A Streamlined Protocol for Single-Molecule Localization Microscopy in Arabidopsis Nuclei

Summary

This protocol presents an efficient workflow for single-molecule localization microscopy imaging of isolated Arabidopsis nuclei, using optimized isolation, fixation, and labeling to achieve reliable nanoscale visualization of chromatin and RNA Polymerase II. This approach can be readily adapted to other chromatin-associated proteins or histone modifications for high-resolution imaging.

Abstract

While confocal fluorescence microscopy has provided valuable insights into chromatin organization in plant nuclei, its diffraction-limited resolution constrains the investigation of chromatin architecture, motivating the use of super-resolution techniques such as Single-Molecule Localization Microscopy (SMLM). Among these approaches, direct stochastic optical reconstruction microscopy (dSTORM) provides nanoscale resolution in individual cells, enabling precise visualization of chromatin domains, histone modifications, and nuclear organization. While such methods are increasingly applied in mammalian systems, their use in plant biology remains limited, largely due to technical challenges in sample preparation.

Here, we present a streamlined and reproducible workflow for SMLM imaging of nuclei isolated from Arabidopsis thaliana. This protocol starts with seedling fixation to preserve nuclear morphology, followed by gentle tissue chopping and centrifugation to enrich intact nuclei. Isolated nuclei are then fluorophore-labeled in liquid medium and immobilized on low-melting agarose pads, a strategy that enhances stability during prolonged single-molecule imaging sessions. These steps collectively minimize background fluorescence, improve labeling consistency, and increase reproducibility across biological replicates.

The resulting preparations provide enhanced clarity for visualizing chromatin modifications and nuclear architecture in plants. By lowering the technical barriers to implement SMLM imaging in Arabidopsis, this protocol provides a versatile means to investigate epigenetic regulation, chromatin organization, and nuclear topological variations at the nanoscale. This work establishes a methodological foundation for applying SMLM to plants, bridging the gap with mammalian cell biology and opening new opportunities to study how nuclear architecture contributes to genome regulation in response to developmental and environmental cues in plant systems.

Introduction

Microscopy has long been a cornerstone for studying chromatin organization and nuclear architecture, revealing how DNA and histone-associated proteins arrange within the 3D nuclear space and influence gene regulation^1,2. Conventional fluorescence microscopy has provided valuable insights into large-scale chromatin domains such as chromosome territories, chromocenters, and nuclear bodies^3,4, feasible in situ, in particular in root tissues for plants^3,5,6. However, these techniques are fundamentally limited by the diffraction barrier of light, restricting spatial resolution to approximately 200 nm laterally and 500 nm axially⁷. This limitation obscures nanoscale details of chromatin folding, histone modification patterns, and transcriptional compartmentalization that underlie epigenetic regulation. To overcome this barrier, super-resolution microscopy (SRM) approaches such as structured illumination microscopy (SIM)⁸, stimulated emission depletion (STED)^9,10, and single-molecule localization microscopy (SMLM)^11,12 were developed to achieve resolutions ranging from 100 nm down to a few nm¹³.

Among SRM techniques, SMLM—including PhotoActivated Localization Microscopy (PALM)¹³ and direct Stochastic Optical Reconstruction Microscopy (dSTORM)—offers exceptional spatial resolution combined with molecular specificity¹⁴. By detecting the positions of individual fluorophores over time, SMLM reconstructs high-precision nanoscale images of chromatin domains, histone marks, and nuclear structures in single cells^15,16. This capability has transformed chromatin research in mammalian systems, where SMLM has been applied to visualize nucleosome clustering, chromatin compaction, and histone modification patterns with unprecedented detail^15,16,17. In contrast, the application of SMLM to plant nuclei remains limited, largely due to technical challenges associated with plant tissue properties—such as high autofluorescence, cell wall rigidity, and the presence of light-scattering compounds like starch and chlorophyll¹⁸. Furthermore, the delicate nature of nuclei during isolation and labeling requires optimized workflows to maintain structural integrity and minimize background fluorescence during prolonged imaging sessions.

The protocol described here was originally developed by Elizabeth Kracik-Dyer and Célia Baroux for stimulated emission depletion (STED) microscopy¹⁹. It was subsequently optimized and adapted to meet the specific requirements of STORM imaging with the integration of knowledge coming from published microbiology applications^20,21,22. The overall goal is to provide a streamlined and reproducible workflow for SMLM imaging of isolated Arabidopsis thaliana nuclei. By integrating optimized fixation, gentle mechanical disruption, and fluorophore labeling in suspension, this method preserves nuclear morphology while efficiently removing cytoplasmic contaminants that often compromise imaging quality. Immobilization of labeled nuclei on low-melting agarose pads ensures mechanical stability during acquisition, allowing for long-term single-molecule imaging at nanoscale precision. Compared to previously described SRM protocols for plant nuclei—particularly those based on STED microscopy^19,23—this workflow uses photoswitchable fluorophores (e.g., AF647, JF549), and reduces background fluorescence thanks to washing steps that discard chlorophyll and tissue debris, and improves labeling homogeneity by adding the labeling molecules directly to the fixed and permeabilized nuclei. Consequently, it enables consistent visualization of chromatin modifications, nuclear compartments, and epigenetic patterns in individual nuclei. By lowering the technical barriers for implementing SMLM in plants, this method broadens the accessibility of nanoscale imaging and provides a foundation for studying how chromatin organization and nuclear architecture contribute to gene regulation in response to developmental and environmental cues.

Protocol

NOTE: The detailed information on the materials and instruments used in this protocol is available in the Table of Materials. Unless otherwise specified, buffers are cleared on apyrogenic 0.2 µm filters. Double-distilled ultra-pure H₂O (ddH₂O) is used throughout this protocol. Centrifugation steps are carried out in 1.5 mL microtubes using a fixed-angle rotor.

1. Seedlings’ fixation and release of nuclei

1. Place a 6-well culture plate on ice under the fume hood, then add to each well approximately 4 mL of ice-cold fixation solution prepared extemporaneously under a fume hood (4% (v/v) methanol stabilized formaldehyde, 0.1% Triton X-100 in 1× Phosphate-Buffered Saline [PBS]).
2. Using clean forceps, remove the seedlings (between 8 and 15 days after germination) that have been grown on solid (0.8% agar) half-strength Murashige and Skoog Medium 2.2 g/L MS salts, 1% sucrose (w/v), pH 5.7, as described for MS seedlings in reference²⁴, and immediately submerge them in the fixation solution. Use one well for approximately 20 seedlings (3 wells were necessary to obtain sufficient material per sample).
3. Put the plate in the vacuum chamber filled with ice and apply a vacuum at pressures cycling between 0.2 bar and 0.8 bar for 15 min to remove trapped air and allow deep penetration of the fixative into the plant tissues²⁵. Slowly release the vacuum, stir the sample, and then apply the vacuum for a further 15 min.
4. Under the fume hood, remove the fixation solution with a pipette, and wash each well 3 times for 10 min under gentle agitation, using 5 mL of ice-cold PBS.
5. Remove PBS and use forceps to transfer the fixed tissue to a 120 mm × 120 mm Petri dish. Add to the fixed seedlings 50 µL of Nuclei Isolation Buffer (NIB), prepared as described by Kracik-Dyer and Baroux (45 mM MgCl₂, 20 mM MOPS, 30 mM sodium citrate, 0.3% Triton X-100. Adjust to pH 7 using 1 N HCl)¹⁹. Chip with a blade.
   NOTE: Microtome blades turned out to be more efficient than conventional razor blades for this task.
6. Scrape chopped tissue to one edge of the Petri dish with the same blade and add 600 µL of NIB, making sure to wash off any tissue stuck to the blade as well.
7. Transfer tissue slurry into the 2 mL Dounce homogenizer tube and stroke 5–10 times with the smaller pestle, and then with the larger pestle until it is smoothly moving inside the tube (it requires again usually 5–10 strokes).
8. Filter the resulting homogenized sample through a 30 µm mesh filter into a 1.5 mL microtube.
9. Centrifuge the suspension for 10 min at 800 × g using the lowest ramp for both acceleration and deceleration. A small white pellet is visible among the green tissue debris.
10. Discard the supernatant and resuspend the pellet in 150 µL of NIB.
11. Centrifuge again for 10 min at 800 × g and remove the supernatant. If there is some visible green debris, wash for a second time. Resuspend in 50 µL to 150 µL of NIB, based on the size of the pellet.
12. Store the resulting isolated nuclei suspension at 4 °C for a maximum of 1 week before performing the staining. However, it is recommended to proceed immediately.

2. (Immuno)staining

NOTE: In the following steps, the protocol is applied to a volume of 50 µL of nuclei suspension, corresponding typically to 20 seedlings. Unless specified, the following steps are performed at room temperature.

1. Centrifuge the solution for 10 min at 800 × g and discard the supernatant. Be careful not to touch the pellet.
2. Resuspend the nuclei by gentle pipetting in 50 µL of blocking buffer (143 mM NaCl, 4% BSA in PBS, stored at 4 °C). Incubate for 30 min under gentle agitation.
3. Centrifuge the solution for 10 min at 800 × g and discard the supernatant without touching the nuclei pellet.
4. Resuspend the nuclei by gentle pipetting in 50 µL of 1× PBS. Centrifuge again for 10 min at 800 × g. Discard the supernatant as above.
5. Resuspend the nuclei in 50 µL of primary antibody solution prepared in immunostaining buffer composed of 143 mM NaCl, 1% BSA, 0.05% Tween 20 in PBS (antibody and working dilutions are detailed in the Table of Materials). Incubate the samples for 60 min under gentle agitation.
6. Proceed with 2 washes (repeat steps 2.3 and 2.4).
7. Resuspend the nuclei in 50 µL of goat anti-rabbit AF647 secondary antibody at 1/200 dilution²⁶. Wrap the tubes in aluminum foil to protect the fluorophore from light. Incubate the tubes for 60 min under gentle agitation.
8. Proceed with 2 washes (repeat steps 2.3 and 2.4).
9. Resuspend the pellet in 50 µL of NIB. Store the suspension in 1.5 mL microtubes wrapped in aluminum foil. The labeling is stable for several weeks at 4 °C.

3. Mounting

1. For STORM imaging, prepare a low-melting agarose solution (LMA 1.5% w/v) by dissolving the powder in 1.25× buffer B composed of 50 mM Tris pH 8, 10 mM NaCl, 10% Glucose. Prepare 1.25× stock solution, and store at 4 °C.
2. Heat at 90 °C for 1 min to dissolve the powder and obtain a clear solution. Cool down and maintain the melted LMA solution at 37 °C before proceeding to the next step to avoid heat denaturation of glucose oxidase (GLOX) buffer (prepare 10x stock solution stored at -20°C composed of 100 mM Tris pH 7.5, 25 mM KCl, 1 mM TCEP pH 7.0, 5 mg/mL Glucose oxidase, 0.4 mg/mL Catalase, 50% v/v Glycerol).
3. After the LMA solution reaches 37 °C, prepare the STORM buffer by adding 100 mM Mercaptoethylamine (MEA) and GLOX to the warm LMA solution to achieve 1× final concentration. Store the MEA 1 M stock solution for up to 1 month at 4 °C, but the use of a freshly prepared solution is recommended. Mix gently.
   NOTE: For optional DNA labeling, add 1.5 nM (final concentration, 1 mM stock solution in DMSO) of the JF549 conjugated Hoechst. Keep the suspension at 37 °C until use.
4. Gently mix the nuclei suspension with nanodiamonds at 1/1500 dilution. Incubate them with the sample for 5 min.
   NOTE: Nanodiamonds serve as fiducial markers to correct the sample drift.
5. Deposit 10 µL of the mix on an ozone-treated 1.5H coverslip (20 min in Ultraviolet Ozone cleaning system) and wait for 3 min to let the nuclei deposit on the coverslip.
6. Remove the excess of liquid, wait 2 min. Add 10 µL of the LMA prepared at step 3.3 to the coverslip and cover immediately with a microscope slide.
7. Seal the coverslip using silicone sealant. Mix equal amounts of solution A and B in a small Petri dish using a micropipette tip. Apply the mix over the edges of the coverslip to seal the sample. Silicone sealant hardens within 5 min.
8. Once mounted, perform imaging of slides without delay, ultimately in the following 4–5 h, to avoid loss of the oxygen scavenging properties of the mounting pad.

4. Imaging

NOTE: Imaging parameters must be carefully optimized according to the nature of the sample, the photo-physical properties of the fluorophores, and the optical configuration of the microscope. It is crucial to ensure a high signal-to-noise ratio for each fluorophore while minimizing spectral bleed-through and cross-excitation between fluorescent channels. Samples were observed with an inverted fluorescence microscope equipped with a high NA 100x oil immersion objective, a focus maintaining unit, and a white LED for bright-field imaging. Fluorescent labels AlexaFluor647 (AF647) were excited at 640 nm and JF549-Hoechst at 561 nm. The emitted light was filtered with a compatible dichroic mirror and band-pass filter(s), and finally, images were collected with a scientific Complementary Metal-Oxide-Semiconductor (sCMOS) camera.

1. Mount the sample slide and find focus. Activate a focus-maintaining module if the system has one (highly recommended for long acquisitions).
2. Find and focus on a nucleus. It is better to have one or a few nanodiamonds near the nucleus of interest to more efficiently correct the drift during the reconstruction.
3. (Optional) Capture diffraction-limited snapshots of bright field and each fluorescent channel at low laser power.
4. When performing two-color imaging, start with the acquisition of the longer wavelength channel and then switch to the shorter wavelength channel (i.e., acquire AF647 channel before JF549).
   NOTE: Example of acquisition parameters used in this work:
   Channel A - epitope labeling - AF647
   Exposure time: 20 ms
   Laser intensity: 640 nm – 2.2 kW/cm²
   Excitation mode: HiLo²⁷
   Channel B - DNA labeling - JF549-Hoechst
   Exposure time: 50 ms
   Laser intensity: 561 nm - 1 kW/cm²
   Excitation mode: Epi
   Number of recorded images: 10’000-30’000 per channel, ensuring that the density of fluorophores per frame remains sparse enough for accurate single-molecule localization, while the cumulative number of localizations was empirically determined using Super-resolution QUantitative Image Rating and Reporting of Error Locations (SQUIRREL)²⁸, which also indicates the sufficient number of frames.

Results

The protocol described above enables the preparation of high-quality Arabidopsis thaliana nuclei suitable for SMLM. Several rounds of optimization were performed to improve nuclear integrity and reduce cytoplasmic and cellular debris that compromise image quality. After each optimization step, the quality of the nuclear preparations was evaluated by confocal imaging using an Olympus IX81 spinning-disk microscope before proceeding to super-resolution acquisition.

Confocal imaging
The procedure described above is indeed compatible with confocal imaging. For this purpose, the LMA solution used in step 3.1 can be based on PBS (or any neutral buffer or medium). MEA and GLOX are only necessary for SMLM. On the other hand, the JF-based Hoechst used for SMLM (step 3.3) is too diluted to be detectable by conventional imaging methods. DNA labeling for confocal or epifluorescence imaging was thus achieved using Hoechst H33258 at 1/500 dilution (stock at 1 mg/mL).

Using the present optimized protocol, nuclei were isolated from 15-day-old Arabidopsis wild-type seedlings. The actively elongating form of RNA Polymerase II, marked by phosphorylation of Ser2 of the heptapeptide tandem repeats on the C-terminal domain (CTD), was immunostained with a specific primary antibody²⁹ and an AF647-conjugated secondary antibody²⁶. Confocal inspection illustrated in Figure 1 shows various isolated and well-preserved nuclei with minimal debris, most of which did not show Hoechst, AF647, or chlorophyll autofluorescence signals. Chromocenters (revealed by Hoechst) and the nucleolus were clearly distinguishable. Notably, RNA Polymerase II signal was absent from both chromocenters and the nucleolus (arrows), consistent with the expected localization of the enzyme in its active, transcriptionally engaged form.

Figure 1: Examples of three different ROIs with extracted nuclei (rows), displaying the active RNA Polymerase II  immunolabeled with AF647-conjugated antibody (red), and DNA counterstained with Hoechst H33258 (cyan). Confocal imaging was used to assess sample quality prior to Single-Molecule Localization Microscopy (SMLM). The preparation contains multiple intact, well-preserved nuclei clearly separated from debris, with chromocenters and nucleoles readily visible (arrows), indicating suitability for subsequent SMLM. DIC: Differential Interference Contrast. Scale bar: 3 μm (all panels). Please click here to view a larger version of this figure.

STORM imaging
For super-resolution imaging, samples were stained for RNA Polymerase II and counterstained for DNA with JF549-Hoechst. Among a few DNA-binding SMLM dyes (e.g., SiR-DNA, JF646-Hoechst), JF549-Hoechst was chosen as a second label in combination with AF647 for two-color imaging. Ten to thirty thousand frames were sequentially acquired for AF647 (RNA Pol II) and JF549 (DNA). Representative results are shown in Figure 2. To simulate diffraction-limited images, the maximum intensity projections (MIP) of all detected blinking events in the raw acquisitions were generated using ImageJ2/Fiji^30,31 (Figure 2, Diffraction-limited MIP). To achieve sub-diffraction-limited localizations of single-molecule signals, the raw image stacks were processed with the ThunderSTORM plugin in Fiji³² (Figure 2, SMLM). The typical analysis workflow is detailed below. The corresponding images were generated using ThunderSTORM average shifted histogram representation.

SMLM data processing by ThunderSTORM was obtained using Gaussian fitting with Weighted Least-Squares (WLSQ) estimation (Figure 3A, top). To further improve the quality and resolution of the reconstructed images, additional filtering over sigma and uncertainty was applied. An appropriate sigma value corresponds to the expected width of a true single-molecule emission, whereas excessively small or large values typically represent noise and overlapping or slightly defocused emitters, leading to reconstruction artifacts, and therefore were filtered out (Figure 3A and histograms in Figure 3B). On the other hand, localization uncertainty values can be used to select the best localizations with respect to the calculated fit accuracy³³. Uncertainty reflects how precisely ThunderSTORM can localize a single molecule based on photon statistics and background noise. Localizations were therefore filtered on their uncertainty for values below 30 nm (Figure 3A and histograms in Figure 3C; Supplementary File 1). Following this second filtering step, detections with poor positional precision were removed, resulting in sharper molecular signals and reduced background noise.

Sample drift correction was carried out using ThunderSTORM cross-correlation or the COMET (Cost-function Optimized Maximal overlap drift EsTimation) algorithm (https://github.com/gpufit/comet). Note that in Figure 3A, the drift was minimal and does not appear on the non-corrected images.

After reconstruction, the nucleolus and the overall distribution of DNA within the nucleus became clearly resolved. Active RNA Pol II formed small clusters throughout the nucleoplasm and was distinctly excluded from compact DNA regions, consistent with its transcriptional activity. JF549-Hoechst produced a strong signal under SMLM conditions; Surprisingly, reconstructed images showed reduced detection of chromocenters compared with confocal images of Hoechst H33258-labeled DNA, although occasional dense DNA foci remained distinguishable (Figure 2 or Figure 3, close-up views).

To quantify the improvement in image resolution during processing, we performed Fourier Ring Correlation (FRC) analysis. FRC provides an objective estimate of spatial resolution by comparing two independent reconstructions of the same dataset and measuring their correlation across spatial frequencies³⁴. This method, therefore, enables direct assessment of resolution gains from the fitted localizations to the fully processed data. In the example shown in Figure 4, the resolution of the AF647 channel improved from 90.1 nm to 47.7 nm after filtering (Figure 4A–C). Similarly, the resolution of the JF549 channel increased from 56.6 nm to 27.9 nm following processing.

The representative results presented here illustrate the effectiveness of the optimized nuclear preparation protocol in preserving nuclear integrity and enabling high-quality SMLM imaging. Confocal inspection confirms that isolated nuclei are largely free of cytoplasmic debris, with chromocenters and nucleoli clearly distinguishable, demonstrating that the protocol maintains structural features critical for downstream super-resolution analysis. The subsequent SMLM reconstructions further highlight the technique’s capacity to resolve sub-diffraction molecular distributions, as exemplified by the clustered localization of active RNA Polymerase II within the nucleoplasm and its exclusion from compact DNA regions. Image processing in ThunderSTORM, including filtering based on sigma and localization uncertainty, as well as drift correction, provides a robust framework to refine single-molecule localizations and reduce background noise. Fourier Ring Correlation analysis quantitatively confirms the resolution enhancement achieved through these steps, offering a reliable metric for assessing the quality of reconstructed images. Together, these results validate the protocol and provide a clear workflow for interpreting SMLM data, allowing users to identify both global nuclear organization and fine-scale molecular patterns.

Figure 2: Two-dimensional STORM imaging of Arabidopsis thaliana Col-0 wild-type nucleus immunostained with an antibody against the active RNA Polymerase II, and counterstained for DNA (JF549-Hoechst). Diffraction-limited maximum intensity projection (MIP) of the acquired streams for AF647 (left), JF549 (middle), and merged channels (right) is shown for an isolated nucleus. Corresponding ThunderSTORM reconstructions are displayed below. A 1.2 µm × 1.2 µm region of interest was selected and enlarged to visualize finer structural details. Scale bars: 1 µm. Please click here to view a larger version of this figure.

Figure 3: SMLM localization filtering workflow. (A) Localizations obtained after applying the weighted least-squares fitting method in ThunderSTORM (top, detected localizations). A 1.2 µm × 1.2 µm region of interest is enlarged to highlight structural details. The image appears blurred and partially saturated due to the presence of poorly fitted or out-of-focus fluorophore localizations. Localizations remaining after sigma and uncertainty filtering (middle). Both AF647 (RNA Pol II) and JF549 (DNA) channels show substantially improved definition compared to the top panel. The drift was corrected using the COMET algorithm (bottom). (B) Corresponding distributions of sigma values for the raw localizations with the selected filtering ranges highlighted in red. Sigma filtering was applied by selecting the dominant peak of the distribution and removing events with excessively small or large sigma values. (C) Corresponding distributions of uncertainty values for the raw localizations with the selected filtering ranges highlighted in red. Localizations with high uncertainty were eliminated. Scale bars: 1 µm. Please click here to view a larger version of this figure.

Figure 4: Fourier Ring Correlation (FRC) analysis of the nucleus presented in Figure 3A. FRC curves estimate the effective spatial resolution for the AF647 and JF549 channels following sequential application of (A) WLSQ analysis, (B) sigma filtering, and (C) uncertainty filtering. After applying all filters, the resolution improved from 90.1 nm to 47.7 nm for the AF647 channel and from 56.6 nm to 27.9 nm for the JF549 channel. The decrease in FRC-derived resolution values indicates enhanced ability to resolve finer structural details. Please click here to view a larger version of this figure.

Supplementary File 1: ThunderSTORM fitting parameters. The raw ThunderSTORM protocol file corresponding to the fitted image includes the plugin version used in the present work, specific camera settings, and detailed settings for the WLSQ fitting method. Further post-processing filtering parameters include the sigma and uncertainty ranges.Please click here to download this file.

Discussion

The preparation of Arabidopsis thaliana nuclei for single-molecule localization microscopy (SMLM) requires several critical steps that strongly influence the final sample quality. As described previously^19,35, thorough and consistent tissue chopping is essential to maximize the recovery of intact nuclei. Using microtome blades instead of standard razor blades produces cleaner and more uniform tissue fragmentation, thereby improving nuclei release. Subsequent washing and centrifugation steps are equally important for reducing debris. These steps are performed carefully, with the supernatant removed slowly using a 200 µL micropipette, followed by a 10 µL micropipette to minimize nuclei loss. When the pellet is particularly compact, increasing the resuspension volume of NIB buffer to 200–250 µL helps prevent aggregation and ensures a more homogeneous suspension.

Optional modifications can further enhance sample purity and imaging quality. A brief sonication step before the final wash removes debris and disperses aggregated nuclei¹⁹; however, excessive sonication can damage the nuclear envelope. Additionally, the imaging buffer can be degassed under vacuum to remove dissolved oxygen, which improves fluorophore blinking—particularly for AF647 and eliminates the need for glucose oxidase. In this approach, the buffer is prepared using freshly made Tris (pH 8) containing only MEA and LMA, since degassed solutions rapidly reoxygenate at atmospheric pressure and are unsuitable for extended storage.

Despite the improvements achieved with the optimized protocol, some limitations remain. In certain preparations, nuclei can still be positioned near debris or adjacent to other nuclei, necessitating screening multiple fields of view to identify isolated examples. For drift correction, fiducial markers such as nanodiamonds provide excellent stability, but they are not always located in proximity to the nucleus of interest. In such cases, computational correction using the COMET algorithm is efficient, although minor chromatic shifts may still be observed.

The protocol also incorporates several improvements over existing methods. Conducting immunolabeling within microcentrifuge tubes, rather than directly on slides, minimizes nuclei loss and drying artifacts, and enables long-term storage of the preparations at 4 °C. Additionally, incorporating the DNA counterstain directly into the agarose pad enhances signal uniformity, reduces background fluorescence, and eliminates extra washing steps that risk dislodging nuclei. As with any immunolabeling protocol, appropriate validation controls should be performed, including epifluorescence, confocal, and super-resolution microscopy, as well as single-labeled samples to assess spectral cross-talk and primary antibody-free controls to confirm labeling specificity and rule out non-specific secondary antibody binding or autofluorescence.

Overall, this pipeline enables consistent high-resolution imaging of plant nuclei while overcoming common plant-specific challenges, including structural complexity and autofluorescence. It allows chromatin organization and epigenetic marks to be resolved at the nanometer scale and can be readily applied to studies of nuclear architecture in Arabidopsis and other plant species, including crops³⁶. Beyond the applications demonstrated here, the presented method could eventually be coupled to locus-specific DNA labelling or to single-molecule fluorescence in situ hybridization (smFISH) that allows visualization of gene expression at a single DNA locus. These approaches employ multiple labeled DNA oligonucleotides to hybridize different portions of DNA or transcript introns³⁷. As a starting point, large intron-rich genes would particularly be suitable to examine the spatial organization of transcriptional outputs in the context of the nuclear architecture.

The present protocol focuses on robust nuclei preparation optimized for chromatin preservation and compatibility with super-resolution microscopy. Nuclei being isolated from bulk tissues, the obtained preparations represent a heterogeneous population originating from multiple cell types and positional contexts within the seedling (shoot and root tissues). Consequently, chromatin features observed by super-resolution imaging should be interpreted with this cellular heterogeneity in mind. However, the method could in principle be further combined with strategies for cell-type-specific nuclei enrichment or nuclei sorting. Indeed, additional steps corresponding to technologies already developed in plants, such as isolation of nuclei tagged in specific cell types -INTACT-³⁶ or fluorescence-based nuclei sorting -FANS-³⁸ could be inserted in the nuclei isolation workflow, likely with optimizations, including the use of milder fixative concentrations.

Altogether, these potential extensions highlight the versatility of the method and provide a framework for investigating chromatin organization and gene expression at the single-molecule level in plant nuclei, following eventual pre-selection of nucleus types.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Anti-RNA polymerase II CTD repeat YSPTSPS (phospho S2) antibody	Abcam	ab5095	Primary antibody used at 1/200 dilution in immunostaining buffer to visualize the active form of the polymerase (Polyclonal antibody raised in rabbit )
Acquisition software	Abbelight	NEO LiveImaging v2.18
Bovine Serum Albumin	Sigma-Aldrich	A7906
Catalase	Sigma-Aldrich	C40
Certified Low Melting Agarose	Bio-Rad	1613111
Coverslips thickness 1.5H round	Marienfeld	117640	Ø: 24 mm
D-(+)-Glucose, anhydrous, 99%	Thermo Scientific	A16828-36
Dichroic mirror	Semrock	Di03-R405/488/561/635-t1	Dichroic mirror for excitation/emission separation
Dulbecco’s Phosphate Buffered Saline 10´	biowest	X0515 - 500	Work in sterile conditions
Epredia Ultra Disposable Microtome Blades	FisherScientific	12191830	Low profile, single use
Falcon 6-well Clear Flat Bottom TC-treated Multiwell Cell Culture Plate, with Lid	Falcon	353046
Formaldehyde solution 37%	Carl Roth	7398	Work under the fume hood
Glucose oxidase	Sigma-Aldrich	G2133
Glycerol	VWR	24388.295
Goat anti-Rabbit IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor Plus 647	Thermo Scientific	A32733TR	Used at 1/200 as secondary antibody coupled to AF647
Hoechst 33258 solution	Sigma-Aldrich	94403	Used at 1/500 to counterstain the DNA, added to the agarose pad
Hydrochloric Acid Solution	Chem-Lab	CL05.0311.1000
JF549-Hoechst	Lavis lab and Open Chemistry team (Janelia)	JF549-Hoechst	Used to counterstain DNA, added to the agarose pad at 1.5 nM final concentration
KIMBLE Dounce tissue grinder set	Sigma-Aldrich	D8938	2 mL tube with both large (0.0030-0.0050 inch) and small (0.0005-0.0025 inch) clearance pestles
Lasers unit	Oxxius	L6Cc	Equipped with six lasers: 405 nm - 100 mW, 488 nm - 200 mW, 532 nm - 500 mW, 561 nm - 300 mW, 640 nm - 500 mW and 730 nm - 30 mW
Magnesium chloride hexahydrate	Carl Roth	HN03
Mercaptoethylamine (MEA)	Sigma-Aldrich	30070
MOPS sodium salt 98%	Fisher Scientific SAS	352590010
Multiband emission filter	Semrock	FF01-446/523/600/677	Multiband emission filtration to block laser scatter light
Nanodiamonds	Adamas Nanotechnologies	NDNV100nmHiWGA2ml	Stock 1 mg/mL
ORCA-Fusion Digital CMOS camera	Hamamatsu	C14440-20UP
Petri dish, 100/20 mm, PS, clear, with vents, sterile	Greiner Bio-One	664161	Used to grow plants on 1/2 MS medium
Petri dish, square, PS, clear, 120/120/17 mm, sterile	Greiner Bio-One	688161	Used when chopping the plant seedlings
pluriStrainer Mini cell strainer	pluriSelect	43-10030-50	Mesh size 30 µm, suitable for 1.5 mL Eppendorf tube
Potassium Chloride	Sigma-Aldrich	P9333
Single band emission filter	Semrock	FF01-698/70	Specific emission filter for AF647 channel
Single band emission filter	Chroma	ET600/50m	Specific emission filter for JF549 channel
Sodium chloride (99,5%)	Euromedex	1112-A
Star-Frost slides 76 x 26 mm	Knittel	VS112711FKB.01
Sterile syringe filterø 33 mm porosity 0.2 µm Luer Lock cone	ClearLine	146560
Super-resolution system	Abbelight	SAFe360	Olympus IX83 equipped with 100x NA1.5 oil-immersion objective
Tri-Sodium Citrate 2H2O Gen-Apex Biomole	Prolabo	PRO-33615.268
Tris Base Molecular Biology Grade	Promega	H5135
Tris(2-carboxyethyl)phosphine hydrochloride (TCEP)	Thermo Scientific	20491
Triton X-100	Euromedex	2000-B
Tween 20	Euromedex	2001-B
Twinsil Speed	Picodent	13001002	Silicone sealant
Ultraviolet Ozone cleaning system UVOCS oven	UVOCS Inc.	T10X10	20 min exposure for coverslip cleaning
Vacuum pump	Vacuubrand	VP 100C
Heraeus Megafuge 16R Centrifuge	Thermo Scientific		Rotor TX-400 75003629. Centrifugation performed with Eppendorf tubes.