This protocol presents a technique for high-resolution mapping of replication sites in structurally preserved chromatin in situ that employs a combination of pre-embedding EdU-streptavidin-Nanogold labeling and ChromEMT.
Principles of DNA folding in the cell nucleus and its dynamic transformations that occur during the fulfillment of basic genetic functions (transcription, replication, segregation, etc.) remain poorly understood, partially due to the lack of experimental approaches to high-resolution visualization of specific chromatin loci in structurally preserved nuclei. Here we present a protocol for the visualization of replicative domains in monolayer cell culture in situ, by combining EdU labeling of newly synthesized DNA with subsequent label detection with Ag-amplification of Nanogold particles and ChromEM staining of chromatin. This protocol allows for the high-contrast, high-efficiency pre-embedding labeling, compatible with traditional glutaraldehyde fixation that provides the best structural preservation of chromatin for room-temperature sample processing. Another advantage of pre-embedding labeling is the possibility to pre-select cells of interest for sectioning. This is especially important for the analysis of heterogeneous cell populations, as well as compatibility with electron tomography approaches to high-resolution 3D analysis of chromatin organization at sites of replication, and the analysis of post-replicative chromatin rearrangement and sister chromatid segregation in the interphase.
DNA replication is a basic biological process required for faithful copying and transmission of the genetic information during cell division. In higher eukaryotes, DNA replication is subjected to tight spatio-temporal regulation, which is manifested in sequential activation of replication origins1. Neighboring replication origins firing synchronously form clusters of replicons2. At the level of optical microscopy, sites of ongoing DNA replication are detected as replication foci of various number and size. Replication foci display specific patterns of spatial distribution within the cell nucleus depending on the replication timing of the labeled DNA3,4, which, in turn, is tightly correlated with its gene activity. Thanks to well-defined sequence of DNA replication, strictly ordered in space and time, replicative labeling is a powerful method of precise DNA labeling not only for the study of replication process per se, but also to discriminate a specific DNA sub-fraction with defined transcription activity and compaction level. Visualization of replicating chromatin is usually performed through the detection of major protein components of DNA replication machinery (either by immunostaining or by expression of fluorescent protein tags5,6) or by the incorporation of modified DNA synthesis precursors7,8,9,10. Of these, only methods based on the incorporation of modified nucleotides into newly replicated DNA allow for the capture of conformational changes in chromatin during replication, and trace the behavior of replicative domains after their replication is completed.
In higher eukaryotes, DNA packaging into chromatin adds another level of complexity to the regulation of basic genetic functions (transcription, replication, reparation, etc.). Chromatin folding affects accessibility of DNA to regulatory trans-factors and DNA conformational changes (double helix unwinding) required for the template synthesis. Therefore, it is generally accepted that DNA-dependent synthetic processes in the cell nucleus require a structural transition of chromatin from its condensed, repressive state to a more accessible, open conformation. Cytologically, these two chromatin states are defined as heterochromatin and euchromatin. However, there is still no consensus concerning the mode of DNA folding in the nucleus. The hypotheses range from a "polymer melt" model11, where nucleosomal fiber behaves as a random polymer for which the packing density is controlled by phase separation mechanisms, to hierarchical folding models postulating sequential formation of chromatin fiber-like structures of increasing thickness12,13. Hierarchical folding models recently gained support from molecular approaches based on the analysis of in situ DNA-DNA contacts (chromosome conformation capture, 3C), demonstrating the existence of the hierarchy of chromatin structural domains14. It is important to note that replication units correlate very well to these chromatin domains15. The major criticism of these models is based on potential artificial chromatin aggregation caused by sample preparation procedures, such as permeabilization of cell membranes and removal of non-chromatin components, in order to improve chromatin contrast for ultrastructural studies while improving chromatin accessibility for various probes (e.g., antibodies). Recent technical advances in selective DNA staining for electron microscopy by DNA-binding fluorophore-mediated photo-oxidation of diaminobenzidine (ChromEMT6) has allowed for the elimination of this obstacle. However, the same considerations hold true for electron microscopy visualization of replicating DNA17,18. Here we describe a technique that allows for the simultaneous high-resolution ultrastructural mapping of newly synthesized DNA and total chromatin in intact aldehyde-crosslinked cells. The technique combines detection of EdU-labeled DNA by Click-chemistry with biotinylated probes and streptavidin-Nanogold, and ChromEMT.
The protocol is optimized for adherent cells and was tested on HeLa, HT1080, and CHO cell lines.
1. Cell labeling and fixation
2. Click-reaction
NOTE: This procedure is modified from a previously published protocol19.
3. Ag-amplification
NOTE: This procedure is modified from Gilerovitch et al., 1995 (see 20,21).
4. Gold toning
NOTE: In order to protect silver nanoparticles from dissolution by OsO4 oxidation in the subsequent procedures, an impregnation with gold is used at this step (Sawada, Esaki, 1994)22.
5. ChromEM
NOTE: This protocol was modified from Ou et al., 201716.
6. Dehydration and epoxy resin embedding
7. Electron tomography
Replication foci in mammalian cell nuclei display distinct patterns of distribution within the nucleus depending on S-phase progression. These patterns correlate with transcriptional activity of the loci being replicated. Since the method presented here utilizes a rather strongfixation procedure, it is fairly straightforward to use replicative pulse labeling for specific detection of chromatin loci in various transcriptional states, even under conditions offering best structural preservation of chromatin obtained by room temperature chemical fixation.
Quality control steps are intended to assure successful completion of the key procedures in this protocol. Initially, successful EdU labeling should be confirmed by Click-reaction with fluorescently labeled azide, demonstrating typical replication patterns if heterogeneous cell population is used (Figure 2). The second important step is streptavidin labeling, which can be strongly influenced by glutaraldehyde fixation. For the evaluation of streptavidin binding efficiency, use AlexaFluor-488-conjugated streptavidin under the same conditions as streptavidin-Nanogold (Figure 3). At this point, some background is expected mainly due to glutaraldehyde autofluorescence as well as incomplete streptavidin wash-out. If the signal-to-noise ratio is unacceptable, try increasing the number and duration of washing steps at step 2.6. Alternatively, add glutaraldehyde quenching with NaBH4 (steps 2.9-2.10) after step 1.5.
The silver enhancement procedure often produces inconsistent results and requires optimization. First, the reaction speed varies dramatically depending on the acacia powder solution batch. The viscosity of the solution is inversely proportional to the time of the silver staining development. It is a good idea to have several aliquots of the same batch and test them on the control samples to determine optimal reaction timing. As the temperature of the reagents and the environment also has a significant impact on the reaction speed, try to perform the reaction at the exact same temperature. If processing more than three samples simultaneously, it is convenient to start the reaction at 30 second intervals to have enough time for washing steps.
A good outcome of the silver enhancement procedure is the dark brown staining of some (not all) nuclei with very faint yellowish cytoplasmic staining (Figure 4). This means the mean size of silver particles is about 20 nm, which is suitable for easy label detection in a broad magnification range of TEM over the chromatin counterstaining. For tomography experiments, the size of the particles should be reduced to about 10 nm by shortening the reaction time. The color in this case should be light brown or reddish. The color of silver staining should be checked before gold toning.
DAB photoconversion and osmification is performed after gold toning to protect silver shells from dissolution. In case of insufficient gold impregnation, silver nanoparticles become eroded and acquire irregular shape, but still remain visible under TEM. The quality of DAB photoconversion is monitored first by DRAQ5 bleaching (in fluorescence mode), then in a brightfield mode, when the cell nuclei in an irradiated field become dark brown (Figure 5). DAB staining intensity should not be very high so that Ag-Au staining remains well detected to allow for the selection of the cells with required replication timing for sectioning.
Ultrathin or semi-thin sections do not require additional staining and can be viewed directly under TEM. The appropriate labeling typically results in well-defined clusters of Ag nanoparticles demarcating replication sites (Figure 6), while background labeling is limited to a few nanoparticles per square micrometer. Semi-thin sections are ready for tomography and do not require addition of gold nanoparticles to the section surface, as Ag nanoparticles present in a section can be used as fiducial marks (Figure 7). In this figure, the effect of prolonged labeling time is shown: individual replication foci fuse into fiber-like structures, delineating higher-order chromatin domains.
Figure 1. The method overview. Cells grown on coverslips are labeled with EdU, fixed with 2.5% glutaraldehyde, permeabilized, and Click-reaction is performed with biotinylated azide. Alternatively, fluorescently labeled azides are used in light microscopic observation. Sites of biotin incorporation are labeled with Nanogold-streptavidin (or fluorescently labeled streptavidin as a control), silver enhanced, and after gold toning subjected to DRAQ5-mediated DAB photoconversion and osmification. The samples are embedded in epoxy resin, sectioned, examined in TEM, and subjected to electron tomography. Please click here to view a larger version of this figure.
Figure 2. Replication patterns in mammalian cells revealed by EdU labeling and Click-reaction (green). S-phase progression from early to late S-phase (from left to right) is manifested by an orderly change in replication patterns: A,B – early S-phase, C – mid S-phase, D,E – late S-phase. DNA is stained with DAPI (red). Scale bar = 10 um. Please click here to view a larger version of this figure.
Figure 3. Replication patterns in mammalian cells revealed by EdU labeling and two-step Click-biotin and streptavidin-AlexaFluor 488 staining after glutaraldehyde fixation. Control sample labeled with streptavidin-AlexaFluor 488 displays optimal labeling efficiency and S/N ratio after glutaraldehyde fixation. Please click here to view a larger version of this figure.
Figure 4. Representative image of cells after silver enhancement procedure (Step 3). Various replicative patterns (early S-phase, arrows; mid S-phase, arrowhead; late, double arrows) are readily visible in a brightfield light microscopy mode. Please click here to view a larger version of this figure.
Figure 5. A typical outcome of DRAQ5-mediated DAB photoconversion. (A) DRAQ5 staining (dashed circle indicates irradiated area). Note strong DRAQ5 bleaching inside the circle. (B) The same area seen in brightfield light microscope. Note stronger DAB precipitation in the nuclei in irradiated area. Inset: arrows indicate early S-phase nuclei where replication patterns labeled with Ag-Au are still readily detectable on the background of DRAQ5-photo-oxidazed DAB staining. Please click here to view a larger version of this figure.
Figure 6. Replication foci in mammalian cell nuclei revealed by EdU labeling and two-step Click-biotin and streptavidin-Nanogold staining after glutaraldehyde fixation. Thin 90 nm sections of an S-phase cell showing clusters of Ag-Au nanoparticles at replication foci (A, red circles). The background level can be appreciated in a non-S-phase cell (B). Arrowheads indicate chromatin fibers of various scale. Please click here to view a larger version of this figure.
Figure 7. 0°-tilt projection of 250 nm section. (A) and virtual tomographic section. (B) of a cell nucleus labeled with EdU for 2 h. See individual replication foci fused into fiber-like structures (arrowheads). Please click here to view a larger version of this figure.
The method described here has several advantages over previously published protocols. First, the use of Click-chemistry for labeling replicated DNA eliminates the necessity of DNA denaturation prerequisite for BrdU detection with antibodies, thus better preserving chromatin ultrastructure.
Second, utilization of biotin as a secondary ligand that is generated after glutaraldehyde fixation and proper quenching of unbound aldehyde groups minimizes chemical modification of the target, thus improving labeling efficiency while reducing background due to endogenous biotin. Biotin-streptavidin also adds versatility as the use of fluorescently labeled streptavidin provides easy quality control at the very early stage of the procedure. The protocol can be further simplified by introduction of FluoroNanogold-labeled streptavidin, however, in our hands these reagents gave somewhat higher background compared to Nanogold-streptavidin, maybe due to larger size of the probes.
Third, a combination of small probes, streptavidin-Nanogold being the largest component of about 5 nm, provides very good penetration into even glutaraldehyde-crosslinked cells. This makes the technique suitable for pre-embedding labeling of optimally ultrastructurally preserved cells and easily compatible with various 3D-electron microscopy techniques, including serial section reconstruction, array tomography, electron tomography, serial block face imaging, and FIB-SEM.
Silver enhancement is the most critical step, requiring precise control of reagent diffusion and washing steps, which is easier to perform with adherent cells. On the other hand, since strong crosslinking with glutaraldehyde can also affect the silver enhancement, the fixation conditions should be fine-tuned in order to balance chromatin structure preservation and silver enhancement efficiency. For this reason, cryo-fixation/cryo-substitution techniques, which require extended and poorly controlled post-fixation, although allowing even better structure preservation, might not be fully compatible with the proposed labeling strategy23.The technique can be further improved by using Au-amplification instead of silver enhancement24. This modification allows to skip the gold toning step, however, in our hands, it gives considerably higher background.
Finally, pre-embedding labeling allows pre-selection of target cells for ChromEM and sectioning based on the replication pattern detectable under brightfield or fluorescent microscope. Moreover, the method can be easily extended to CLEM, including various super-resolution techniques. This opens new horizons in studies of chromatin higher order structure and dynamics in various physiological states. The approach we described here can be used for studies of chromatin organization at sites of replication, for analysis of post-replicative rearrangement of chromatin, including high-resolution imaging of chromatid segregation in interphase, and for replicative labeling of large chromosomal domains and studies of higher-order chromatin folding of specific fractions of the genome (euchromatin or heterochromatin, labeled based on replication timing). These types of imaging techniques are also important as a reference point for the data generated by alternative approaches.
The authors have nothing to disclose.
This work was supported in part by RSF (grant #17-15-01290) and RFBR (grant #19-015-00273). The authors thank Lomonosov Moscow State University development program (PNR 5.13) and Nikon Center of Excellence in correlative imaging at Belozersky Institute of Physico-Chemical Biology for access to imaging instrumentation.
Reagent | |||
5-ethynyl-2`-deoxyuridine (EdU) | Thermo Fisher | A10044 | |
2-(4-Morpholino)ethane Sulfonic Acid (MES) | Fisher Scientific | BP300-100 | |
AlexaFluor 555-azide | Termo Fisher | A20012 | |
biotin-azide | Lumiprobe | C3730 | |
Bovine Serum Albumine | Boval | LY-0080 | |
DDSA | SPI-CHEM | 26544-38-7 | |
DMP-30 | SPI-CHEM | 90-72-2 | |
DRAQ5 | Thermo Scientific | 62251 | |
Epoxy resin monomer | SPI-CHEM | 90529-77-4 | |
Glutaraldehyde (25%, EM Grade) | TED PELLA, INC | 18426 | |
Gum arabic | ACROS Organics | 258850010 | |
Magnesium chloride | Panreac | 141396.1209 | |
NaBH4 | SIGMA-ALDRICH | 213462 | |
NMA | SPI-CHEM | 25134-21-8 | |
N-propyl gallate | SIGMA-ALDRICH | P3130 | |
PBS | MP Biomedicals | 2810305 | |
Silver lactate | ALDRICH | 359750-5G | |
Streptavidin-AlexaFluor 488 conjugate | Termo Fisher | S11223 | |
Streptavidin-Nanogold conjugate | Nanoprobes | 2016 | |
tetrachloroauric acid | SIGMA-ALDRICH | HT1004 | |
Tris(hydroxymethyl)aminomethane (Tris) | CHEM-IMPEX INT'L | 298 | |
Triton X-100 | Fluka Chemica | 93420 | |
Instruments | |||
Carbon Coater | Hitachi | ||
Copper single slot grids | Ted Pella | 1GC10H | |
Cy5 fluorescence filter set (Ex620/60 DM660 Em700/75) | Nikon | Cy5 HQ | Alternatives: Zeiss, Leica, Olympus |
Diamond knife Ultra Wet 45o | Diatome | DU | Alternatives: Ted Pella |
Fluorescent microscope | Nikon | Ti-E | Alternatives: Zeiss, Leica, Olympus |
High-tilt sample holder | Jeol | ||
Rotator | Biosan | Multi Bio RS-24 | |
Transmission electron microscope operating at 200 kV in EFTEM mode, with high-tilt goniometer | Jeol | JEM-2100 | Alternatives: FEI, Hitachi |
Tweezers | Ted Pella | 523 | |
Ultramicrotome | Leica | UltraCut-E | Alternatives: RMC |
Software | |||
Image acquisition | Open Source | SerialEM (https://bio3d.colorado.edu/SerialEM/) | |
Image processing | Open Source | IMOD (https://bio3d.colorado.edu/imod/) |