Critical re-evaluation of experimental settings with fluorescent probes for bioorthogonal chemistry with palladium antimony inactivation-resistant catalyst Authors: Jana Kurcova, Marie Kudlickova Peskova, Lukas Jeremias, Sofiia Orlova, Jan Bilek, Paulina Takacsova, Zbynek Heger, Vladimir Pekarik Published: 2026 ABSTRACT AbstractBioorthogonal reactions, both in vitro and in vivo, show great promise in fluorescent proteins labelling and for the development of bioorthogonal anticancer therapies. In this work, we focus on palladium-catalysed deprotection chemistry. Despite its potential, many questions remain unanswered. The most critical issues include the long-term stability of the palladium catalysts in biological milieu, the lack of efficient methods of reaction monitoring etc. Herein, we report the design of palladium catalyst containing phosphorus, arsenic, and antimony ligands. The catalytic activity has been evaluated for up to 7 days in vitro in the presence of bovine serum and for up to 9 h in living cells. Catalyst´s performance was assessed in the presence of cysteine and glutathione, as well as the abiotic thiophenol. The catalytic activity was measured using 14 different fluorescent probes, likely the largest probe library used to date. Furthermore, we have evaluated the cellular retention of fluorescent probes used for the detection of catalysts activity, revealing that most probes or fluorescent products are poorly retained in cells, thereby limiting long-term monitoring of the catalytic activity in physiological conditions in vivo. The results show that triphenylantimony is unparalleled ligand with respect to catalyst longevity under bioorthogonal conditions as well as in the tolerance to thiols. Long-term experiments show that propargylethers are excellent probes in vitro while allylcarbamates are more useful for in vivo studies. Our results also indicate that each probe (or prodrug) – catalyst couple is unique, and a search for a universal catalyst might be futile.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-57111-2.Keywords: Bioorthogonal chemistry, Palladium, Triphenylantinony, Tsuji-Trost reaction, Fluorescent probeSubject terms: Biochemistry, Cancer, Chemical biology, Chemistry, Drug discovery Introduction Recent years have witnessed rapid development in bioorthogonal reactions that can be carried out in biologically relevant environments, including live cells or tissues. These reactions can be applied to protein and DNA labelling; however, the major focus has shifted toward bioorthogonal therapies, in which a caged prodrug is converted into an active compound directly within target cells that contain a requisite catalyst. This strategy is of considerable interest in oncology, as intratumoral production of active species can potentially replace systemic drug administration, thereby reducing many of the deleterious side effects associated with conventional chemotherapy. A wide range of bioorthogonal reactions employing non-metal chemistry, such as strain-promoted [3 + 2] dipolar cycloaddition, and inverse electron-demand Diels–Alder (IEDDA). reactions of strained cycloalkenes with tetrazines have been described1,2. Reactions involving tetrazine have been found to exhibit ultrafast kinetics and biocompatibility. Tetrazine is a strong fluorescence quencher. After a reaction with a strained dienophile such as trans-cyclooctene, norbornene or a vinyl tag is converted into dihydropyridazine, activating the fluorescence. The tetrazine moiety was effectively attached to many fluorochromes such as coumarins, BODIPY dyes, xanthenes, cyanines, aggregation induced emission (AIE) probes. Visible tetrazine-isolated fluorogenic probes maintain low background fluorescence due to electronic energy transfer (EnT) between the fluorophore (serving as the energy donor) and tetrazine (acting as the energy acceptor). These bioorthogonal reactions has been used for imaging of biological dynamics in live cells, such as DNA synthesis, RNA metabolism, imaging of site-specifically labelled proteins, visualisation of glycan dynamics and localisation, tracking lipid uptake and transport, monitoring intracellular processes and detecting low molecular weight analytes. The technique has been further modified into click-to-release methodology, allowing bioorthogonal release of the fluorochrome or caged pro-drug3–6. An alternative approach relies on metal catalysts to promote chemical conversions, with transition metals playing a central role. Among these noble metals such as Au7,8, Ru9, and especially, Pd have emerged as particularly prominent. Pd catalyses a broad array of chemical reactions – Heck, Kumada, Hiyama, Negishi, Stille, Buchwald-Hartwig, Sonogashira, Suzuki-Miyaura, and Tsuji-Trost. Many of these reactions proceed only in organic solvents and are therefore unsuitable for bioorthogonal chemistry, which must be carried out in water, as even small amounts of organic solvents are incompatible with the cell viability. Nevertheless, several Pd-catalysed reactions have been adapted for in vivo applications, including intramolecular Heck coupling10, Suzuki-Miyaura11–13, and Buchwald-Hartwig13. The Tsuji-Trost reaction remains the most widely used under biologically relevant conditions. Two general classes of catalysts are commonly used: (i) homogeneous catalysts, which are directly dissolved in cell culture media or encapsulated into a suitable nanocarrier, such as protein complex or biodegradable polymeric nanoparticles, and (ii) heterogeneous catalysts, which consist of Pd nanoparticles or complexes immobilised on inert carriers. Heterogeneous catalysts have mainly been used for extracellular prodrug or probe activation. The group of Unciti-Broceta has published several studies utilising Pd0 nanoparticles bound to polystyrene microspheres14–17. A similar strategy has been explored by Bradley group12. An intriguing concept involves in vivo generation of Pd catalyst in the form of Pd nanosheets produced within cancer cells-derived exosomes18, which improves cellular uptake of the catalyst. The use of homogeneous catalysts in bioorthogonal reactions is well documented; however, during an extensive literature survey, we identified several critical issues in the experimental designs, some of which we address in this work. A comprehensive summary of the previously described probes and reaction conditions is provided in Supplementary Table 1. (ST1) We begin with the assumption that any system with clinical potential must operate in purely aqueous media and exhibit long-term catalytic activity, ideally persisting for several days. Yet, the vast majority of previously described experiments assess catalyst performance for only 30–60 min, an interval that is clearly insufficient. Moreover, catalysts intended for biological use must withstand inactivation, particularly by biogenic thiols. Accurate evaluation of catalytic activity in biological systems requires careful selection of both the reporter probe and the reaction conditions. To develop an effective testing framework, we identify and optimise three key aspects of experimental design: (i) probe structure, (ii) cellular retention of both the probe and the uncaged compound, and (iii) the catalyst itself. In this study, we evaluate the catalytic activity of several catalytically-active Pd complexes in phosphate-buffered saline (PBS) and in the presence of 10% bovine serum, using up to 14 different fluorescent activable probes, and assess catalytic activity in the presence of biogenic thiols cysteine and glutathione, as well as in the presence of abiotic thiophenol. All experiments were carried out over an extended period of up to 7 days to delineate long-term catalytic activities essential for efficient bioorthogonal therapeutic intervention. These in vitro experiments are supplemented by the study of fluorescent probes activation in living cells. Finally, we highlight complications associated with long-term monitoring of catalysts activity in biological milieu, stemming from the limited ability of live cells to retain the fluorescent deprotected dyes. We believe that the presented data provide a solid foundation for advancing the promising concept of bioorthogonal chemistry toward novel chemistries for the development of cutting-edge diagnostic and therapeutic modalities. The results also highlight possible complications related to the transition of the field toward bioorthogonal therapy where many pro-drugs will most like exhibit similar problems addressed in this manuscript. Results and discussion Pd catalysts design and synthesis Triphenylphosphine (TPP) is one of the most commonly used ligands in Pd-catalysed reactions. An extensive literature survey revealed that trifurylphosphine ligand (TFP), in the form of dichlorobis(tri(2-furyl)phosphine)palladium(II) (Pd(TFP)2Cl2), is among the most potent catalysts reported for in vivo applications10. Replacement of TPP with TFP or triphenylarsine (TPAs) has been shown to dramatically enhance the performance of the Stille reaction19. PdTPAs complexes also exhibit strong catalytic activity in Suzuki cross-coupling reactions20 and, more broadly TPAs, serve as effective ligands for Pd and Pt21,22. The crystal structure of dibromobis(triphenylarsine)palladium(II) has been elucidated23. Pd and Pt complexes bearing arsine ligands have been proposed as potential anticancer therapeutics24–27. To further explore this trend, we included triphenylantimony (TPSb) as a ligand, yielding a new Pd catalyst. However, the structure of the compound is not as expected [Pd(SbPh3)2Cl2] but [Pd(Ph)(SbPh3)2Cl]. This compound has been characterised previously28–31. The exact structure of this complex may be somewhat ambiguous due to possible Sb–C bond cleavage28. All catalysts were synthesised by reacting bis(acetonitrile)dichloropalladium(II) with an excess of the ligand in acetonitrile (ACN), resulting in the formation of a precipitate that was isolated. In line with this, we attempted to prepare of triphenylbismuth (TPBi) Pd complex; however, the reaction with TPBi consistently led to the formation of black precipitate of metallic Pd, and was thus excluded from the further experiments. During experiments, we observed that Pd(TPP)2Cl2 and Pd(TPAs)2Cl2 complexes exhibit low solubility in DMF and extremely poor solubility in water. To improve aqueous solubility, essential for biological experiments, we produced a series of complexes with dipicolinic acid32–35(Fig. 1). Dipicolinic acid forms a meridional tridentate ligand with Pd and ACN, where the weakly coordinated ACN can be readily substituted with N- and P-donor ligands. In solution, these complexes exist in equilibrium between monomeric and dimeric form33. Schematic representation of designed Pd catalysts. The abbreviated names used throughout the publication are shown in bold. The mixed dipicolinate/P, As, Sb ligand catalysts exist in a dynamic equilibrium of monomer and dimer similar to that observed for Pd(dipic)ACN. Characterisation and catalytic activity of Pd catalysts We evaluated the catalytic performance of Pd complexes using several standard reactions commonly used to assess catalytic activity of Pd catalysts. The reduction of nitrophenol in the presence of sodium borohydride (BH) is widely used to demonstrate the catalytic activity of metal nanoparticles in hydrogenation reactions36 and is now considered a benchmark model reaction37. In this reaction, the bright yellow nitrophenolate ion is reduced to colourless aminophenol. Surprisingly, only Pd(dipic)TFP showed a measurable reduction in absorbance at 400 nm, alongside the Pd(OAc)2 control (Fig. 2a). Another benchmark reaction is the reduction of azo dyes38,39, previously exploited to detect protein-Pd complexes40. In the BH-mediated discoloration of methyl orange (Fig. 2b) and Orange G (Fig. 2c), Pd(dipic)TFP again exhibited the highest activity; however, in contrast to the nitrophenol assay, all other catalysts displayed varying degrees of activity. Pd is also known to promote benzofuran-2-boronic acid homocoupling41. This was the only reaction carried out in ACN instead of water, and at elevated temperature of 60 °C. All tested catalysts showed excellent activity, as evidenced by strong fluorescence signals (Fig. 2d) with the exception of Pd(dipic)TFP, which had performed exceptionally well in the previous assays. Reductive dehalogenation represents another common reaction catalysed by Pd42. In the reduction of weakly red fluorescent iodoBODIPY in the presence of BH, the reaction yields brightly green fluorescent H.BODIPY43. In this assay, both TFP-containing catalysts, PdTFP and Pd(dipic)TFP, together with Pd(OAc)2 outperformed all remaining catalysts (Fig. 2e). Collectively, these results confirm that we successfully produced active catalysts. However, it proved difficult to correlate the activity of individual catalysts with experimental conditions of these assays, indicating a unique relationship between each substrate and catalyst. Comparing the catalytic activities of Pd catalysts. (a) Catalytic reduction of nitrophenol; (b) Decomposition of methyl orange and (c) Orange G; (d) homocoupling of benzofuran-2-boronic acid in acetonitrile; (e) Reductive dehalogenation of iodoBODIPY; (f) Suzuki-Miyaura reaction between 2,1,3-benzothiadiazole (BTD) and phenylboronic acid carried out in PBS. Fluorescent probes design A variety of fluorescent probes have been developed for the detection of Pd in pharmaceutical products, wastewater, and in vivo. Several functional groups are commonly used to convert fluorescent molecules into Pd sensors or to transform pharmacologically active substances into prodrugs (depicted in Fig. S1), with amines, hydroxyls, and carboxyls being the most prominent. Most Pd-mediated deprotection strategies rely on Tsuji-Trost deallylation or depropargylation of caged molecules. Biological systems impose important constraints on probe design, primarily due to endogenous enzymatic activities that can lead to Pd-independent probe activation. High intracellular levels of lipases and esterases effectively preclude the use of probes based on esters or carbonates. Such unwanted bioactivation has been demonstrated for seminaphthorhodafluor (SNARF)44 and benzothiazole coumarin45, and similar instability has been reported for allyl and propargyl carbonates of gemcitabine, where the modified prodrug has similar toxicity comparable to the parental compound, suggesting rapid intracellular clevagel15. Despite these limitations, several allyl-carbonate probes have been used in biological systems, including naphthalimide46, SNARF47, dihydroxantheneindolium48, resorufin49, DDAO50, while Zhou et al. employed an allyl ester of BODIPY mesocarboxylate51. In all reported cases, probes activation, supposed to work in cell cultures or Zebrafish (Danio rerio) embryos, was assessed exclusively at short time points (30 min), and neither long-term stability nor spontaneous enzymatic activation was examined. Organic carbonates are also moisture-sensitive and prone to spontaneous hydrolysis, particularly under basic pH conditions. To evaluate spontaneous activation in the presence of serum, we used resorufin allylcarbonate (Suppl. Table 1, Entry15) in 10% bovine, goat and rabbit sera, monitoring fluorescence over 2.5 h (Fig. S2). Probe activation was rapid in rabbit and goat serum, reaching plateau within 10 min, whereas activation in bovine serum was markedly slower, reaching the plateau after 2 h, similar to the behaviour in 100 mM NaHCO3 solution at pH 9. Unexpectedly, the maximal fluorescence differed substantially among sera. To determine whether this reflected partial conversion, we measured the fluorescence of unmodified resorufin under the same conditions (Fig. S2c) and found that all sera significantly quenched its fluorescence. This unexpected observation demonstrates that reporter assays can be strongly biased by biogenic components. To avoid such artefacts in subsequent experiments, we analysed the emission spectra of several commonly used fluorophores, previously converted into Pd probes, in the presence of bovine serum albumin and/or in the presence of nucleic acids, two abundant biomacromolecules. The fluorescence of fluorescein (Fig. S3a), hydroxypyrene trisulphonic acid (HPTS, Fig. S3b), V-shaped hydroxyl coumarin52(Fig. S3c), and rhodol (Fig. S3d) was largely unaffected by RNA, DNA or albumin. In contrast, the fluorescence of hydroxyl Nile red (Fig. S4a), NDCM-OH53 (Fig. S4b), BINC2-Fluor54(Fig. S4d), and hydroxystyryl tricyanofurane (HS-TCF)55(Fig. S4e) was strongly enhanced by albumin. SNARF fluorescence (Fig. S4c) was reduced by RNA, DNA and albumin, while hydroxystyryl N-ethylbenzothiazolium iodide (Fig. S4f) showed pronounced fluorescence enhancement in the presence of DNA, an important consideration when using cell or tissue lysates in bioorthogonal activation assays. Together, these results highlight the significant potential for assay bias when evaluating catalytic activities of Pd-catalyst using probes derived from these fluorochromes. Due to the rapid activation of resorufin allyl carbonate by animal serum, as well as its spontaneous activation at elevated pH, we excluded allyl carbonate probes from all subsequent experiments. To expand the range of bright water-soluble probes, we prepared a Pd-reactive probe from HPTS. HPTS is easily propargylated by reacting it with propargyl bromide in DMF in the presence of K2CO3, yielding the probe HPTS-PE (Fig. S5a). The absorption maximum of the probe is blue-shifted from 455 nm to 400 nm (Fig. S5b). The emission maximum shifts from 515 nm for HPTS to 425 nm for HPTS-PE (Fig. S5c), providing a sufficiently large spectral shift suitable for distinguishing the probe from the released fluorophore. Next, we evaluated the selectivity of the probe toward various metal ions in the presence of reducing agent BH. The probe was effectively activated by Pd, Rh and Au, whereas Pt, Ru, Cu and Ag showed moderate activity. Other tested metal ions were unable to activate HPTS (Figure S5d). We further evaluated deprotection in the absence of BH. In this assay, Pd was the only metal capable of uncaging HPTS (Figure S5e). A second probe used in this study is rhodol propargyl ether (ROL-PE), synthesized from diethylamine rhodol by propargylation with propargyl bromide. This compound was previously reported as an Au3+ probe, undergoing a rearrangement that forms a new furan ring56. However, we found that Pd triggers depropargylation, enabling the use of ROL-PE as a Pd probe. Additional probes based on propargyl esters and allyl carbamates were prepared according to published procedures and will be used and discussed throughout the paper. Resistance of the Pd-catalysts to thiol inactivation Catalysts operating in biological environments are exposed to numerous compounds that can cause catalyst poisoning and inactivation. Among these, thiols are the most relevant as they are abundant in cells and play key roles in maintaining redox homeostasis. Therefore, we evaluated the resilience of selected Pd catalysts against biogenic thiols cysteine and glutathione, as well as the abiotic thiol thiophenol, using 0, 0.5, 1, and 2 equivalents of each thiol relative to the catalyst. Three probes were used in these assays: umbelliferone propargyl ether (U-PE) (Suppl. Table 1, Entry 23), which has been extensively used in our previous work57, seminaphthorhodol propargyl ether (SNARF-PE), prepared as a propargyl ether rather than the allyl carbonate used in the original publication47, and newly developed, highly water-soluble HPTS-PE. PdTFP displayed similar catalytic activity with all three probes U-PE (Fig. S6-S8), SNARF-PE (Fig. S9-S11), and HPTS-PE (Fig. S12-S14). In each case, deprotection by PdTFP was progressively inhibited as the thiol concentration increased, with complete inactivation of the catalyst observed at 2 equivalents of thiol. The remaining catalysts showed different responses to thiols depending on the probe used. For U-PE, deprotection by PdTPP was abolished by 0.5 equivalent of either cysteine or glutathione (Fig. S7), whereas thiophenol caused a gradual, concentration-dependent reduction in activity. In contrast, PdTPSb-mediated deprotection was progressively inhibited by cysteine and thiophenol, while remained rather unaffected in the presence of 2 equivalents of glutathione (Fig. S8). SNARF-PE deprotection by PdTPP was strongly suppressed by all biogenic thiols at 0.5 equivalent. Thiophenol caused a more gradual inhibition (Fig. S9). Interestingly, PdTPSb was completely resistant to both cysteine and glutathione, with only thiophenol exerting a dose-dependent inhibitory effect (Fig. S10). For HPTS-PE, PdTPP activity was strongly suppressed by all thiols at 0.5 equivalent (Fig. S13). Deprotection by PdTPSb decreased proportionally with increasing thiol concentrations (Fig. S14). Overall, these results underscore the complex interplay between catalyst, substrate, and the specific thiol present, highlighting the importance of probe- and thiol-dependent effects when assessing catalyst robustness. Long-term catalytic activity in the presence of bovine serum in vitro Reaction conditions used in many published studies are often biologically irrelevant. Numerous reports employ only water or PBS, or use cell culture media without serum supplementation. Reaction times are typically limited to minutes or a few hours. However, the concept of bioorthogonal therapy presumes delivery of Pd-based catalysts to tumour cells within a living organism, an environment that contains essentially 100% serum. For therapeutic application, a catalyst would therefore need to retain its activity over extended periods, ideally several days to weeks. To address this requirement, we assessed long-term catalytic activity under serum-containing conditions. Reactions were performed directly in 96-well plates, with each catalyst tested in triplicate over a period of 7 days. Between measurements, plates were stored at room temperature in a humid chamber to prevent evaporation. Catalytic activity was evaluated in PBS and in DMEM/F12 cell culture medium supplemented with 10% of foetal bovine serum (FBS). Specific results obtained for specific types of reactions are described below. Suzuki-Miyaura reaction The reaction is a C-C bond-forming coupling between an organoboron and aryl halide. An aqueous reaction between phenylboronic acid and 4-bromo-2,1,3-benzothiadiazole has previously been shown to generate green fluorescent product in breast cancer cells13. We evaluated the activity of Pd catalysts in PBS (Fig. 2f), where the reaction proceeded efficiently and was completed within 2 days. PdTPP and PdTPAs were the least active catalysts, whereas their dipicolinic acid analogues exhibited markedly improved activity. However, the reaction completely failed in serum-containing medium, and no green fluorescence was observed in cells (data not shown). Boronic acid derivatives readily form esters with biogenic diols, including mono- or polysaccharides. This property has been widely exploited in the development of saccharide sensors58. Therefore, it is plausible that phenylboronic acid, used as substrate in this fluorescent Suzuki-Miyaura reaction, is rapidly sequestered by polysaccharides present in serum. Our findings are in agreement with those of Destito et al.59, who reported that formation of a blue fluorescent product from the reaction between 2-thiophenylboronic acid and p-iodobenzaldehyde is reduced in the presence of glucose and completely abolished by glutathione. Tsuji-Trost reaction The Tsuji-Trost reaction is the most commonly used strategy for releasing an active compound from a propargylated or allylated precursor. We evaluated a broad panel of probes and monitored their deprotection for up to 7 days. All tested catalysts exhibited moderate to excellent activity in PBS. For U-PE deprotection in PBS, Pd(dipic)TFP showed the highest efficiency, followed by Pd(OAc)2 (Fig. 3a). In the presence of 10% FBS, however, almost all catalysts were rapidly inactivated, and only PdTPSb retained long term activity (Fig. 3b). A similar pattern was observed with resorufin propargyl ether18(Res-PE) (Fig. 3c andd). The fluorescence even in the presence of PdTPSb does not approach the fluorescence in PBS for two likely reasons. The catalyst reached partial conversion of Res-PE only and the released resorufin fluorescence is suppressed by FBS as mentioned before. In vitro deprotection of U-PE and Res-PE using a panel of Pd-based catalysts. (a) and (c) show Tsuji-Trost reaction kinetics in PBS, while (b) and (d) demonstrate reaction kinetics in PBS supplemented with 10% FBS. X-axis: Time in hours. Y-axis: Fluorescence intensity (a.u.). Umbelliferone excitation 365 nm, emission 450 nm, gain 50 for (a, b) and resorufin excitation 560 nm, emission 595 nm, gain 50 for (c, d). Interestingly, bis(alloc)bis(ethoxy)carboxyrhodamine 110 (BABEcRh110) was most effectively deprotected by PdTFP in the presence of FBS, and by both PdTFP and PdTPSb in PBS (Fig. S14). Activation of cresyl violet bis(allylcarbamate)60(BACV) proceeded the most efficiently by PdTPSb, regardless of the presence or absence of serum (Fig. S15). Bis(propargyloxy) dichlorofluorescein61(BP-DCF) was activated in serum only by PdTPSb, whereas in PBS the most productive catalyst was Pd(dipic)TFP (Fig. S16). HPTS-PE was likewise activated by PdTPSb in the presence of serum. Notably, this was the only probe that reached fluorescence levels comparable to those observed in PBS, where Pd(dipic)TFP was the most effective catalyst (Fig. S17). Propargyloxystyryl indolium iodide62(HSI-PE) again showed the highest activation by PdTPSb in serum and by Pd(dip)TFP in PBS (Fig. S18). Rhodol propargyl ether (ROL-PE) was activated in serum exclusively by PdTPSb, whereas in PBS the most effective catalyst was Pd(OAc)2, followed by PdTFP (Fig. S19). CM480B probe, a coumarin precursor that undergoes intramolecular cyclization upon deprotection to yield a coumarin dye63, was activated only by PdTPSb in serum, while in PBS Pd(OAc)2, and PdTFP showed comparable activity (Fig. S20). SNARF-PE was activated by PdTPSb in serum, whereas Pd(OAc)2 and PdTFP were the most effective catalysts in PBS (Fig. S21). We also examined the activation of ethidium bromide bis(allylcarbamate)9(BAEB) in vitro. Due to the low intrinsic fluorescence of free ethidium bromide (EB), the reactions were supplemented with 1 µg/mL of salmon sperm DNA. EB is an intercalating dye whose fluorescence increases dramatically upon binding to nucleic acids and can be excited under UV or at visible light (Fig. S22). Only very modest activation was observed in PBS in the presence of PdTPSb. In FBS, all reactions displayed a gradual decline in fluorescence, with no detectable probe activation (Fig. S23). In conclusion, a mixed-ligand catalyst Pd(dipic)TFP exhibited superior activity in PBS for the activation of U-PE, Res-PE, BP-DCF, HPTS-PE, HSI-PE. The high catalytic activity of the Pd complexes containing tridentate dipicolinic acid ligand in PBS might be caused by a lower stability of Pd-O coordination bond, which is also reflected by the formation of two isomers (Fig. 1), allowing easier bonding of the Pd with the alkyne moiety. In the presence of FBS, PdTPSb was consistently the most effective catalyst, with the sole exception of BABEcRh110, which was the most efficiently converted by PdTFP. This resilience could be the consequence of Pd coordination to a benzene ring after the TPSb fission. In all cases, the presence of bovine serum leads to a dramatic reduction of catalytic activity of all studied catalyst. The level of activation of all probes with all catalysts after 24 and 48 h are summarised in Fig. 4 as a fold-change. Fold increase of the fluorescence of various probes with the whole panel of catalysts after (a, c) 24 h, and (b, d) 48 h of the reaction time. The catalytic activation was carried out in (a, b) in PBS or in (c, d) DMEM cells culture media containing 10% of foetal bovine serum. The values are presented as a heat-map for clarity. Cytotoxicity of Pd catalysts During the preliminary experiments, we have noticed the toxicity of some catalysts. That is why we carried out the MTT viability assay of a few selected catalysts and ligands (Fig. S24) displaying a dramatic toxicity of Pd(TPSb)2Cl2. Having an approximate range of the toxicities, we have conducted more precise viability assay in the presence of all catalysts and ligands with CellTiter Glo Cell Viability assay, which measures cell viability by quantifying ATP, a direct marker of metabolically active cells (Fig. 5). The experiments confirmed marked toxicity of PdTPSb as well as Pd(dip)TPSb. The IC50 for PdTPSb was 2.48 µM (95% CI 1.83–3.08 µM) and 2.89 µM (95% CI 2.14–3.66 µM) for Pd(dipic)TPSb. In contrast, the other catalysts were well tolerated up to 25 µM, at which viability decreased by only about 20%. Interestingly, TPSb ligand does not show any significant toxicity itself. We presume that the lack of the toxicity is mainly due to very poor solubility of the TPSb in aqueous media. The same most likely apply for the lack of toxicity of the TPAs and of the derived Pd-catalysts, which are all very poorly soluble. The solubility of the antimony-based Pd-catalysts in water is, for unknown reasons, greatly improved, allowing cellular uptake leading to the elevated toxicity, probably by generating reactive oxygen species (ROS) or interfering with the mitochondrial respiratory chain. This finding significantly influenced the design of subsequent cell culture experiments, as much lower concentrations of PdTPSb had to be used compared to other catalysts. Cytotoxicity analysis of selected Pd-based catalysts and ligands. Experiments were carried out using HEK-293 cells treated with Pd-based catalysts and ligands for 24 h. The viability was measured with CellTiter Glo ® assay. Catalytic activity in living cells Having established that PdTPSb is the most effective catalyst in the presence of serum and the most resistant to biogenic thiols, we next evaluated the ability of the catalysts to deprotect various fluorescent probes inside cells. HEK-293 cells were incubated with each probe overnight, washed with PBS, and then further incubated for 6 h in complete medium containing the catalyst. Cellular experiments revealed that BABEcRh110 was effectively deprotected by PdTFP and Pd(dipic)TFP (Fig. 6). These results contrast sharply with the in vitro assay, where significant activity was observed only for PdTPSb. HPTS-PE activation occurred with both Pd(dipic)TFP and PdTPSb (Fig. S25). In this case, activation by PdTPSb aligns with the in vitro cell-free data; Pd(dipic)TFP also showed activity in vitro, although to a much lower extent than PdTPSb. Activation of BACV (Fig. S26) and BAEB (Fig. S27) followed the same pattern, with deprotection observed only for PdTFP and Pd(dipic)TFP. The discrepancy between in vivo and in vitro activity of PdTPSb might be related to concentration of the catalyst. We have shown that PdTPSb is toxic to cells (Fig. 5) and used lower catalyst concentration, which might influence the outcome. The ability of PdTFP catalysts to outperform other catalytic compounds in vivo might be also related to the physiological conditions inside cells. One possible reason might be the redox state. The intracellular concentration of the oxygen is greatly reduced in comparison to the in vitro assays. Alternatively, the live cells actively produce reducing agents (NADH, NADPH, FADH2), which can promote the cycling of Pd atoms in between Pd2+ and Pd0, necessary for the catalytic activity. Activation of BABEcRh110 in cells. HEK-293 cells were incubated with 25 µM BABEcRh110 overnight. After removal of the medium containing the probe, the cells were incubated with a fresh medium containing 10% of FBS and the Pd catalysts at the concentration indicated at the image. Surprisingly, activation of BP-DCF, SNARF-PE, Rhodol-PE, and several other probes could not be detected in cells after 6 h of incubation with Pd-catalysts. This discrepancy was unexpected, particularly since each probe has been reported in the literature and all responded to Pd catalysis in cell-free conditions. We were therefore prompted to investigate why no detectable fluorescence appeared under these conditions, despite their reactivity having been proven in previous experiments. Cellular retention of fluorescent dyes The absence of any detectable signal in cells pre-treated with the probes after 6 h of incubation with Pd catalysts led us to investigate whether live cells are capable of retaining the fluorescent dyes. Notably, nearly all compounds that failed to produce observable cellular fluorescence were hydroxylated fluorophores converted into their corresponding propargyl ethers. Thus, the simplest explanation is that these probes are not retained by cells for extended periods or that the deprotected fluorescent dyes are eliminated from cells. To evaluate the probe retention inside cells is rather challenging due to lack of detectable fluorescence of many probes. Instead, we focussed on the retention of the fluorescent dyes themselves. The cells were treated with individual fluorescent compounds for up to 14 h, and fluorescence images were collected at multiple time points. Before each imaging step, the dye-containing medium was replaced with fresh culture medium without the dye. All images were acquired using identical exposure settings, which resulted in intentionally overexposed images at time zero. In every case, cells appeared brightly fluorescent immediately after removal of the dye-containing medium. However, a substantial decline in fluorescence intensity was evident within 30 min. For rhodamine 110, residual fluorescence was detected after 4 h but was completely absent by 6 h (Fig. 7a, d). Fluorescein fluorescence disappeared entirely within 2 h (Fig. 7b, e), while hydroxyl Nile red showed no detectable signal after just 1 h (Fig. 7c, f). Similar results were obtained for rhodol (Fig. S28), SNARF (Fig. S29), resorufin (Fig. S29), and BINC-2 fluor (Fig. S30). Collectively, these results clearly demonstrate that although hydroxy-based probes perform effectively in cell-free conditions, their utility in living cells imaging is severely limited. The rapid loss of the released fluorophores from live cells makes these probes unsuitable for long-term in vivo experiments and restricts their applicability to very short time window. Cellular retention of fluorescent dyes in live HEK-293 cells. The cells were treated with (a) 10 µM Rhodamine 110 for 17 h, (b) 10 µM fluorescein for 14 h, (c) 10 µM hydroxyl Nile Red for 3.5 h. After the media exchange, the cells were photographed at given time points with constant exposure settings. The medium was always replaced with a fresh one just before image acquisition. The fixed exposure time is responsible for the overexposure at T 0 h. The inset shows phase contrast images of the same field. The images were onverted to heat-maps and the maximal fluorescence intensities plotted into graphs showing fast fluorescence intensity decline of all three fluorochromes (d) rhodamine 110, (e) fluorescein, and (f) hydroxyl Nile red. The cell’s inability to retain fluorescent dyes and potentially probes for an extended period of time prompted us to explore strategies for stabilising dyes inside cells. A straightforward approach would be covalent linking of the dyes or probes to a cationic nanocarrier. However, this is only feasible for compounds bearing a suitable functional moiety, such as free carboxyl or amino groups. Carboxylated dyes include 5(6)-carboxy xanthenes such as fluorescein, rhodamine or rhodol derivatives. Unfortunately, most available probes lack such a functional groups. In contrast, HPTS and HPTS-PE are highly acidic molecules carrying three sulphonyl groups, suggesting that they might form a strong complexes with cationic compounds. We have prepared complexes of HPTS with spermine, spermidine, chitosan, and polybrene, used them to treat HEK293 cells. Surprisingly, the polybrene complex exhibited nearly undiminished fluorescence for up to 10 h (Fig. S32) while the fluorescence of HPTS alone became undetectable after 30 min (data not shown). However, this treatment required very high concentrations of both the dye and polybrene, which may limit its applicability, especially for compounds with even mild cytotoxicity. Conclusions Palladium-catalysed bioorthogonal reactions show great promise in many areas of biological chemistry, however, Pd-catalysts suffer from a rapid inactivation in biological systems, a fact that is not widely emphasized in many publications. We identified triphenylantimony as a new ligand for Pd catalysts, conferring exceptional resistance to thiol poisoning and retaining catalytic activity for up to 7 days in the presence of 10% FBS in cell culture media. However, we discovered that catalysts efficiency in the presence of serum does not necessarily correlate with the performance in living cells. Although PdTPSb is the most effective catalyst in serum-containing environments, PdTFP proved substantially more efficient in deprotecting several probes in living cells. Our data reveal a distinct and probe-specific relationship among the caged compound, the catalyst, and potential cellular inhibitors. This complexity suggests that designing a universal catalyst for in vivo applications may be unrealistic; instead, each probe or prodrug may require a tailored catalyst optimised for its particular chemical structure and for the intended intracellular and extracellular context. Another, often neglected, issue is related to the effective long-term monitoring of catalytic activity in vivo. We demonstrate that hydroxylated fluorophores are poorly retained in live cells, making them unsuitable as long-term reporters in vivo. Therefore, monitoring catalytic activity inside cells may be more reliable when using dyes with strong affinity for nucleic acids, such as EB or cresyl violet, which are stabilised intracellularly through complex formation with RNA or DNA. Alternatively, hydroxylated probes could be covalently attached to polymeric carriers (e.g. polyethylene imine or chitosan) that are retained more effectively within cells. Although the cytotoxicity of PdTPSb currently precludes its in vivo use, this finding points toward an alternative approach to the catalyst design aimed at reducing its toxicity. An additional possibility is the encapsulation of PdTPSb into nanocarriers, which could shield cellular compounds from direct interaction with catalysts, thus mitigating its harmful toxicity. Materials and methods The chemicals were obtained from commercial suppliers and were used without further purification. The1 H, 13C and19 F NMR spectra of samples were measured in the CDCl3 solution on a spectrometer Bruker Avance NEO 500 MHz (700.25 MHz for1 H, 176.08 MHz for13 C nucleus equipped with a 5 mm cryoprobe Prodigy. For measurement 5 mm NMR tubes Norell 508 HP7 have been used. For the measurement and data processing SW Bruker BioSpin Topspin version 4.08 (Linux OS) and Topspin 3.6 (Windows 10), as well as SW ACD Labs ChemSketch Spectrus were used. All measurements are performed in accordance with the rules of GLP. Absorption, excitation and emission spectra and fluorescence were measured using the Spectrophotometer Infinite M200 PRO (Tecan, Männendorf, Switzerland) in Costar® (UV plate, 96 wells; Corning Incorporate, Corning, NY, USA). Fluorescent images were collected on Olympus IX53 inverted fluorescent microscope. Nitrophenol reduction All the catalysts were analysed in 200 µL reaction volume containing 0.2 mM p-nitrophenol, 5 mM NaBH4, and 10 µM Pd-catalyst. The absorbance at 400 nm was measured after 1 h of incubation at ambient temperature. Orange G reduction All the catalysts were analysed in 200 µL reaction volume containing 0.2 mM orange G, 5 mM NaBH4, and 10 µM Pd-catalyst. The absorbance at 400 nm was measured after 1 h of incubation at ambient temperature. Methyl orange reduction All the catalysts were analysed in 200 µL reaction volume containing 0.2 mM methyl orange, 5 mM NaBH4, and 10 µM Pd-catalyst. The absorbance at 465 nm was measured after 1 h of incubation at ambient temperature. Benzofuran-2-boronic acid homocoupling The reaction was carried out in 200 µL reaction volume with 200 µM benzofuran-2-boronic acid, 10 µM Pd catalyst, and 10 mM diisopropylethylamine (DIPEA) in acetonitrile for 60 min at 60 °C. All reactions were performed in triplicate. After the end of the reaction, 150 µL of reaction mix was transferred to a 96-well plate and the fluorescence was measured (Excitation: 320 nm, Emission: 360 nm, Gain: 50). Serum activation of resorufin allyl carbonate In a 96-well plate was prepared 200 µL of the reaction mix containing 50 µM resorufin allyl carbonate in PBS or in 10% rabbit, goat and bovine serum in PBS or in 100 mM NaHCO3. The fluorescence was measured at indicated intervals with excitation at 535 nm and emission 595 nm with gain 50. As a control we measured fluorescence of 50 µM resorufin dissolved in the same serum solutions. Reductive dehalogenation of iodoBODIPY IodoBODIPY was synthesized according to Park et al.64.Reductive dehalogenation was carried out according to Keum et al.43 as described previously40. In 200 µL H2O in a 96-well plate with 10 µM of the Pd-catalyst and 50 µM I.BODIPY and 0.5 mM NaBH4. The fluorescence was measured after 1 h incubation at ambient temperature with the excitation at 418 nm, emission 515 nm, and gain 40. Suzuki–Miyaura reaction The reaction was carried out in a 96-well plate in 200 µL reaction volume in PBS or in DMEM/F12 + 10% FCS cell culture medium. The reaction consists of 0.5 mM 4-bromo-2,1,3-benzothiadiazole, 0.5 mM phenylboronic acid, and 20 µM Pd-catalyst. The fluorescence with excitation 365 nm, emission 485 nm, gain 60 was monitored for 7 days. The plate was kept in between measurements in a humidified chamber to prevent solvent evaporation. Resistance of Pd catalysts to thiol inactivation Each studied Pd catalyst (PdTFP, PdTPP, and PdTPSb) was added to a 50 µM solution of a respective probe in PBS to obtain a final catalyst concentration 10 µM. L-Cysteine, L-Glutathione (biogenic thiols), and thiophenol (abiotic thiol) were introduced at 0, 0.5, 1, and 2 molar equivalents relative to the catalysts concentration to evaluate potential concentration-dependent inhibition. Importantly, the thiols must be added into the reaction prior to catalyst introduction, as Pd complexes exhibit immediate catalytic activity upon contact with the substrate. Fluorescence kinetics were monitored using the Spectrophotometer Infinite M200 PRO (Tecan, Männedorf, Switzerland) in a time course of 0–216 h. Activation of U-PE was recorded using excitation wavelength 365 nm, emission wavelength 450 nm, gain 70; SNARF-PE – excitation wavelength 580 nm, emission wavelength 635 nm, G80; HPTS-PE – excitation wavelength 490 nm, emission wavelength 530 nm, gain 50). All assays were conducted in 96-well microplates Costar® (UV plate, 96 wells; Corning Incorporate, Corning, NY, USA), which were maintained in a humidified chamber between measurements to minimize the solvent evaporation and concentration drift. Long-term in vitro probes activation A 50 µM probe solution in PBS or in DMEM/12 + 10% FBS was prepared. 200 µL of the solution was transferred to each well in 96-well plate, and to each well were added 2 µL of 1 mM Pd-catalyst, providing a final concentration of 10 µM. The fluorescence was measured immediately after the catalyst addition, followed by measurements after 30 min, 1, 2, 4, 6 h and each consecutive day. The plates were in between measurement kept in a humid chamber to prevent drying. The experiments were carried for up to 7 days. Spectrophotometric characterization For each measurement, 1 µL of the tested compound (10 mM in DMF) was added to 200 µL of the respective solvent to obtain a final concentration of 50 µM. Solvents included aqueous buffers (100 mM NaH₂PO₄, pH 5.5; PBS, pH 7.4; 100 mM NaHCO₃, pH 9), hydroxypropyl-β-cyclodextrin (10% w/v in H₂O), and a panel of organic solvents (methanol, ethanol, isopropanol, n-butanol, acetonitrile, and DMSO). The interactions with biomacromolecules were tested in PBS supplemented with yeast RNA, salmon sperm DNA, or bovine serum albumin (each at 1 mg/mL). All measurements were carried out in 96-well microplates Costar® (UV plate, 96 wells; Corning Incorporate, Corning, NY, USA), employing Spectrophotometer Infinite M200 PRO (Tecan, Männedorf, Switzerland). Probes activation in living cells HEK-293T cells (ATCC, CRL-3216) were cultivated in complete cells culture medium consisting of DMEM/F12 supplemented with 10% FBS, 1% Penicillin-Streptomycin and 2 mM glutamine. The HEK-293T cells were seeded at 40–50% confluence. After cells attachment, the cells were treated with 25 µM probe in complete medium over night. Following morning, the cell culture media containing the probe was removed, cells were washed with PBS, and fresh complete medium containing the Pd catalysts was applied. The cells were incubated in presence of the catalysts for 6–9 h. After the medium with catalyst was removed, cells were washed with PBS and new complete medium was added and fluorescent microscopy images were collected. The palladium catalysts were used at 50 µM concentrations except for PdTPSb and Pd(dipic)TPSb, which were 12.5 µM. The low concentration was used to prevent immediate cells death due to high cytotoxicity of Sb-based compounds. HPTS-PE complexation with polybrene to increase intracellular retention Due to poor intracellular retention of highly negatively charged HPTS-PE, a complex with cationic polymer polybrene was prepared and used for the probe delivery. 20 mM solution of HPTS-PE in water was mixed with 10 mg/mL solution of polybrene in water. The compounds were incubated for 15 min and 0.5 mL of complete cell culture media was added to yield 250 µM HPTS-PE and 750 µM polybrene (related to the monomer). The mix was transferred to 24-well plate with HEK293T cells and incubated overnight. Cytotoxicity of the palladium catalysts – MTT assay The cytotoxicity assay was done with the use of colorimetric reduction of water soluble 3-(4,5-di methyl thiazol-2-yl)−2,5-diphenyltetrazolium bromide (MTT) to insoluble formazan. During the assay were HEK293 cells cultivated in a complete medium consisting of DMEM/F12 supplemented with 10% FBS, 1% Penicillin-Streptomycin and 2 mM glutamine in a humidified CO2 incubator at 37 °C. The HEK293 cells seeded in triplicate to an anticipated confluence of 30% in a flat-bottom tissue culture treated 96-well plate. The cells were allowed to attach to the bottom of the wells overnight. The following morning the cell culture media were removed, and fresh complete medium containing the Pd catalysts at final concentration of zero, 500 nM, 1 µM, 5 µM, 10 µM, 15 µM, 20 µM, 25 µM, 30 µM in 150 µl of complete media were applied. The cells were incubated with the mixture for 48 h without medium exchange. After the medium with catalyst was removed, 100 µL of fresh complete medium containing 0.25 mg/mL of MTT was added, and the cells were incubated for 1.5 h. After the medium with MTT was removed, cells were carefully washed with PBS, and 100 µL of solubilisation solution (0.04 M HCl in isopropanol with 0.1% NP-40) was added to the wells. The absorbance of the formazan was measured at 570 nm on the Spectrophotometer Infinite M200 PRO. Cytotoxicity of the palladium catalysts - CellTiter-Glo luminescent cell viability assay The cytotoxicity assay was performed using CellTiter-Glo® Luminescent Cell Viability Assay, a method based on luminescent quantification of ATP produced by metabolically active cells. During the assay, HEK293 cells were cultivated in a complete medium consisting of DMEM/F12 supplemented with 10% FBS, 1% Penicillin-Streptomycin, and 2 mM glutamine in an incubator set at 5% CO2 and 37 °C. The HEK293 cells were seeded in triplicate, 10 000 cells were seeded into each well. For these experiments, a flat-bottom tissue culture 96-well plate was used. The cells were allowed to attach to the bottom of the wells overnight. After the cells attached, the cell culture medium was replaced with fresh one and the treatment was applied. Pd catalysts or ligands were diluted in medium and added to cells at a final concentration of 0.5 µM, 1 µM, 2.5 µM, 5 µM, 7.5 µM, 10 µM, 15 µM, 20 µM, and 25 µM. The cells were further incubated for 24 h in an incubator set at 5% CO2 and 37 °C. After incubation, the medium with tested compound was removed and replaced with 50 µl of fresh medium. Subsequently, the same amount of CellTiter-Glo® working solution was added. The plates were placed on the shaker for 2 min, 30 rpm, in the dark. After that, another 10 min incubation in the dark followed, and cells lysate was transferred onto white 96- well plates. The luminescence was measured using a multifunctional reader Synergy H1 (Agilent). IC50 for cytotoxic catalysts PdTPSb and Pd(dipic)TPSb were determined by nonlinear regression analysis using a four-parameter dose-response model. Cellular retention of florescent dyes Cells were incubated with 10 µM solution of the fluorescent dye in complete media for 16 h. Then, the medium was replaced with a fresh complete medium without the dye and cells were photographed. The light intensity and the exposure times were maintained the same throughout the experiment. The procedure was repeated at indicated time intervals. Fluorescent images were analyzed using ImageJ software. Fluorescence intensity was visualized as a pseudocolor heat maps using a fixed look-up table (LUT) applied equally to all samples, and the values for maximum intensities of respective images were exported and compared. Chemical syntheses The synthetic procedures and compounds characterisations are provided in Supplementary Materials. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Acknowledgements This work was supported by the Czech Science Foundation (project no. 23-04740 M). J.K. expresses her sincere gratitude to Internal Agency of Faculty of AgriSciences, Mendel University in Brno (project no. IGA24-AF-IP-015) for funding. CIISB research infrastructure project LM2018127 funded by MEYS CR is gratefully acknowledged for the financial support of the measurements at the Josef Dadok National NMR Centre. JK expresses her sincere gratitude to PhD Talent Brno. Author contributions M.K.P. conceived, planned, and carried out the experiments, J.K. conceived, planned, and carried out the experiments, L.J. carried out the NMR characterisation of the compounds. S.O. performed fluorescent microscopy experiments, J.B. carried out the experiments, P.T. carried out the experiments, Z.H. contributed to the interpretation of the results, acquired the financial support for the project, participation in manuscript refining, V.P. oversight and leadership responsibility for the research activity planning and execution, chemical synthesis experiments, wrote the manuscript. All authors discussed the results and commented on themanuscript. Data availability The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files. Should any raw data files be needed in another format or any additional information are required they are available from the corresponding author upon reasonable request. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Jana Kurcova and Marie Kudlickova Peskova contributed equally to this work. Contributor Information Zbynek Heger, Email: zbynek.heger@mendelu.cz. Vladimir Pekarik, Email: pekarikv@mail.muni.cz. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Data Availability Statement The authors declare that the data supporting the findings of this study are available within the paper and its Supplementary Information files. Should any raw data files be needed in another format or any additional information are required they are available from the corresponding author upon reasonable request.