showcase the effect of missing angular frequency information on complex (chromatin) or punctate (RNAPII) labeled features, supplementary to Fig. 1c. (a-a’’) Reconstruction of the same dataset with correct angle (k0) settings (a), and with a false k0 parameter settings for either one (a’) or two (a’’) angles, respectively. Note that the loss of lateral resolution enhancement is particularly evident in the corresponding Fourier plots with frequency extensions missing for the affected angles (second column). Accordingly, the labeled features in the reconstructed images (insets) are less sharp. (b, c) Reconstruction of punctate nuclear RNAPII signals with correctly fitted k0 value (b) compared to a dataset where the k0 fitting failed for two of the three angle directions (c). Detailed view of the reconstructed lateral midsection highlight elongated features of nuclear signals in c, but not in b (boxed insets). Single cytoplasmic background signals (originating from unspecific bound antibody complexes) show ‘starfish’ extensions in the faulty dataset (circular inset in c) while similar signals are round in the well-reconstructed dataset (arrows, inset in b). While not easily noticeable in the reconstructed image, the absence or presence of hatching becomes apparent in the corresponding Fourier plot. Of note, enhancement of the axial resolution, as well as contrast increase by out-of-focus signal suppression are hardly affected, as highlighted by the comparison of the the orthogonal sections and the corresponding orthogonal Fourier plots in b’ and c’. Scale bars: 5 μm and 1 μm (insets). Images were acquired on a GE OMX V3 Blaze instrument equipped with an Olympus 60x/1.42 PlanApo N objective and PCO edge 4.2 sCMOS cameras.

Supplementary Figure 3 Illustration of refractive index mismatch induced artifacts

(a) Mounting medium refractive index (RI) matched to the sample will exaggerate or suppress the z-ghosting artifact, as observed in lateral and orthogonal views of sub-diffraction sized FluoSphere green (Ex. 488λ) coverslip beads (Ref. calibration slide #1). As the acquisition medium’s refractive index (RI) is varied from too low (left) to optimally matched (middle) to too high (right) reconstruction of the depicted area is seen to increase in quality nearest the sample’s reference oil. Orthogonal views highlight that an extreme mismatch (>5 units change from reference RI) produces images with prominent intensity dips (arrowheads) in z-slices adjacent to sample real signal. Additionally, undershooting the sample’s RI-match can result in a noticeable ghost image in z-slices atop real signal while overshooting this match results in a ghost image of beads repeated in z-slices beneath real signal (dotted arrows). (b) Effect of increasing RI mismatch on a punctate nuclear staining pattern (EdU pulse replication labeling and detection with click chemistry in C127 cells). Note the intermixing of real and ghost signals in the orthogonal views with increasing mismatch. (c) RI mismatch on very bright isolated features (here 200 nm diameter TetraSpeck beads) produces characteristic hexagonal patterns of the out-of-focus echo signal, highlighted in the detailed view of individual z-planes through a selected bead (right). (d) Image series supplementary to Fig. 1e demonstrating ghosting on an MDCK cell immunostained for tubulin (Alexa Fluor 488) and desmoplakin C-terminus (Alexa Fluor 568) to emphasize the ease in diagnosing intensity dips from an RI-mismatch in images prior to any thresholding. Upper row of panels shows the lateral (top panel) and orthogonal (bottom panel) cross section of the reconstructed dataset with the full dynamic range, while the lower row shows the same data after intensity cut-off at zero. Arrowhead (first column) indicates position of the orthogonal view. Scale bar: 5 μm. Images in (a-c) were acquired on a GE OMX V3 Blaze instrument with PCO edge 4.2 sCMOS cameras; images in (d) were acquired on a GE OMX V4 Blaze instrument equipped with an Olympus 100x/1.40 SApo objective and Photometrics Evolve EMCCD cameras.

Supplementary Figure 4 Illustration of z-wrapping artifacts

(a) Schematic diagram indicating the top (A), central (B), bottom (C), and full (D) sub-sections used for the reconstruction. Coverslip is on the bottom. (b) Orthogonal views of the resulting reconstructions. Note the artifactual patterns generated in (A) and (C) sections. Tick marks in each indicate positions of images shown in the next panel. (c) Representative slices of each reconstruction, corresponding to the z positions marked in (b). Letters represent the corresponding sub-section of the entire nucleus as shown in (a), while numbers represent the corresponding z-position shown in (b). Arrowheads denote z-wrapping artifacts, or ghost signals from other z-slices in the reconstructed segment that occur on the opposite end of a z stack to where the stack has finished acquisition on a prominent structural feature. These include shadowing and moderate ‘honeycomb’ patterning (inset). Scale bars: 10 μm and 2 μm (inset). Images were acquired on a GE OMX V3 Blaze instrument equipped with an Olympus 60x/1.42 PlanApo N objective and PCO edge 4.2 sCMOS cameras.

Supplementary Figure 5 Strategic considerations for imaging method selection and trade-off finding

(a) Schematic diagram illustrating factors influencing the strategic decision of imaging modality selection and governed by the research question to be addressed. Also illustrated are the contributions of experimental factors affecting the relative balance between contrast and photon budget, and where these feed into technical considerations to optimize resolution, context, speed, and sample preservation. Note that improving one area will necessarily lead to trade-offs in others. (b) Schematic representation of potential factors increasing the signal and reducing unspecific background to increase the overall image contrast, which is balanced with photodamage and fluorophore bleaching not exceeding some critical threshold. Too much bleaching reduces SNR and thus the (modulation) contrast in the raw data, and will induce artifacts if the ratio is weighted too heavily towards generating high signal at the start of the dataset acquisition. The level of bleaching allowed depends on the acquisition order and may be case-dependent; however, from experience rates of up to 50% seem acceptable in most cases. The ‘Channel Intensity Plot’ function in SIMcheck13 offers a useful tool for quantifying acquisition bleaching and intensity variations affecting SNR and modulation contrast in SIM raw data.

Supplementary Figure 6 Illustration of ‘ghosting’ effect in PSFs dependent on immersion oil refractive index, wavelength, and imaging depth

3D-SIM images of sub-diffraction PSF beads in indicated wavelengths were collected at either coverslip level (bottom, 0 μm) or the top (8 μm) of slides prepared as described, and intensity measurements averaged from 10 separate beads are plotted on the right with intensity on the y-axis and relative z-position from the center on the x-axis. The leftmost set of bead images are displayed with full bit-depth after reconstruction for clarity of artifacts, and the right set are displayed after cutting off intensities at 0 to represent images most likely seen by the user. Arrowheads indicate ‘ghost’ signals that are likely to be interpreted as ‘false-positive’ in a real sample situation. Each set of beads was imaged in a different immersion oil RI indicated on the left and above each graph and reconstructed with color-specific ‘near-optimal’ OTFs, i.e. from PSFs acquired with RIs of 1.510 (blue), 1.512 (green), 1.514 (red), 1.516 (far-red). The relative size of the lobes in the intensity graph correspond to the prominence of the spherical aberration artifacts seen in the raw images, and are dependent both on wavelength and immersion RI. Low RI’s generate more artifacts in the longer red wavelength while the reverse is true for blue wavelengths. The optimal matching of imaging conditions and OTF for each wavelength is indicated by the colored outline around certain bead images. Bars: 1 μm. Images were acquired on a GE OMX V3 Blaze instrument equipped with an Olympus 60x/1.42 PlanApo N objective and PCO edge 4.2 sCMOS cameras.

Supplementary Figure 7 SIM acquisition of 3D bead samples highlighting variations in the ‘safe z-range’ and color channel matching under different imaging conditions

Comparisons of multicolour gel-embedded 0.2 μm TetraSpeck beads acquired using varying RI immersion oils (indicated on left) and reconstructed using different combinations of OTFs (coloured legends, also on left). The left image column shows xz-views (maximum projection) of superimposed channels (after image registration). The other columns show individual blue, green and red emission channels of the same xz-projections. In all images the coverslip is at the bottom and the y-axis shows ascending z-distance deeper into the gel. Insets show higher magnification views of individual beads with accompanying spherical aberration artifacts as a function of distance from the coverslip. Note the significant doubling/tripling of bead images in the xz-axis, especially when using an oil of too low RI (1.516) for this sample (top row). Superimposed images also reveal colour dispersion in the aberrated regions, particularly after reconstruction using channel-optimised OTFs (middle image row), rather than matched OTF sets (bottom row). Ghosting effects are most extensive in the red channel and least problematic in the blue channel. Dotted lines mark the depth optimum, colored areas around indicate the corresponding safe z-range for each channel, i.e. the range of depths within which ghosting effects are minimal. The extent and z-position of the safe z-range varies with acquisition RI, wavelength, and the OTFs used for reconstruction. Details on assembling gel-embedded bead samples are available in the Supplementary Method. Scale bars: 5 μm and 0.5 μm (inset). Images were acquired on a GE OMX V4 Blaze instrument equipped with an Olympus 100x/1.40 SApo objective and Photometrics Evolve EMCCD cameras.

Supplementary Figure 8 Matching of refractive indices of immersion oils used for sample acquisition and OTF measurements can compensate for PSF asymmetry

(a) Matrix of acquisitions of the same DAPI-stained C127 nucleus with varying immersion oil RIs (y-axis) reconstructed with OTFs generated from matching or varying immersion oil RIs (nOTF, x-axis). Images are shown with full bit depth and no discarding of negatives to emphasize the dynamic range of the reconstruction. The immersion oil RI generating the most symmetrical PSF on the system used is 1.511, therefore the ‘asymmetry index’ is calculated as deviations from that metric based on the immersion RI used. This clearly shows that matching the immersion oil RI during acquisition of both the OTF and the raw image results in higher contrast and a better reconstruction than using an OTF generated with the ideal RI to reconstruct and image collected with a different immersion oil RI. (b) Example of 100nm red ·ead layer showing a similar effect. The extent of mismatch compensation, however, is affected by wavelength, as PSFs of longer wavelengths are more easily distorted than shorter wavelengths (prominence of side lobes, see Supplementary Fig. 6). Thus, matching refractive indices n oil and n OTF results in a smaller window of acceptable compensation. Scale bars: 10 μm (a), 5 μm (b). Images in (a) were acquired on a GE OMX V3 instrument equipped with an Olympus 100x/1.4 PlanApo objective and Cascade II:512 EMCCD cameras, and images in (b) were acquired on a GE OMX V3 Blaze instrument equipped with an Olympus 60x/1.42 PlanApo N objective and PCO edge 4.2 sCMOS cameras.

Supplementary Figure 9 Schematic representation of matching spherical aberrations of OTF and sample for multi-channel SIM imaging

(a) Representation to scale of the light path through sample components. For a high NA oil objective, the opening angle is approx. 70°. Fixed and variable refractive indices n of the different components are indicated. (b) Representation of determining optimal OTF sets based on sample depth (based on empirical data acquired on an OMX V3 Blaze system). The y-axis represents depth, or the distance of the sample region from the coverslip, in μm. The x-axis indicates the relative change in refractive index (RI) of the immersion oil from a pre-determined, system-specific reference – generally the RI of oil that gives a symmetrical PSF in the green channel at the coverslip. The optimal OTF for the 488 nm excitation channel on this system was determined (by observed symmetry of the PSF) be at an immersion RI of 1.512 (defined as the zero point on the x-axis, as the reference value may deviate on other systems). These measurements are valid for samples mounted in glycerol with a RI of 1.47 at 23 °C, with solid lines indicating the depth at which a given immersion RI will produce a symmetrical PSF in glycerol, and would thus best match to an OTF measured at the coverslip. E.g. following the green match line for 488 nm excitation, for best matching in 4 μm depth the RI should be increased by +0.002 relative to the reference (circle, grey arrow); in this particular case an oil with RI of 1.514 should be used. Dashed lines represent the equivalent values for water-based immersion media (such as OptiMEM) at 23 °C, and dotted lines represent equivalent values for water-based immersion media at 37 °C. Note that the higher refractive index between immersion and mounting medium causes a stronger tilt along the y-axis, while a temperature increase effectively changes viscosity and thus the RI of the immersion oil, leading to a deviation from the nominal RI typically indicated for 23 °C, and shifts along the x-axis. (c, d) Representations of the match lines and the corresponding safe z-ranges of good reconstruction quality when using color optimized OTFs (c) as compared to using a set of channel specific OTFs acquired with the same RI (in this example 1.514) shifting the optimum towards the red channel (d). Note that in the latter case the match lines for all channels co-align and the corresponding safe zones display a much wider overlap region. (e) Representation of the ideal case of having no RI mismatch achieved when combining silicone oil immersion objective and a RI matched mounting medium (e.g. 50% glycerol), and both matching the RI of the biological specimen. Theoretically this would allow aberration-free imaging throughout the working distance of the objective. In praxis, however, light scattering and specimen-inherent RI variations will become limiting, especially when the light must penetrate through many layers of biological tissue, for which correction requires more advanced adaptive optics.

Supplementary Figure 10 Evaluation of Z-stack images and corresponding 3D Fourier plots reveal inappropriate SIM beam alignment

Assessment of central (a) and peripheral (b) z-slices of reconstructed bead layer image stacks, that were acquired under otherwise ideal conditions (i.e., bright and photostable sample, minimal RI mismatch, well-matched OTF, etc.). (a) The central section of the corresponding 3D Fourier plots (function provided with SIMcheck 1.1) show an even frequency distribution with no obvious abnormalities or loss in resolution (top row) whether the central illumination beams (0th order) of all three angles are aligned (left) or misaligned (right). However, assessing the reconstruction of out-of-focus z-planes (b) displays an angle-specific hatched noise pattern in the acquisition of the misaligned system (right, inset). Hatching in z-direction can also be observed in the magnified orthogonal view (c, right panel), and by accompanying angle specific amplitude peaks in the peripheral region of the 3D Fourier plot (arrows). Scale bars: 5 μm and 0.5 μm (insets). (d) Spatial frequency components can also be viewed and evaluated via maximum projection of the 3D Fourier plots to reveal issues with SIM beam alignment. Here both lateral and orthogonal view show additional interferences as ‘dots’ in frequency space (arrowheads, right panel), highlighting that despite overall resolution being comparable to an aligned instrument (left), the misaligned instrument (right) is generating suboptimal SIM images. Images were acquired on a GE OMX V3 Blaze instrument equipped with an Olympus 60x/1.42 PlanApo N objective and PCO edge 4.2 sCMOS cameras.

Supplementary information

Supplementary Figures and Table

Supplementary Manual, Supplementary Table 1, Supplementary Method, and Supplementary Figures 1–10. (PDF 3220 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Demmerle, J., Innocent, C., North, A. et al. Strategic and practical guidelines for successful structured illumination microscopy. Nat Protoc 12, 988–1010 (2017). https://doi.org/10.1038/nprot.2017.019

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1038/nprot.2017.019

Further reading

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing