General remarks

All reagents were purchased from commercial sources and used as received. Anhydrous solvents were procured from Acros Organics and used as received. Nuclear magnetic resonance spectra were acquired on a Bruker 400 or 600 instrument using TopSpin 4.2.0 and analysed with MestreNova 14.2. The 1H NMR chemical shifts are reported in parts per million relative to SiMe4 (δ = 0) and were referenced internally with respect to residual protons in the solvent (δ = 7.26 for CDCl3, δ = 1.94 for CD3CN, δ = 3.31 for CD3OD, and δ = 2.50 for (CD3)2SO). Coupling constants are reported in hertz. The 13C NMR chemical shifts are reported in parts per million relative to SiMe4 (δ = 0) and were referenced internally with respect to solvent signal (δ = 77.16 for CDCl3, δ = 1.32 for CD3CN, δ = 49.00 for CD3OD and δ = 39.52 for (CD3)2SO). High-resolution mass spectra were acquired on a timsTOF Pro TIMS-QTOF LC/MS spectrometer (Bruker Daltonics GmbH) by using electrospray ionization and analysed using Xcalibur v.4.2 software. IUPAC names of all compounds are provided and were determined using CS ChemBioDrawUltra v.16.0.

Computational modelling

All calculations were performed using Gaussian 09 at the B3LYP/DGTZVP level of theory as well as at the M06-2X/DGTZVP level. An implicit solvation model (IEFPCM) was used to simulate the effect of an aqueous environment. All stationary states were characterized by harmonic analysis at the same level of theory. All minima displayed zero imaginary frequencies and all transition states gave one imaginary frequency along the C–O bond elongation coordinate. Energies were corrected by zero-point energy.

Optical spectroscopy

Stock solutions were prepared in DMSO (spectrophotometric grade >99.9%) at concentrations of 5 mM, 1 mM and 50 μM and stored at –20 °C. Spectroscopic measurements were conducted in PBS. Ultraviolet–visible spectra were acquired using a Multiskan SkyHigh Microplate Spectrophotometer (ThermoFisher Scientific) and quartz cuvettes from ThorLabs (10 mm path length).

Buffered aqueous solutions in the pH range of 2 to 8 were prepared by mixing citric acid (0.1 M) and sodium dihydrogen phosphate NaH2PO4 (0.2 M) in ultrapure water. Buffered aqueous solutions in the pH range of 9 to 11 were prepared by mixing sodium bicarbonate (0.1 M) and sodium carbonate (0.1 M) in ultrapure water. Buffered aqueous solutions in the pH range of 12 to 13 were prepared by mixing potassium chloride KCl (0.2 M) and sodium hydroxide (0.2 M); 5 μM solutions of the dyes in the buffered aqueous solutions were prepared and the absorbance spectra were recorded after 1.5 h in triplicates using 96-well plates (Corning) on the MultiSkan SkyHigh microplate reader. The obtained ultraviolet–visible spectra were background corrected and the absorbance maxima of each pH value were normalized. Normalized absorbance values were plotted against the pH value and fitted (non-linear fit, sigmoidal, 4PL) using Prism 9.

Solutions of different dielectric constant were prepared by making water-dioxane mixtures containing 10% (ε = 72.02), 20% (ε = 63.50), 30% (ε = 54.81), 40% (ε = 45.96), 50% (ε = 36.89), 60% (ε = 28.09), 70% (ε = 19.73), 80% (ε = 12.19) or 90% (ε = 6.23) dioxane (by volume) in water52.

We prepared 10 μM solutions of the dyes in the dioxane–water mixtures and the absorbance spectra were recorded in triplicates using 96-well plates (Corning) on a MultiSkan SkyHigh microplate reader. The obtained ultraviolet–visible spectra were background-corrected and the absorbance maxima of each dioxane-water mixture were normalized. Normalized absorbance values were plotted against the dielectric constant and fitted (non-linear fit, sigmoidal, 4PL) using Prism 9.

Extinction coefficients were obtained by measuring the absorbance spectra at various concentrations between 1 µM and 15 µM. The absorbance maxima were plotted against the corresponding concentration and fitted (simple linear regression) using Prism 9.

Fluorescence spectra were acquired using an FS5 Spectrofluorometer operated with Fluoracle software (Edinburgh Instruments) equipped with an SC-25 cuvette holder or SC-40 plate reader. Absolute fluorescence quantum yields were determined using 1 µM solutions employing an integrating sphere (SC-30, Edinburgh instruments). All spectroscopic measurements were performed in triplicate and at room temperature.

Photobleaching experiments were performed using a FS5 Spectrofluorometer equipped with an SC-25 cuvette holder. Protein adducts were measured at a 50 nM concentration and the emission slit width was adjusted individually for each dye to achieve a power of 1.2 mW (547 nm, 12 nm; 646 nm and 654 nm, 22.5 nm; 751 nm, 29.9 nm). The emission slit width was set to 0.75 nm and the fluorescence was measured every second for 10 min while keeping the shutter always open. The power was measured at the cuvette holder using a PMD100D compact power and energy meter console equipped with a S120VC standard photodiode power sensor (UV-Extended Si, 200–1,100 nm, 50 mW, Thorlabs GmbH).

Cloning

All plasmids were cloned by Gibson assembly. DNA encoding SNAPf or H2B-SNAPf was amplified from pSNAPf-H2B control plasmid (Addgene no. 101124) to generate an insert. To generate a backbone containing an organelle-specific targeting group and a fluorescent protein, the plasmids pmTurquoise2-ER (Addgene no. 36204) and pmTurquoise2-Golgi (Addgene no. 36205) were used. Primers for amplification (minimum 15 overlapping base pairs) were generated using SnapGene and modified manually to minimize secondary structures, self-dimers and repeated motifs. The designed primers were supplied by Microsynth AG (Switzerland) and are reported in Supplementary Tables 2 and 3. Vector and insert fragments were amplified by PCR using a Phusion High-Fidelity PCR Master Mix with HF buffer from New England Biolabs (NEB). The fragments were analysed by agarose gel electrophoresis and template DNA was digested using DpnI. PCR fragments were purified using a QIAquick PCR purification kit (Qiagen) according to the manufacturer’s instructions. The Gibson assembly reaction was performed using the Gibson Assembly Master Mix (NEB) according to the manufacturer’s protocol. DH5α competent Escherichia coli cells (NEB, C2987I) were transformed with the assembly product by heat shock following the manufacturer’s instructions, streaked onto lysogeny broth (LB) agar plates containing kanamycin (50 μg ml–1), and incubated at 37 °C for 24 h. Single colonies were selected and grown in LB liquid medium containing 50 μg ml–1 kanamycin at 37 °C for 16 h. Plasmid DNA was isolated using the QIAprep spin miniprep kit (Qiagen) according to the manufacturer’s instructions. The correct sequence of the gene of interest was verified by the Sanger sequencing service of Microsynth using the standard sequencing primers CMV-F, SV40pA-R, and EGFP-C. All plasmids produced in this paper are available on Addgene (no. 197494–98).

Expression and purification of fSNAP protein

pET24b-6His-fSNAP (Addgene no. 106999) was transformed into E. coli strain Rosetta2(DE3) (Merck, 71397) and streaked on an agar plate with kanamycin and chloramphenicol resistance. A single colony was picked, inoculated in 80 ml LB cultures containing 50 μg ml–1 kanamycin and 25 μg ml–1 chloramphenicol, and grown at 37 °C for 24 h. We innoculated 40 ml of the starter culture into 4 l of medium and shaken at 37 °C to an optical density at 600 nm (OD600) of 0.8. Expression was induced by the addition of 0.5 mM isopropyl β-d-thiogalactopyranoside, and cells were grown at 18 °C for 16 h. Cells were harvested by centrifugation, resuspended in HEPES buffer (300 mM NaCl, 20 mM HEPES, pH = 7.5), and supplemented with turbonuclease (20 μl) and a protease inhibitor cocktail tablet (Roche). Cells were lysed by sonication (70% power, 10 s pulse/10 s pulse off for 2 min, 30 s) and the lysate was cleared by centrifugation. The protein was purified by Ni-His-affinity column chromatography and the fractions were analysed by SDS–PAGE. The fractions containing the protein were concentrated using Amicon Ultra-4 and Ultra-15 centrifugal filter units (MWCO 10 kDa) and further purified by size-exclusion chromatography. The correct size and purity of the protein were verified by SDS–PAGE analysis. Purified SNAP protein was stored in aliquots in a buffer containing 150 mM NaCl, 20 mM HEPES, pH 7.4, 1 mM DTT at −80 °C.

Protein-binding turn-on assays

Prior to biological assays with purified SNAP-tag protein, the buffer was exchanged to PBS using Zebra Spin Desalting Columns (Thermo Scientific) according to the manufacturer’s protocol. SNAP-dyes (1 mM in DMSO) were either added to PBS alone or to a 5 μM SNAP-protein solution in PBS. The final concentration of SNAP-dyes was 2.5 μM and the resulting mixtures were incubated at 37 °C for 1.5 h.

The turn-on assay with actin was performed according to a published procedure45. Actin ligand 17-actin (1 mM in DMSO) was added to either supplemented actin buffer alone or in the presence of 0.4 mg ml−1 G-actin (catalogue no. AKL99, Cytoskeleton). The final ligand concentration was 2.5 μM. The supplemented actin buffer contained 5 mM Tris-HCl pH 8.0, 0.2 mM CaCl2 (from General Actin Buffer, catalogue no. BSA01, Cytoskeleton), 50 mM KCl, 2 mM MgCl2, 5 mM guanidine carbonate and 1.2 mM ATP (from Actin Polymerization Buffer, catalogue no. BSA02, and ATP, catalogue no. BSA04, Cytoskeleton). Samples were incubated at 37 °C for 2 h.

The DNA-binding assay was performed according to a published procedure47. The hairpin-forming 28-bp DNA oligonucleotide was purchased from Microsynth (5’-CGCGAATTCGCGTTTTCGCGAATTCGCG-3’) and dissolved in Tris-buffered saline (TBS, 50 mM Tris-HCl, 150 mM NaCl, pH 7.4) at 1 mM concentration. The folding of the DNA into the secondary hairpin structure was achieved by heating the DNA at 75 °C for 2 min followed by slowly cooling to 25 °C. Nucleus ligand 18-DNA (1 mM in DMSO) was incubated with either TBS (50 mM Tris-HCl, 150 mM NaCl, pH 7.4) alone or in the presence of 50 μM hairpin DNA at room temperature for 1 h. The final ligand concentration was 1 μM.

Absorbance spectra were recorded in 96-well plates (Corning) on the Multiskan SkyHigh microplate reader and fluorescence measurements were performed on the FS5 Spectrofluorometer (Edinburgh Instruments) equipped with a SC-40 plate holder. All spectroscopic measurements were carried out in triplicates and at room temperature.

Cell culture and fluorescence imaging

HeLa cells (CLS, 300194CP5) were grown in Dulbecco’s Modified Eagle Medium supplemented with foetal bovine serum (10%) and penicillin-streptomycin (1%) at 37 °C in a 95% humidity atmosphere under a 5% CO2 environment. The cells were grown to 90% confluency before seeding at a density of 15–20 000 cells per millilitre onto Ibidi μ-Slide 8-well glass-bottom plates 48 h before the imaging experiment. Cells were transfected with the plasmids H2B-SNAPf-mTurquoise2 (Addgene no. 197494), LifeAct-SNAPf-mTurquoise2 (no. 197497), TOM20-SNAPf-mTurquoise2 (no. 197496), SNAPf-mTurquoise2-KDEL (no. 197498) or SNAPf-mTurquoise2 (no. 197495) using jetPrime according to the manufacturer’s protocol 24 h before imaging. Cells were incubated with the respective probes in FluoroBrite Dulbecco’s Modified Eagle Medium for 1.5 h and imaged directly.

Confocal microscopy

Confocal imaging was performed with a Nikon W1 spinning disk microscope operated with NIS Elements AR software equipped with a CMOS camera (Photometrix). Brightfield imaging was performed with a white LED. Laser lines and filters were set up for the appropriate channel as described in Supplementary Table 5.

Images were collected using a CFI Plan Apochromat Lambda D oil immersion objective (60×, NA = 1.4). Channels were imaged sequentially. The microscope was operated using NIS Elements. Imaging experiments were performed at 37 °C in a 5% CO2 environment. Images were analysed by Fiji/ImageJ.

Determination of the absolute configuration of (+)−6 and (–)−6

Circular dichroism spectra were measured with a Chirascan V100 CD spectrometer (Applied Photophysics) and operated using the Chirascan v.4.2 software. For circular dichroism measurements, 60 μM solutions of (+)-6 and (–)-6 in PBS were used. Optical rotation measurements were performed with a Jasco P-2000 polarimeter and operated using the Spectra Manager CFR software.

For TD-DFT calculations, we first performed a systematic conformational search at the B3LYP/DGTZVP level of theory with an implicit solvation model (IEFPCM) by varying all rotatable bonds in 60° steps. We then applied the Boltzmann distribution to the set of low-energy minima obtained by using the free energy differences, and considered the structures above the 0.1% population threshold for the TD-DFT calculation. Time-dependent density functional theory calculations were performed at the CAM-B3LYP/DGTZVP level of theory with IEFPCM as the solvation model.

Fluorescence lifetime imaging

Fluorescence lifetime imaging was performed on a Leica SP8 inverse FALCON confocal laser scanning microscope and operated using the Leica LAS X Navigator. Images were collected using an HC PL APO corr CS2 oil objective (63×, NA = 2.4) and the RHOD fluorescence filter set (excitation = 546/10, emission = 585/40). Phasor analysis of the FLIM data was performed with the Leica LAS X Phasor License.

Single-molecule localization microscopy

HeLa cells (CLS, 300194CP5) were transfected with β-tubulin-SNAP plasmid using jetPrime according to the manufacturer’s protocol 48 h before imaging. Cells were incubated with probe 11 (100 nM) for 14 h, then detached and transferred to a new Ibidi slide. Cells were left for 10 h to adhere to the glass surface, and the medium was exchanged to FluoroBrite prior to imaging. Single-molecule imaging was performed on a Ti2 eclipse inverted microscope (Nikon) equipped with a water-cooled iXon 888 Ultra EMCCD camera (Andor) at 37 °C in a 5% CO2 environment. Acquisitions were performed at 638 nm (90 mW, 30 ms) and a single band 708/75 emission filter was used. 2000 frames were acquired. Acquisitions were collected using a NIKON 100x TIRF Apo Plan SR oil objective (NA = 1.49). All laser and camera shutters were controlled by a NIDAQ oscilloscope (National Instruments) controller unit. The single-molecule signal was fitted with 2D Gaussian point spread functions using ThunderSTORM. The localization threshold was set to 1.1*std(Wave.F1).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.