Ethics statement
The research complies with all ethical regulations. Immune cell isolation from human blood and analysis of human primary tissue was approved by the ethics committee of the University Clinic Erlangen (reference numbers 22-321-Bp and 24-349-Bp).
Cell culture
MCF10A/AT cells were cultured in T-25 flasks (Greiner Bio-one, catalogue number (cat. no.) 690175) containing Dulbecco’s modified Eagle’s medium/F12 (Gibco, reference number (ref. no.) 21041-025) medium supplemented with 5% horse serum (Gibco, ref. no. 16050-112), 20 ng ml−1 of epidermal growth factor (Gibco, ref. no. PHG0311), 0.5 μg ml−1 of hydrocortisone (Sigma-Aldrich, ref. no. H0396), 100 ng ml−1 of cholera toxin (Sigma-Aldrich, cat. no. C8052-5MG), 10 μ ml−1 of insulin (Sigma-Aldrich, cat. no. I1882-100MG) and 1% penicillin–streptomycin (Sigma-Aldrich, cat. no. P0781-100ML). Cells were maintained at 37 °C in a humidified atmosphere with 5% CO2. On reaching 70%–80% confluency, cells were washed with 5 ml of Ca2+/Mg2+-free Dulbecco’s phosphate-buffered saline (DPBS) (Gibco, ref. no. 14190-094) and harvested by enzymatic dissociation by brief incubation with 0.05% trypsin ethylenediaminetetraacetic acid (EDTA) solution (Gibco, ref. no. 2500-054) for 8–10 min. Following trypsin neutralization with complete medium, cells were centrifuged at 3,000g for 5 min, and the resulting pellet was resuspended in 5 ml of complete growth medium. For experimental assays, cells were seeded into eight-well chamber slides (ibidi, ref. no. 80807-90) at a 1:10 dilution and incubated for 24 h prior sample preparation. For the TGFβ-treated variants of MCF10A/AT, cells were cultured in media containing 5 ng ml−1 of TGFβ (Bio-Rad, ref. no. PHP143B) for four subsequent passages. Then, the cells were continued in culture using complete media without TGFβ.
A549 cells were cultured in T-25 flasks containing RPMI-1640 medium (Gibco, ref. no. 31870-025) supplemented with 10% fetal bovine serum (Gibco, ref. no. A31605-02), 1% penicillin–streptomycin and 1% GlutaMAX (Gibco, ref. no. 35050-38). Cells were maintained at 37 °C in a humidified atmosphere with 5% CO2. On reaching 70%–80% confluency, cells were washed with 5 ml of Ca2+/Mg2+-free DPBS and harvested by enzymatic dissociation by brief incubation with 0.05% trypsin EDTA solution for 8–10 min. Following trypsin neutralization with the complete medium, cells were centrifuged at 3,000g for 5 min, and the resulting pellet was resuspended in 5 ml of complete growth medium. For co-culture experiments, cells were seeded into eight-well chamber slides at a 1:10 dilution and incubated for 24 h before experimental manipulations.
Isolation and expansion as well as sample preparation of primary NK cells
NK cells were isolated from peripheral blood using the NK cell isolation kit (Stem Cell Technologies, cat. no. 19665) following the manufacturer’s instructions, which involves density gradient centrifugation and magnetic bead separation. The isolated NK cells were then expanded using the NK cell expansion kit (Stem Cell Technologies, cat. no. 100-0711), which includes the base medium (Stem Cell Technologies, cat. no. 100-0712), supplement (Stem Cell Technologies, cat. no. 100-0715) and coating material (Stem Cell Technologies, cat. no. 100-0714). Briefly, a 24-well plate (Stem Cell Technologies, cat. no. 38044) was coated with the coating material and incubated for 2 h at room temperature and then rinsed with Ca2+/Mg2+-free DPBS. Isolated NK cells were seeded onto the coated plate at a density of 1 × 106 cells ml−1 in ImmunoCult NK Cell Expansion Medium and incubated at 37 °C and 5% CO2. On day 3 or 4, an additional expansion medium was added, and the cells were further incubated. On days 7 and 10/11, the cells were harvested via Ca2+/Mg2+-free DPBS and reseeded onto coated eight-well plates (2 × 105 cells μgml−1) in a fresh expansion medium and incubated for 72 h to prepare the samples for fixation.
Co-culture of NK cells and A549 cells
A549 cells were seeded at a 1:10 concentration onto NK-cell-surface-coated eight-well plates and incubated at 37 °C in a humidified atmosphere containing 5% CO2 for 24 h to allow for proper adherence and growth. Separately, previously seeded and expanded NK cells were resuspended for a passaging and co-culture experiment. For co-culture experiments, resuspended cells in fresh media were subsequently added to the A549 cells in a concentration of 2 × 105 cells ml−1. The co-culture was incubated at 37 °C in 5% CO2 for a duration of 5 min, during which NK cells adhered to the A549 cancer cells. The interaction between NK and A549 cells was closely monitored using light microscopy. Once the physical interactions between NK and A549 cells were observed, the samples were immediately fixed.
Preparation of tissue slices
Samples were obtained from the Central Biobank of the University Clinic Erlangen. Cryosections were fixed using 4% paraformaldehyde for 20 min and transferred to #1.5 glass slides. Before imaging, a chamber (ibidi sticky slides, ref. no. 80427) was placed on top of the glass slides and filled with PBS to facilitate staining, imager strand addition and washing steps.
Metabolic incorporation of Ac4ManNAz and copper-free click chemistry
For the metabolic labelling of sialic acids, the MCF10A/T panel was seeded into slides, as described above. After 3 h, the cell culture medium is supplemented with 50 µM of Ac4ManNAz and incubated for 72 h. After incubation, cells were washed twice with Ca2+/Mg2+-free DPBS and then refreshed with standard growth medium supplemented containing 50 µM of DBCO conjugated to docking strand R6 for 2 h in the incubator to facilitate the click reaction. Then, cells were fixed and stained with lectins.
Primary neuronal cultures
Hippocampi of E18 Sprague Dawley rats (Transnetyx, SKU number SDEHP) were washed thrice in ice-cold Hanks’s balanced salt solution (Thermo Fisher, cat. no. 14175095) with 1% penicillin–streptomycin (Thermo Fisher, cat. no. 15140122). Tissue was then incubated for 10 min in 0.05% trypsin EDTA (Gibco, cat. no. 15400054) at 37 °C. Trypsin was removed and the tissue was washed ten times in prewarmed Hanks’s balanced salt solution with 1% penicillin–streptomycin. Hanks’s balanced salt solution was replaced by 1 ml of warm Neurobasal Medium (Thermo Fisher, cat. no. 12348017) supplemented with 1% penicillin–streptomycin, 1% GlutaMAX (Thermo Fisher, cat. no. 35050061) and 2% B27 (Thermo Fisher, cat. no. 1750404). Cells were then mechanically dissociated by pipetting up and down about 30 times with a 200-µl pipette. Also, 100,000–150,000 cells were seeded on glass-bottomed Petri dishes (WPI, cat. no. FD35-100) coated with 10 µg ml−1 of poly-D-lysine (Sigma-Aldrich cat. no. p6407-5mg) and 1 µg ml−1 of laminin (Merck, cat. no. L2020-1MG) in 2 ml of Neurobasal Medium supplemented with 1% penicillin–streptomycin, 1% GlutaMAX and 2% B27. The medium was replaced the next day, and afterwards, half of the medium was replaced every 3 days.
Isolation, activation and preparation of primary human CD4+ T cells
CD4+ T cells were isolated from blood samples of healthy donors. First, peripheral blood mononuclear cells were collected after Ficoll gradient isolation (Ficoll-Paque PLUS, VWR). CD4+ T cells were then separated using commercially available CD4+ T cell isolation kit for human samples (Miltenyi), according to manufacturer’s instructions. Here 0.5 × 106 CD4+ cells were resuspended in Iscove’s modified Dulbecco’s medium (Gibco) supplemented with 10% fetal bovine serum (PanBiotech) and 1% penicillin–streptomycin (Gibco). To activate the cells, anti-human CD3 (α-CD3, 1 µg ml−1, Ultra-LEAF Purified anti-human CD3 Antibody, BioLegend) together with anti-human CD28 (α-CD28, 2 µg ml−1, Ultra-LEAF Purified anti-human CD28 Antibody, BioLegend) and recombinant human IL-2 (20 ng ml−1, ImmunoTools) were also provided. The cells were cultured for 3 days at 37 °C and 5% CO2. For control, non-stimulated and freshly isolated CD4+ T cells from the same donor were used. Samples were then fixed using 4% paraformaldehyde for 10 min at room temperature and washed three times with PBS before lectin labelling.
Isolation, stimulation and preparation of primary human neutrophils
Human neutrophils were isolated using a commercially available kit (MACSxpress Whole Blood Neutrophil Isolation Kit, Miltenyi), according to the manufacturer’s instructions. Then, 106 neutrophils were prepared in RPMI-1640 medium (Gibco) supplemented with 10% fetal bovine serum (PanBiotech) and 1% penicillin–streptomycin (Gibco). For stimulation conditions, recombinant human tumour necrosis factor-alpha (100 ng ml−1, ImmunoTools) was also added and the cells were cultured for 2 h at 37 °C and 5% CO2. Non-stimulated neutrophils were kept in the same conditions for culture, but without tumour necrosis factor-alpha. Samples were fixed using 4% paraformaldehyde for 10 min at room temperature and washed three times with PBS before lectin labelling.
Sample fixation
Samples were washed with DPBS and Ca2+/Mg2+ (Gibco, ref. no. 14040-091) three times. Then, 4% paraformaldehyde, diluted from a 16% stock solution (Thermo Scientific, ref. no. 28908) to 4% working solution via Ca2+/Mg2+-free DPBS was added to wells and cells were incubated at room temperature for 15 min. Then, cells were washed three times with Ca2+/Mg2+-free DPBS. After fixation, cells were permeabilized with 0.1% Triton X (Alfa Aesar, cat. no. A16046) for 10 min at room temperature, followed by four DPBS and Ca2+/Mg2+-free washing steps.
Lectin labelling
A lectin cocktail (2.5 μg μg ml−1 of each lectin) prepared in 1× Tris buffer (Fisher Bioreagents, ref. no. M-15836) was applied to the cells at room temperature for 30 min, allowing specific binding to glycan targets on the cellular surface. Following incubation, cells were washed three times with Ca2+/Mg2+-free DPBS. For the MCF10A panel cells, pre-permeabilization is performed using 0.1% Triton X-100 for 10 min at room temperature, followed by four washing steps with Ca2+/Mg2+-free DPBS and incubation with lectins as above. Before imaging, another permeabilization step using 0.1% Triton X-100 was performed for 10 min at room temperature, followed by four washing steps with Ca2+/Mg2+-free DPBS. For primary neurons, a permeabilization step using 0.1% Triton X-100 was performed for 10 min at room temperature, followed by four washing steps with Ca2+/Mg2+-free DPBS before imaging. For primary immune cells and tissue sections, samples were permeabilized with 0.2% Triton X-100 for 10 min at room temperature, followed by four washing steps with Ca2+/Mg2+-free DPBS. To test the effect of fixation, we inverted the protocol sequence and tested live-cell staining with WGA, followed by fixation. No detectable difference was observed (Supplementary Fig. 12).
Optical setup
DNA-PAINT imaging was carried out on an inverted microscope (Nikon Instruments, Eclipse Ti2) with the Perfect Focus System. Objective-based total internal reflection fluorescence (TIRF) mode was used, using a high-numerical-aperture objective (Nikon Instruments, Apo SR TIRF ×100, numerical aperture of 1.49, oil) and the Nikon TIRF module. A 560-nm laser (MPB Communications, 1 W) was used for excitation and coupled into the microscope via the TIRF module. The power of the laser beam was controlled in free space using a filter wheel (Thorlabs, FW212CNEB). The laser beam was passed through a clean-up filter (Chroma Technology, ZET561/10) and coupled into the microscope objective using a beamsplitter (Chroma Technology, ZT561RDC). Fluorescence was spectrally filtered using an emission filter (Chroma Technology, ET600/50m, and ET575LP) and imaged on a scientific complementary metal–oxide–semiconductor camera (Hamamatsu Orca Fusion) without further magnification, resulting in an effective pixel size of 130 nm after 2 × 2 binning. TIRF illumination was used for all the measurements with a laser power of ~33 mW above the objective. The central area of 1,152 × 1,152 pixel2 (576 × 576 after binning) of the camera was used as the region of interest. Raw microscopy data were acquired using μManager (v. 2.0.3).
DNA sequences
Docking and imager strand sequences are previously optimized for orthogonal binding specificity. The sequences used for DNA-PAINT were as follows:
R1 (WGA) with 5 × R1 sequence TCCTCCTCCTCCTCCTCCT
R2 (SNA) with 5 × R2 sequence ACCACCACCACCACCA
R3 (PHA-L) with 7 × R3 sequence CTCTCTCTCTCTCTCTCTC
R4 (AAL) with 7 × R4 sequence ACACACACACACACA
R5 (PSA) with 5 × R5 sequence CTTCTTCTTCTTCTT
R6 (DBCO) with 5 × R6 sequence AACAACAACAACAACAA.
All DNA oligonucleotides were purified via high-performance liquid chromatography and obtained from Metabion.
Imager strand preparation for DNA-PAINT imaging
Imaging buffer was prepared by combining 50 ml of Ca2+/Mg2+-free DPBS with 0.0146 g of EDTA (PanReac, cat. no. 60-00-4), 1.461 g of sodium chloride (Alfa Aesar, cat. no. A12313), and 10 µl of Tween-20 (MP Biomedicals, cat. no. 103368) in a 50-ml Falcon tube. All the components were thoroughly mixed until completely dissolved. The prepared buffer was stored at 4 °C until use in subsequent imager strand preparation. Imaging strands for DNA-PAINT were prepared by diluting 1 µM of the stock solutions in the buffer, the optimal imager concentration to achieve sparse blinking ranged from 0.075 to 0.5 nM, adjusted for sample types and targets, focusing on sparse single-molecule signals. Samples were incubated with a 1:3 dilution of 90-nm gold nanoparticles (Absource, cat. no. CG-90-20), which were used as fiducial markers for drift correction and channel alignment.
Multiplexed DNA-PAINT
DNA-PAINT imaging was conducted via six subsequent imaging rounds with only one imager type in each round. For each target, 20,000 frames of single-molecule blinks are captured with a frame time of 100 ms. A laser power of ~33 mW at the objective was used. Between imaging rounds, thorough washing for at least four times with Ca2+/Mg2+-free DPBS was performed, ensuring no residual signal from the previous imager solution remained before introducing the next imager solution.
Number of samples
For the MCF10A panel, data from two independent seedings are shown. Each seeding was staining independently. Data from at least two cells per staining are shown. For neurons, data from two independent seedings are shown. Each seeding was staining independently. Data from at least one cell per staining are shown. For NK cells, data from four independent isolations are shown. Each isolation was stained independently. Data from at least two cells per staining are shown. For T cells, data from one isolation are shown. Isolated cells were stained in two independent samples. Data from at least two cells per staining are shown. For neutrophils, data from two isolations are shown. Each isolation was stained in two independent samples. Data from at least one cell per staining are shown. For tissue sections, data from two independent slices are shown. Each slice was stained in two independent samples. Data from at least two FoVs are shown.
Statistics and reproducibility
No statistical methods were used to predetermine the sample sizes. No data were excluded from the analyses. The investigators were not blinded to allocation during experiments and outcome assessment.
Postprocessing
Raw fluorescence data were reconstructed using the Picasso software package27 (v. 0.7.4). For identifying distinct blinks in each frame, an intensity threshold of 5,000 was used. Drift correction was performed using redundant cross-correlation, followed by precise fiducial-based drift correction with gold nanoparticles. The six channels were aligned using gold nanoparticles as fiducial markers. The region of interest is segmented using the polygon pick function from Picasso (no signal from outside the cell is included in the analysis). All downstream analysis after segmentation until the calculation of cluster centres was performed using a custom version of Picasso in Python. An estimate of experimental localization precision is calculated for each individual channel of images using NN-based analysis68. For clustering, a radius of two times the experimental NN-based analysis precision was used. The minimum number of localizations inside a cluster was two. Time signatures of blinking events inside each cluster were analysed to account for unspecific sticking events. For a single cluster, if a time bin of 200 frames (1% of the total length of the stack) contains more than 90% of all the events in that cluster, the respective cluster is rejected. Thus, single and/or atypically extended events are rejected as false localization. Cluster centres are calculated and are used as the location of glycan targets.
For further analysis, two approaches were used. First, we calculated the NN distances across all glycan localizations, both within channels and between different channels. Histograms of the distances to the first NN were plotted and the peak of the distributions were collected in a 5 × 5 matrix (6 × 6 for datasets with DBCO). Second, we performed a classification of lectin binding sites that are frequently occurring in close spatial proximity within the same channel and across channels using a cut-off radius of 5 nm. Each observed combination of two or more lectin binding sites was assigned a different class. The distribution of classes per square micrometre was plotted and the cellular location of these classes was mapped. To handle these multidimensional data, we performed standard linear dimensionality reduction using PCA in Python.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
