Control of Tetrazine Bioorthogonal Reactivity by Rotaxanation

Abstract

Rotaxanation is an efficient method to control the tetrazine‐mediated inverse electron demand Diels–Alder (IEDDA) reaction. Tetrazine rotaxanes were synthesized in high yield by crown ether active template applied to the nucleophilic aromatic substitution of tetrazines. Kinetics of the bioorthogonal reaction with tetrazine rotaxanes were much slower than those with the corresponding threads. Interestingly, disassembly of the mechanical bond upon the application of the right stimulus activates IEDDA. Control of IEDDA in live cells was proved by means of a β‐galactosidase sensitive tetrazine rotaxane: Enzymatic digestion allowed the Diels–Alder reaction, which resulted in a fluorescent compound.

Control of tetrazine‐mediated inverse electron demand Diels–Alder (IEDDA) reaction has been achieved by rotaxanation. Tetrazine rotaxanes were synthesized in high yield and they displayed a much slower kinetics compared to the corresponding thread. Mechanical bond disassembly upon the right stimulus activates reactivity. Indeed, IEDDA could be selectively activated inside senescent surrogate cells by a β‐galactosidase‐responsive rotaxane.

Studying the intricate biochemical processes inside cells is a challenging task. It is essential to rely on a special set of selective high‐yielding reactions that can take place in a complex matrix such as a biological environment, at a considerable rate and without affecting endogenous functional groups. Those reactions constitute the foundations of the bio‐orthogonal chemistry,^{[ 1} , ² , ³ , ⁴ , ^{5 ]} which was developed as a remarkable tool for labeling biomolecules and for studying living systems, but has been later applied even to materials science, drug delivery or molecular biology.^{[ 4 ]} Among the numerous bio‐orthogonal reactions that have been developed, tetrazine‐mediated inverse electron demand Diels–Alder (IEDDA) reaction, usually with strained alkynes and alkenes, is one of the most widely used due to its fast reaction rate, excellent bio‐orthogonality, and great biocompatibility.^{[ 6 ]} However, even after such an overwhelming success of tetrazine ligation, the control of bio‐orthogonal reactivity in IEDDA has not been adequately developed yet. So far, only three approaches to control the reactivity of the tetrazine ring towards IEDDA reaction have been described. Firstly, reduction to unreactive dihydrotetrazine (Figure 1a) has been explored by Devaraj et al.,^{[ 7} , ⁸ , ^{9 ]} Fox et al.,^{[ 10} , ¹¹ , ^{12 ]} and others.^{[ 13} , ^{14 ]} The right stimulus (light irradiation or enzymes) allows the re‐oxidization of the tetrazine ring, and therefore it activates the IEDDA reaction. Another interesting approach is the inhibition of the reactivity of the tetrazine ring by strain (Figure 1b), developed by Mizukami et al.^{[ 15 ]} Indeed, the high distortion displayed by some macrocyclic tetrazines, causes the inhibition of the Diels–Alder reaction which can be eventually reactivated by cleaving the macrocycle, releasing the strain. Finally, the third approach takes advantage of host–guest chemistry (Figure 1c). Liu et al. proved that high‐affinity supramolecular caging of a given tetrazine compound inside a naphthotube, slowed down the kinetics of the IEDDA.^{[ 16 ]} Reactivation in this case is triggered by an even better guest that displaces the tetrazine out of the receptor.^{[ 17} , ¹⁸ , ^{19 ]}

Figure 1.

Strategies to control tetrazine‐mediated IEDDA: a) Reduction and reoxidation of the tetrazine ring; b) IEDDA is inhibited in strained macrocyclic tetrazines; c) Caging of tetrazine compounds inside a naphtotube slows down their reactivity; d) In this work, control of IEDDA by rotaxanation is reported.

Herein, we suggest an alternative approach: Controlling the reactivity of the tetrazine by rotaxanation (Figure 1d). Rotaxanes are one of the three different types of mechanical bond,^{[ 20 ]} together with catenanes and clippanes.^{[ 21 ]} They are composed by a macrocycle and a molecular axle or thread, which are linked together due to the presence of two bulky stoppers on both ends of the thread, that prevents the macrocycle from slipping away. Rotaxanes have found applicability in several different areas of chemistry, such as molecular electronics,^{[ 22 ]} polymer chemistry,^{[ 23 ]} catalysis,^{[ 24 ]} or molecular machines.^{[ 25 ]} A very interesting property of rotaxanes is that the macrocycle can act as an exotic kind of protecting group for the thread. Indeed, there are reported examples of compounds protected from nucleophilic substitution,^{[ 26} , ²⁷ , ^{28 ]} or from enzymatic digestion by rotaxanation.^{[ 29 ]} We hypothesize that kinetics of IEDDA reaction could be hampered by the macrocycle, which avoids the optimal approach of the dienophile. In this sense, rotaxane control would be reminiscent to Liu's approach (Figure 1c),^{[ 16 ]} as well as it could be considered the extreme case of well‐known steric inhibition by bulky substituents on the tetrazine ring.^{[ 30} , ^{31 ]}

Interestingly, decades of progress in rotaxane chemistry, allowed for an efficient synthesis of the tetrazine rotaxanes. In fact, they were successfully synthesized thanks to the so‐called crown ether active template (CEAT),^{[ 32 ]} recently reported by Leigh et al.,^{[ 33} , ^{34 ]} They proved that [2]rotaxanes can be synthesized in one step, from primary amines, crown ethers and a range of electrophiles. In this regard, we have explored the dynamic nucleophilic aromatic substitution of tetrazines with phenols and thiols,^{[ 35} , ³⁶ , ³⁷ , ^{38 ]} and more recently, we found that alkyl amines react efficiently with those phenol and thiol tetrazine dimers, only on one side of the tetrazine due to kinetic effects.^{[ 39 ]} It seems reasonable to think that combining the latest with CEAT, then tetrazine rotaxanes should be easily synthesized (Figure 2a). Indeed, after a small optimization study (Table S1), we found that ideal conditions were concordant with those previously reported by Leigh.^{[ 33} , ^{34 ]} An equimolar mixture of 3,5‐dimethoxyphenol dimer of tetrazine 1 as electrophile, 24‐crown‐8 ether, and 3,5‐bis(trifluoromethyl) benzylamine 2 as nucleophile in toluene (0.1 M) at 25 °C, led to 92% yield of rotaxane 3 and > 100:1 rotaxane:axle selectivity (Figure 2b). Not surprisingly, the monochloro‐tetrazine phenol 4 gave worse yield than diphenolic compound 1 (Figure 2c). We had reported a combined mechanism of the nucleophilic substitution of the amine in such case: Both the chloride and the phenol can act as leaving groups,^{[ 39 ]} being the latter case non‐productive for the rotaxanation. Thiol dimer 5 led to lower yields (Figure 2d), although the advantage of sulfide compounds is that they could be oxidized to the corresponding sulfones 7 with oxone,^{[ 40 ]} which leads to more electronic poor tetrazine rings (Figure 2f).

Figure 2.

a) Synthesis of tetrazine rotaxanes was carried out by a combination of CEAT and nucleophilic aromatic substitution of tetrazines. b) O,N‐Tz rotaxane synthesis led to very good yields. c) Mono‐chloro tetrazine compounds are less efficient than diphenolic compounds for the synthesis of rotaxanes. d) S,N‐Tz rotaxanes gave lower yields than the phenolic counterparts. e) Single‐crystal X‐ray molecular structure of 3. CCDC deposition number for 3: 2470276 f) Oxidation of S,N‐Tz rotaxane 6 was successfully carried out with Oxone.

Kinetics of the IEDDA reaction between the tetrazine compounds and bicyclononyne (BCN) in THF‐PBS (1:1) at 37 °C was followed by UV–vis spectroscopy at 520 nm, due to the disappearance of the typical red–orange color of the tetrazine as the reaction progresses (Figures S1–S6). As expected, the second order rate constant of the phenol‐amine tetrazine 3‐thread (k₂ (thr) = 0.046 M⁻¹·s⁻¹) was on the lower range of reported IEDDA rates (Figure 3a),^{[ 5} , ^{6 ]} due to electronic effects, although it is comparable to other reported reactions of BCN with diphenyl‐tetrazine,^{[ 41} , ^{42 ]} and also similar to strain promoted alkyne azide cycloaddition (SPAAC) reactions.^{[ 5} , ^{43 ]} Fortunately, the more electronic poor sulfone 7‐thread was expectedly faster (k₂ (thr) = 0.44 M⁻¹·s⁻¹) (Figure 3b).^{[ 44} , ^{45 ]} It is well known that BCN leads to slower reactions than trans‐cyclooctene (TCO).^{[ 6} , ⁴⁶ , ^{47 ]} Indeed, reaction of TCO with both 3‐thread and 7‐thread were notably faster than those with BCN (Figures S11–S12), and could be probably even faster with other TCO derivatives.^{[ 48 ]} However, most importantly, in all cases cycloaddition is slower in the rotaxanes by two orders of magnitude (Figure 3a,b). The rotaxanation protecting efficiency was calculated as [k(thread)‐k(rotaxane)]/k(thread).^{[ 49 ]} The control of IEDDA by rotaxanation is always higher than 99%.

Figure 3.

Kinetic studies of the reaction with BCN of rotaxanes versus axles, carried out in 0.5 mM of tetrazine compound THF/PBS (1:1) at 37 °C: a) O,N‐Tz rotaxane 3 offers an excellent protection although rate constants are in the lower range of IEDDA; b) Sulfone rotaxane 7 also inhibits Diels–Alder reaction, and rate constants are more appealing; c) Single‐crystal X‐ray molecular structure of the product of the reaction between 3 and BCN obtained after 4 weeks in acetonitrile. CCDC deposition number for 3_BCN: 2470277.

Our next goal was the design of systems for the activation of IEDDA by disassembly of the mechanical bond in cells upon the application of the right stimulus.^{[ 50} , ⁵¹ , ⁵² , ^{53 ]} With that aim in mind, tetrazine rotaxane 8 was synthesized (Figure 4a). Rotaxanation was carried out with the per‐acetylated precursor in toluene, giving 74% yield of 8Ac, followed by hydrolysis of the acetates. The latter proves that, even when high‐yielding rotaxanation requires precursors soluble in toluene, water solubility or hydrophilicity can be modulated afterwards by post‐functionalization. 8 displays tetraphenylethene (TPE) as one of the stoppers. TPE is the most paradigmatic example of the aggregation‐induced emission (AIE) phenomenon.^{[ 54} , ⁵⁵ , ^{56 ]} The fluorescence of this kind of compounds is greatly enhanced whenever the rotation of phenyl groups is restricted, for instance upon aggregation due to poor solubility (Figures S15–S18). In any case, fluorescence of TPE in 8 (and in 9) is quenched by the tetrazine ring, and it will only be activated as a result of the IEDDA reaction (Figure S19).^{[ 57} , ⁵⁸ , ⁵⁹ , ⁶⁰ , ^{61 ]} The other stopper is β‐galactose, which is susceptible to digestion by β‐galactosidase (β‐gal). This enzyme is, among other things, an important biomarker for eukaryotic senescent cells,^{[ 62 ]} and as a consequence, it has been successfully used for detection of senescent cells or to activate pro‐drugs.^{[ 13} , ¹⁴ , ⁶³ , ⁶⁴ , ^{65 ]} In cells with overexpression of β‐gal, the sugar stopper will be removed and therefore the macrocycle will escape, giving semi‐thread 9, in which the IEDDA reaction with BCN is much faster. Such Diels–Alder reaction yields 10, which is fluorescent (Figure 4c). It is worth mentioning that the initial rotaxane 8, with the sugar stopper is more soluble in aqueous environments than the final compound 10, which benefits the intensity of the AIE fluorescence of TPE.

Figure 4.

Control of tetrazine bio‐orthogonal reactivity by rotaxanation in live cells: a) Rotaxane 8 was designed to be disassembled by the action of β‐galactosidase (β‐gal), boosting the kinetics of IEDDA with BCN, which leads to the fluorescent compound 10. The second‐order rate constant shown on the left, was measured with per‐acetylated rotaxane 8Ac; b) quenching of the fluorescence of TPE by the tetrazine ring results in 9 being non‐fluorescent in PBS‐DMSO (99.5:0.5). However, after the Diels–Alder reaction with BCN, the resulting compound 10 shows green fluorescence; c) fluorescence intensity variation from 0 min (addition of BCN) to 80 min for each experimental condition in transfected cells. Initial background emission is probably due to autofluorescence of cellular compounds because excitation was carried out at 365 nm; d) intensity/area variation with time of the emission at 525 nm for 10 cells of each transfection condition.

Firstly, to prove that the enzyme was able to disassemble the mechanical bond, to a solution of 8 (0.84 mM) in PBS‐DMSO (99:1) was added β‐gal (10 U mL⁻¹), and the outcome followed by HPLC‐MS (Figure S20). The chromatogram of an aliquot taken after 90 min, showed no rotaxane left in the reaction mixture, but it contained semi‐thread 9 instead, which evidences that enzymatic digestion is able to disassemble the mechanical bond.

Next, we tested tetrazine‐ligation control by rotaxanation in live cells. HEK293t cells were transfected with different plasmids. On one hand, as a cellular senescence surrogate, cells were co‐transfected with two plasmids: “pCDNA4/TO/LacZ”, which encodes the enzyme β‐galactosidase, and “pBNJ13.1‐BK1088mScarlet” a BK ion channel construct containing a red fluorescent protein, mScarlet, inserted at position 1088.^{[ 66 ]} The presence of mScarlet helps identify successfully transfected cells using fluorescence microscopy because of its red fluorescence (Figures S22–S24). On the other hand, as a negative control, cells were transfected with the same “BK1088mScarlet” plasmid to keep the monitoring of transfection efficiency consistent. Instead of using “pCDNA4/TO/LacZ”, we included an empty vector plasmid, “pcDNA3” (Figures S25–S27). This ensured that the total amount of plasmid DNA used was the same in both conditions. Rotaxane 8 (final concentration: 5 µM) was added to cells, followed by BCN (final concentration: 50 µM). It is worth mentioning that cell viability was also tested, and no considerable toxicity at those concentrations was observed (Figure S28). Fluorescence intensity measurements were obtained from 10 individual cells for each transfection condition, considering time zero when BCN was added. (Figure 4c). Fluorescence values were normalized by dividing them by the cell area. This helped to account for differences in cell size. The normalized fluorescence intensities were averaged. The mean values were then plotted over time for each transfection condition. As it is clearly seen in Figure 4d, fluorescence in β‐gal containing cells starts to increase after 20 min and keeps growing with time. However, in the control cells there is no appreciable change, not even after 80 min. As a consequence, at the end of the experiment, the difference in fluorescence intensity between both kinds of cells is quite notable (Figure 4c).

In summary, a novel way to control the tetrazine‐mediated IEDDA reaction through rotaxanation has been developed. For such purpose, the synthesis of tetrazine rotaxanes was implemented by taking advantage of the CEAT and the nucleophilic aromatic substitution of tetrazines. Yields above 90% were achieved. In all cases, IEDDA kinetics is considerably slowed down in the rotaxane compared to the corresponding thread. In such a way, mechanical bond disassembly activates the bio‐orthogonal reactivity. Indeed, a responsive rotaxane displaying a stopper designed to be cleaved by enzymatic digestion by β‐galactosidase was synthesized: We proved that tetrazine‐mediated IEDDA could be selectively activated inside senescent surrogate cells.

We believe this approach is synthetically simple, robust and extremely versatile: Many other stimuli could be selected, kinetics could be improved by choosing more deactivated tetrazine threads, which could then be mechanically interlocked by means of any of the multiple reported methodologies for the synthesis of rotaxanes. Different topologies or rotaxane stoichiometries could enhance protection. And obviously, other classic properties of rotaxanes such as shuttling could be applied. There are plenty of possibilities ahead to be explored.

Supporting Information

General experimental details and procedures, X‐ray crystallographic details, NMR spectra, and biological experiments description. The authors have cited additional references within the Supporting Information.^{[ 67} , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ^{78 ]}

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Fumero‐Medina C., Pérez‐Pérez Y., Pérez‐Márquez L. A., Álvarez de la Rosa D., Giráldez T., Paz N. R., Pasán J., Lahoz F., Tejedor D., Scoccia J., Carrillo R., Angew. Chem. Int. Ed. 2025, 64, e202514796. 10.1002/anie.202514796

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.