A general design of caging-group-free photoactivatable fluorophores for live-cell nanoscopy

The controlled switching of fluorophores between non-fluorescent and fluorescent states is central to every super-resolution fluorescence microscopy (nanoscopy) technique, and the exploration of radically new switching mechanisms remains critical to boosting the performance of established, as well as emerging super-resolution methods. Photoactivatable dyes offer substantial improvements to many of these techniques, but often rely on photolabile protecting groups that limit their applications. Here we describe a general method to transform 3,6-diaminoxanthones into caging-group-free photoactivatable fluorophores. These photoactivatable xanthones (PaX) assemble rapidly and cleanly into highly fluorescent, photo- and chemically stable pyronine dyes upon irradiation with light. The strategy is extendable to carbon- and silicon-bridged xanthone analogues, yielding a family of photoactivatable labels spanning much of the visible spectrum. Our results demonstrate the versatility and utility of PaX dyes in fixed and live-cell labelling for conventional microscopy, as well as the coordinate-stochastic and deterministic nanoscopies STED, PALM and MINFLUX.

F luorescence nanoscopy has revolutionized our ability to visualize (living) cells by extending the limits of optical imaging to single-digit nanometre resolution, and by enabling minimally invasive observation of the internal nanoscale structures and dynamics of biological samples with molecular specificity [1][2][3][4] . Central to these techniques are chemically specific fluorescent labels and the intrinsic control between fluorescent (on) and non-fluorescent (off) states of the fluorophores. This sequential off-on transition is key to separating adjacent fluorophores at molecule-scale proximities. Photoactivatable or caged dyes-in which the off-on transition is irreversible and triggered by light-render these nanoscopy techniques very powerful, because they eliminate the need for specific imaging buffers and high intensities of UV light. Such requirements are prevalent in single-molecule-based microscopy, such as photo-activated localization microscopy (PALM) or stochastic optical reconstruction microscopy (STORM), to drive commonly used fluorophores (for example, cyanines) between non-fluorescent and fluorescent states [5][6][7] , as well as enabling high-density single-particle tracking [8][9][10] . Most recently, photoactivatable dyes have been used to reduce the fluorescence background in DNA-PAINT 11 and to increase the number of cellular structures that may be simultaneously imaged in stimulated emission depletion (STED) microscopy through channel duplexing 12 .
Rhodamine dyes have emerged as some of the most widely employed fluorophores in fluorescence microscopy and nanoscopy due to the remarkable tunability of their optical and chemical properties [13][14][15] , cell membrane permeability 16 , photostability 17 and brightness 18 . In particular, silicon rhodamines 19 are often favoured for their intrinsic redshifted emission, fluorogenic behaviour 20 and live-cell compatibility [21][22][23] . However, the reported caging strategies for rhodamines rely on 'locking' the dyes in a non-fluorescent form, either through installation of photolabile protecting groups on the nitrogen atoms (such as with nitroveratryloxycarbonyl 7,24,25 or nitroso 26 groups) or by synthetic transformation of the lactone ring into the corresponding cyclic α-diazoketones 9,27 . The former strategy restricts the attainable substitution patterns, reduces water solubility and yields stoichiometric amounts of potentially toxic by-products upon photoactivation. The latter strategy, meanwhile, suffers from varying uncaging efficiencies and the concomitant formation of non-fluorescent side products, whose abundance depends on the medium and substitution pattern 27,28 .
Accordingly, caging-group-free, compact photoactivatable and biocompatible fluorophores are highly desirable in fluorescence microscopy and nanoscopy applications, enabling lower-molecular-weight labels, provided that the photoactivation is rapid, complete and free of by-products. Recently, the photoactivation of a Si-pyronine analogue was demonstrated, where the fluorophore was initially masked with an exocyclic double bond at the 9-position of the Si-xanthene scaffold 29 . Upon UV irradiation in aqueous solution, protonation of the exocyclic double bond yielded the fluorescent 9-alkyl-Si-pyronine. The resulting cationic fluorophore, however, was susceptible to formation of non-fluorescent nucleophilic addition products with thiols and water, limiting its applicability.
Inspired by the long-established radical photochemistry of benzophenone and other diarylketones, we have now designed, and report herein, a class of functionalized xanthones, which, upon oneor two-photon excitation, convert efficiently and cleanly into the corresponding dihydropyran-fused pyronine dyes. These photoactivatable xanthone (PaX) dyes can be prepared from readily available starting materials via a straightforward and efficient three-step synthetic route, also compatible with carbon-and silicon-bridged The controlled switching of fluorophores between non-fluorescent and fluorescent states is central to every super-resolution fluorescence microscopy (nanoscopy) technique, and the exploration of radically new switching mechanisms remains critical to boosting the performance of established, as well as emerging super-resolution methods. Photoactivatable dyes offer substantial improvements to many of these techniques, but often rely on photolabile protecting groups that limit their applications. Here we describe a general method to transform 3,6-diaminoxanthones into caging-group-free photoactivatable fluorophores. These photoactivatable xanthones (PaX) assemble rapidly and cleanly into highly fluorescent, photo-and chemically stable pyronine dyes upon irradiation with light. The strategy is extendable to carbon-and silicon-bridged xanthone analogues, yielding a family of photoactivatable labels spanning much of the visible spectrum. Our results demonstrate the versatility and utility of PaX dyes in fixed and live-cell labelling for conventional microscopy, as well as the coordinate-stochastic and deterministic nanoscopies STED, PALM and MINFLUX.

NATUrE CHEMISTry
analogues, to yield a family of fluorophores spanning much of the visible spectrum. In particular, PaX-derived Si-pyronine dyes display good live-cell compatibility, resilience to nucleophiles, and an unprecedented photostability for orange-emitting (TAMRA-like) fluorophores. We highlight the utility of PaX dyes and labels in optical microscopy and nanoscopy techniques, in fixed and living cells, including STED, photo-activated localization microscopy (PALM) and minimal photon fluxes (MINFLUX).

Results and discussion
Synthetic design and proposed mechanism of photoactivation. In our search for minimalistic photoactivatable fluorophores, we reasoned that the concept of employing photochemical reactions to assemble or 'lock' fluorophores, rather than 'unlocking' photocleavable caging elements, would provide an improved alternative to caged rhodamine dyes (Fig. 1a)-a strategy similar to photochromic diarylethenes 30 . Diarylketones are known photoinitiators of radical reactions 31 due to their high inherent rate of intersystem crossing (via spin-orbit coupling) and their triplet states with diradical character 32,33 . We hypothesized that their photochemistry would be extendable to 3,6-diaminoxanthones, which are utilized as precursors in the synthesis of rhodamines [34][35][36] . With the introduction of a suitable intramolecular radical trap onto the xanthone scaffold, a juxtaposition of a radical source (diaryl ketone) and a radical trap (styrene) could be exploited to photoassemble 9-alkoxypyronine fluorophores through a light-triggered cascade (Fig. 1b) 37 .
To investigate the effects of substitution of the radical acceptor, we first synthesized a series of photoactivatable Si-xanthones (1-7; Fig. 1c). The target compounds were prepared by an Ir-catalysed, chelation-assisted, ortho-selective C-H borylation of the diaryl ketone (A) 38 (13,14) and actin labelling (15). b, STeD (left) and PALM (right) images of microtubules in COS-7 cells labelled by indirect immunofluorescence with a secondary antibody bearing 13. Preactivation to 13-CF for STeD imaging was achieved with widefield illumination (AHF analysentechnik AG, 4,6-diamidino-2-phenylindole filter set F46-816). c, Actin structures of the periodic membrane cytoskeleton in the axon of fixed primary hippocampal neuron cultures labelled with 15 and mounted in Mowiol. Preactivation to 15-CF for STeD imaging was achieved with widefield illumination (AHF, enhanced green fluorescent protein (eGFP) filter set F46-002) followed by a 518-nm laser. Image data were smoothed with a 1-pixel low-pass Gaussian filter.

NATUrE CHEMISTry
CuBr 2 -pyridine system 39 in the presence of KF (details are provided in the Supplementary Information). A series of alkene substituents were then installed using standard Suzuki-Miyaura cross-coupling reaction conditions. Compounds 1-6 showed a strong absorption band (ε ≈ 10 4 M −1 cm −1 ) at ~400 nm, characteristic of Michler's ketone and its analogues (Supplementary Table 1 presents the photophysical characterization). Upon irradiation in protic media (for example, phosphate buffer, 100 mM, pH 7), compounds 1-6 underwent rapid and complete conversion to give highly fluorescent 'closed-form' (CF) products with TAMRA-like spectral properties (1-CF to 6-CF; Fig. 1d and Supplementary Fig. 1). Liquid chromatography mass spectrometry (LC-MS) analysis of the reaction mixtures revealed no by-products for most samples. The measured quantum yields of photoactivation (Φ PA ) ranged from 1 × 10 −2 to 6 × 10 −2 (Supplementary Table 1). The rate of photoactivation was slowest for vinyl-substituted compound 1, increased with additional substitution of the alkene, and was highest for compound 4, possibly due to a favourable orientation of the alkene induced by the α-methyl substituent (Fig. 1e). To confirm the formation of predicted 9-alkoxypyronine product 1-CF, a solution of compound 1 in methanol was irradiated with a 405-nm light-emitting diode (LED) in a batch photoreactor (see Supplementary Information for details), and the resulting product was isolated and fully characterized by NMR and high-resolution mass spectrometry (HR-MS) analysis ( Supplementary Fig. 2), confirming the expected dihydropyran ring fusion. A solvent-dependent protonation step was confirmed by conducting photolysis in methanol-d 4 (resulting in deuterium incorporation at the benzylic position) and by the absence of efficient photoactivation in aprotic solvents such as 1,4-dioxane. Deoxygenating the solvent increased the rate of photoactivation, confirming the role of the xanthone triplet state. Photolysis of 1 (4.8 µM) in the presence of millimolar concentrations of the radical trap 4-hydroxy-TEMPO resulted in the formation of a PaX-TEMPO adduct ( Supplementary Fig. 3); however, the radical clock probe 7

NATUrE CHEMISTry
showed no evidence of cyclopropane ring-opening upon photoactivation ( Supplementary Fig. 4).
Screening the photoactivation properties of 11 over a range of biologically relevant pH values ( Fig. 1g and Supplementary Fig. 7a) revealed a six-fold decrease in the photoactivation rate in acidic media (pH 4.3) as compared to neutral, and only small rate changes at basic pH values (up to 9.0). The low pH-dependence of the activation rate is similar to previous observations on the protonation of the benzophenone triplet excited state 40,41 , supporting the assumed involvement of this diradical in the activation mechanism. At high pH values, slow hydrolysis of 11-CF was observed ( Supplementary  Fig. 7b); however, there was no difference in the absorption and emission spectra of 11-CF and no change in product composition was detected by LC-MS up to pH 8.5 ( Supplementary Fig. 7c), indicating little observable pH-sensitivity for this dye across the biologically relevant pH range. Furthermore, photoactivation of 11 proceeded cleanly in buffered solutions (pH 7) containing 2 mM mercaptoethylamine or glutathione ( Supplementary Fig. 8), anticipating a lack of unwanted radical or electrophilic reactivity towards biomolecules, and a potential orthogonality with the single-molecule localization microscopy (SMLM) blinking buffers used for cyanine dyes 5 . Finally, we assessed the photostability of 11-CF, benchmarking it against a series of commercially available

NATUrE CHEMISTry
dyes with similar spectral properties (Fig. 1h), and found that 11-CF outperformed all of the tested fluorophores (for details, see Supplementary Information and Supplementary Figs. 9 and 10).

Caging-group-free photoactivatable labels for nanoscopy.
Encouraged by the versatility of the PaX mechanism, we proceeded to construct targeted labels for fluorescence microscopy and nanoscopy. For indirect immunolabelling (with secondary antibodies or nanobodies), an amino-reactive N-hydroxysuccinimide (NHS) ester (13) and a thiol-reactive maleimide (14) derivative of PaX 560 were prepared (Fig. 2a) (15) of PaX 560 was assembled (Fig. 2a). Thanks to their remarkable photo-fatigue resistance, we reasoned that PaX dyes would be strong candidates for STED imaging. We tested their performance by indirect immunofluorescence labelling of microtubules in fixed COS-7 cells. The fluorescent form of the dye was generated in situ (405-nm photoactivation) before STED imaging with 561-nm and 660-nm light for excitation and STED, respectively. Super-resolved images of microtubules were successfully acquired for antibody conjugates of PaX 560 (13; Fig. 2b), as well as of PaX 525 and PaX+ 560 (17,18; Supplementary Fig. 11b), demonstrating their compatibility with STED nanoscopy. The specificity of PaX 560 -phalloidin (15) for actin was validated in fixed neuron cultures in which the periodic membrane cytoskeleton structure of the axon was visualized by STED (Fig. 2c).
We next tested the performance of our photoactivatable labels in SMLM 42,43 . With this aim, PALM imaging was carried out on indirectly immunolabelled microtubules, and super-resolved images could be obtained for antibody conjugates bearing 13 (Fig. 2b) and  16-18 (Supplementary Fig. 11c). Thanks to the efficient photoactivation mechanism, very low powers (<100 µW) of activation light were required. Importantly, all samples were imaged in phosphate buffered saline (PBS) or in Mowiol, without the need for special blinking buffers or photostabilizing agents.
To further benchmark the utility of PaX labels for PALM imaging, indirect immunofluorescent labelling of nuclear pore complexes (NPCs) was conducted with a primary anti-NUP-98 antibody and secondary anti-rabbit nanobodies bearing 14 ( Supplementary  Fig. 12a,b). The PALM images of NPCs (Fig. 2d) were comparable in quality to those acquired through more demanding methods (for example, qPAINT 44 ). Alternatively, the large-sized (~150 kDa) primary antibodies could be avoided to improve labelling precision 45 in cell lines expressing an mEGFP 46 (~27 kDa) fusion to NUP107 when combined with anti-GFP nanobodies labelled with 14 (Fig. 2e).

Targeted labels for live-cell imaging.
To evaluate the compatibility of the PaX photoactivation mechanism with live imaging, we first prepared PaX 560 constructs (Fig. 3a) containing mitochondria-targeting triphenylphosphonium (19) and lysosome-targeting pepstatin A (20) moieties, as these selected organelles represent the extreme pH values found within the cell (pH 7.8 for the mitochondrial matrix and pH 4.5 in the lysosomal lumen). COS-7 cells were co-incubated with 19 and MitoTracker Deep Red and imaged with confocal microscopy before and after photoactivation with 355-nm light (Fig. 3b and Supplementary Video 1). The resulting fluorescence of 19-CF co-localized strongly with the MitoTracker signal (Pearson correlation coefficient r = 0.94). Similarly, COS-7 cells concurrently labelled with the pepstatin A conjugate 20 and the lysosome-targeting fluorophore SiR-lysosome 20 demonstrated colocalization after photoactivation with r = 0.84. These results confirmed that the photoactivation mechanism is compatible with live-cell imaging in both high-and low-pH cellular environments.
Self-labelling protein tags, such as HaloTag and SNAP-tag, are well-established tools for targeting synthetic fluorophores to specific proteins in live-cell imaging 47 . Seeking to exploit this targeting strategy, we prepared the HaloTag-specific chloroalkane derivative (21) and the SNAP-tag specific O 6 -benzylguanine derivative (22) of PaX 560 (Fig. 3a)

NATUrE CHEMISTry
Upon covalent linking of PaX 560 -Halo (21) with the HaloTag protein, we observed a 7.8-fold increase in the photoactivation rate of the dye (Supplementary Fig. 14a-d). Complete reaction of 21 with HaloTag with only a slight excess (~1.1 equiv.) of the protein was confirmed by mass spectroscopy (Supplementary Fig. 14b,d).
No major fluorescence intensity changes were observed for 21-CF covalently bound to HaloTag in comparison to free 21-CF in buffered solution (Supplementary Fig. 14e,f). However, a similar labelling efficiency and a greater fluorogenic response were observed upon binding of PaX 560 -SNAP (22) to SNAP-tag, with an 11-fold increase in photoactivation rate ( Supplementary Fig. 13a-d) Fig. 15e,f).
We then assessed the feasibility of two-photon activation of PaX 560 -Halo (21) with 810-nm near-infrared (NIR) light, as shifting the excitation wavelength from UV to the NIR range reduces phototoxicity and increases imaging depth in tissues. U2OS cells stably expressing a vimentin-HaloTag fusion construct 48 were labelled with compound 21 and imaged with a confocal microscope equipped with a subpicosecond pulsed laser (Fig. 3c). The activation rate constant was determined for selected areas of the same sample by mono-exponential fitting of the activation rates measured with variable powers of a UV laser for one-photon activation (355 nm) or a subpicosecond pulsed laser for two-photon activation in the NIR (810 nm). Two-photon activation was confirmed by the nearly quadratic (1.84) dependence on the power of the excitation light (Fig. 3d). A pre-activated region of the same sample was further imaged using STED (at 660 nm) to resolve vimentin filaments with subdiffraction resolution, confirming that live-cell STED was readily possible with compound 21 (Fig. 3e). Live-cell STED time-lapse imaging further highlighted the cell dynamics after photoactivation (Supplementary Video 2).
Photoactivatable fluorophores can also be utilized together with regular 'always-active' fluorescent dyes having similar spectral properties for colour duplexing within a single excitation/detection channel, effectively doubling the number of available imaging channels in a confocal or STED system 12 , provided that bleaching of the 'always-active' dye does not result in cell damage. To demonstrate this possibility with PaX labels, U2OS cells stably expressing a vimentin-HaloTag fusion protein were concurrently labelled with PaX 560 -Halo (21) and an Abberior LIVE 560 tubulin (AL-560) probe, then imaged by confocal microscopy using a single detection channel (Fig. 4a-e). First, the AL-560-labelled tubulin filaments were visualized (Fig. 4a), followed by AL-560-photobleaching with high-intensity 560-nm excitation light (Fig. 4b). Compound 21 was then, in turn, photo-activated with a 405-nm laser to reveal the 21-CF-labelled vimentin structure (Fig. 4c).
We explored the utility of the self-labelling protein tag substrates 21 and 22 for live-cell SMLM. U2OS cells stably expressing a SNAP-tag fusion with NUP107 49 were labelled with PaX 560 -SNAP (22) and imaged with PALM (Fig. 5a). The reconstructed image shows largely complete circumferential labelling, which is remarkable given the one-to-one dye-to-protein ratio, highlighting the efficient labelling and efficient detection of activated PaX. Similarly, U2OS cells stably expressing HaloTag fusion proteins with NUP96 45 (another NPC protein) were labelled with PaX 560 -Halo (21). The reconstructed image (Fig. 5b) resolved the structural elements of the NPCs with even greater efficiency. Fixation of live-labelled samples (with HaloTag and SNAP-tag fusion proteins) also allowed PALM imaging with similar contrast (Supplementary Fig. 16). Thus, the established fixation and permeabilization treatments used to preserve NUP structures 45 for super-resolution imaging do not affect the performance of PaX labels.

Multiplexing of PaX labels by selective photoactivation.
Given the difference in photoactivation rates for the PaX dyes, we surmised that two complementary labels could be used for multiplexing purposes by sequentially applying a lower and a higher dose of activation light, to first convert one fluorophore (for example, PaX 560 ) while preserving the more difficult to activate (for example, PaX 480 ) until higher light doses are applied. We tested this first by confocal imaging (Supplementary Fig. 17a-e) and next by two-colour single-detector PALM imaging (Supplementary Fig. 17f) in fixed cells. We further demonstrated sequential activation in live cells ( Supplementary  Fig 18) by confocal imaging, using the organelle-(PaX 560 -Mito 19 or PaX 560 -Lyso 20) and HaloTag-specific (PaX 480 -Halo, 23) labels.
Utilizing PaX labels in MINFLUX nanoscopy. Finally, we tested the PaX labels in MINFLUX nanoscopy 1,4 , a recent technique that localizes individual fluorophores using an excitation beam with an intensity minimum (zero). Fixed HeLa-Kyoto cells expressing mEGFP fused to NUP107 were labelled with anti-GFP nanobodies bearing 14 and imaged by MINFLUX (Fig. 6a), yielding images of largely complete NPCs (Fig. 6b). On average, molecules were localized 106 times, utilizing 116 photons in the final MINFLUX iteration, and accounting for a mean label precision of 3.7 nm (s.d.).

Conclusion
We have introduced a general design strategy for caging-group-free, bright-and live-cell-compatible photoactivatable dyes, suitable for a wide range of optical microscopy and nanoscopy techniques, including PALM, STED and MINFLUX. The unique structural feature of these PaX dyes is the combination of a light-responsive 3,6-diaminoxanthone core functionalized with an intramolecular alkene radical trap, to give a highly compact and intrinsically uncharged, intact cell-membrane-permeable label. Under one-or two-photon activation, these compounds rapidly assemble into highly photostable fluorescent pyronine dyes. By changing the substitution pattern of PaX dyes, the photoactivation kinetics as well as the spectral properties can be tuned, allowing for both multiplexed pseudocolour as well as conventional multicolour imaging. The utility and versatility of PaX dyes is illustrated with a diverse range of target-specific probes and labelling strategies, for fixed-and live-cell super-resolution fluorescence microscopy experiments. We expect that our methodology will further stimulate the development of photoactivatable probes and sensors for biological imaging and material science. Further improvements to PaX fluorophores Articles NATUrE CHEMISTry will benefit applications in MINFLUX imaging and the recently proposed MINSTED nanoscopy 50 .

Online content
Any methods, additional references, Nature Research reporting summaries, source data, extended data, supplementary information, acknowledgements, peer review information; details of author contributions and competing interests; and statements of data and code availability are available at https://doi.org/10.1038/ s41557-022-00995-0.

Methods
Detailed procedures for the synthesis of all compounds and their characterizations, as well as methods sample preparation, live and fixed-cell labelling for microscopy and nanoscopy, are provided in the Supplementary Information. Image acquisition conditions for confocal and STED (Supplementary Table 3) and PALM  (Supplementary Table 4), as well as detailed procedures for image processing and rendering, are provided in the Supplementary Information.

Statistics and reproducibility.
All biochemical or spectroscopic data were obtained in triplicate with similar results. All staining/labelling of cells was performed in triplicate. Cells for microscopy were selected at random during the imaging session; sufficient microscopy images were collected, from experience, to ensure their representation of the sample.
Neuronal culture preparation and labelling. Cultures of dissociated rat hippocampal primary neurons were prepared from postnatal P0-P1 Wistar rats of either sex and cultured on glass coverslips coated with 100 µg ml −1 poly-ornithine (Merck KGaA) and 1 µg ml −1 laminin (BD Biosciences). Procedures were performed in accordance with the Animal Welfare Act of the Federal Republic of Germany (Tierschutzgesetz der Bundesrepublik Deutschland, TierSchG) and the Animal Welfare Laboratory Animal Regulations (Tierschutzversuchsverordnung). According to the TierSchG and the Tierschutzversuchsverordnung, no ethical approval from the ethics committee is required for the procedure of euthanizing rodents for subsequent extraction of tissues. The procedure for euthanizing P0-P1 rats performed in this study was supervised by animal welfare officers of the Max Planck Last updated by author(s): Apr 18, 2022 Reporting Summary Nature Portfolio wishes to improve the reproducibility of the work that we publish. This form provides structure for consistency and transparency in reporting. For further information on Nature Portfolio policies, see our Editorial Policies and the Editorial Policy Checklist.

Statistics
For all statistical analyses, confirm that the following items are present in the figure legend, table legend, main text, or Methods section.

n/a Confirmed
The exact sample size (n) for each experimental group/condition, given as a discrete number and unit of measurement A statement on whether measurements were taken from distinct samples or whether the same sample was measured repeatedly The statistical test(s) used AND whether they are one-or two-sided Only common tests should be described solely by name; describe more complex techniques in the Methods section.
A description of all covariates tested A description of any assumptions or corrections, such as tests of normality and adjustment for multiple comparisons A full description of the statistical parameters including central tendency (e.g. means) or other basic estimates (e.g. regression coefficient) AND variation (e.g. standard deviation) or associated estimates of uncertainty (e.g. confidence intervals) For null hypothesis testing, the test statistic (e.g. F, t, r) with confidence intervals, effect sizes, degrees of freedom and P value noted For manuscripts utilizing custom algorithms or software that are central to the research but not yet described in published literature, software must be made available to editors and reviewers. We strongly encourage code deposition in a community repository (e.g. GitHub). See the Nature Portfolio guidelines for submitting code & software for further information.

Data
Policy information about availability of data All manuscripts must include a data availability statement. This statement should provide the following information, where applicable: -Accession codes, unique identifiers, or web links for publicly available datasets -A description of any restrictions on data availability -For clinical datasets or third party data, please ensure that the statement adheres to our policy The data supporting the findings of this study are available within the paper and its Supplementary Information and are available from the corresponding author upon reasonable request.