Double Click: Unexpected 1:2 Stoichiometry in a Norbornene–Tetrazine Reaction

Abstract

We report a new reactivity for the inverse electron demand Diels–Alder (iEDDA) reaction between norbornene and tetrazine. Instead of simple 1:1 condensation between norbornene- and tetrazine-conjugated biomolecules, we observed that dimeric products were preferentially formed. As such, an olefinic intermediate formed after the addition of the first tetrazine unit to norbornene rapidly undergoes a consecutive cycloaddition reaction with a second tetrazine unit to result in a conjugate with a 1:2 stoichiometric ratio. This unexpected dimer formation was consistently observed in the reactions of both small-molecule norbornenes and tetrazines, as well as oligonucleotide conjugates. When norbornene was replaced with bicyclononyne to bypass the formation of this olefinic reaction intermediate, the reactions resulted exclusively in rapid formation of the expected 1:1 stoichiometric conjugates.

Introduction

The inverse electron demand Diels–Alder (iEDDA) reaction between tetrazine and various alkenes has been widely used in bioconjugation reactions across biology, imaging, and therapeutics.¹⁻⁸ Due to its catalyst-free nature and fast kinetics under physiological conditions,⁹ the iEDDA reaction is a powerful alternative to other commonly used biorthogonal ligation methods, such as CuAAC,¹⁰ SPAAC,¹¹ and Staudinger ligation¹² approaches. Among different subtypes of iEDDA reactions,^1,2 the tetrazine–norbornene cycloaddition has been used for biological applications including site-specific incorporation of unnatural amino acids into proteins,¹³ glycosylation of membrane proteins,¹⁴ development of cell-instructive hydrogels,¹⁵ site-specific RNA labeling in mammalian cells,¹⁶ post-transcriptional functionalization of RNAs,¹⁷ and postsynthetic DNA and RNA modifications.¹⁸⁻²¹

We have recently explored the application of the tetrazine–norbornene ligation reaction to synthesize multivalent oligonucleotides (mONs). The synthesis of mONs by traditional solid-phase DNA synthesis methods is challenging, leading to poor yield and low purity for larger multimers. Although some inline methods have been described for mON synthesis,²²⁻²⁵ they might not be ideal for some applications, including mONs with higher valency and longer sequence composition. In this context, we planned to use a postsynthetic tetrazine–norbornene ligation approach to achieve multivalent scaffolds in an efficient way. Norbornene is particularly attractive for this application since it is fully compatible with the conditions of oligonucleotide synthesis and deprotection.

In the classical norbornene–tetrazine iEDDA reaction, 1,4-cycloaddition of tetrazine to norbornene yields a highly strained bicyclic intermediate 1, which then eliminates N₂ via a retro-Diels–Alder reaction to afford intermediate 2 (Scheme 1). This intermediate isomerizes to 1,4-dihydro products 3/4 and/or is oxidized to pyridazine 5.¹ In contrast, we found that the reaction predominantly yielded a product with 1:2 stoichiometry in the context of both oligonucleotides and small-molecule derivatives.

Scheme 1. Mechanism of the Classical Norbornene–Tetrazine iEDDA Reaction.

Results and Discussion

Our strategy to build mON scaffolds included a two-component reaction between tetrazine-functionalized ONs and a multivalent norbornene core molecule. The tetrazine ONs were synthesized by reacting 5′-amino-functionalized ONs with methyltetrazine NHS ester (Tables S1, and S2 and Scheme S1). A DNA synthesizer was used to assemble a branched dimeric core Heg-Nor2 with two terminal norbornenes (Figure S1), using a norbornene phosphoramidite, which was synthesized in one step from 5-norbornene-2-methanol²¹ (Scheme S2).

We expected to obtain a divalent ON scaffold during the reaction between tet-PS-ON and Heg-Nor2 (Figure 1 and Table S1).

Figure 1.

(A, B) Schematic for conjugation between a tetrazine ON (tet-PS-ON: 5′-tet-C6-G*G*G*T*C*a*g*c*t*g*c*c*a*a*t*G*C*T*A*G-3′) and divalent norbornene core (Heg-Nor2). tet indicates methyltetrazine, C6 is a six-carbon linker, *indicates phosphorothioate, uppercase bases are 2′-MOE, and lowercase italic indicates DNA bases. (C) ESI–MS data for the conjugate showing multiple product formation after 24 h of incubation at room temperature.

But, to our surprise, this approach predominantly yielded trivalent and tetravalent ONs (Figures 1 and S2–S6). To explore this, we conducted concentration-dependent reactions between Heg-Nor2 (100 μM) and tet-PS-ON (200, 400, or 800 μM, which corresponds to 1, 2, or 4 equiv of ONs per norbornene), following the progress of the reaction by liquid chromatography–mass spectrometry (LC–MS). We found that the tri- and tetravalent products started to form even under conditions when only a single equivalent of the tetrazine partner was present relative to each norbornene (Figures S2 and S3). Predominant tri- and tetravalent product formation was observed when higher equivalents of the tetrazine partner were used (Figures 1 and S4–S6). Importantly, the mass spectra showed loss of approximately 28 Da for each tetrazine-derived moiety added, consistent with the loss of dinitrogen associated with the tetrazine iEDDA reaction.

Since this type of reactivity has not been previously reported, we questioned whether some aspect of the structures of the oligonucleotide reactant tet-PS-ON or the core structure Heg-Nor2 could be responsible for the unexpected reactivity. To test these possibilities, we next reacted tet-PS-ON with 5-norbornene-2-methanol only. Consistent with our initial findings above, the product of this reaction was predominantly divalent as confirmed by LC–MS (Figure S7), confirming that the dual reactivity is not an artifact of the di-norbornene core (Heg-Nor2) used earlier.

We then turned our attention to ruling out artifacts from the tetrazine partner. The oligonucleotide tet-PS-ON used for this study is a fully phosphorothioate–DNA gapmer with 2′-methoxyethyl (MOE) wings. To further simplify the chemistry, we next eliminated any involvement of 2′-MOE or the phosphorothioate (PS) bonds during this reaction by testing simpler, phosphodiester-linked DNA oligonucleotides tet-PO-ON1 and tet-PO-ON2 with Heg-Nor2. Despite the removal of the PS and MOE groups, we observed the same reactivity pattern for these substrates yielding a similar mixture of mono-, di-, tri- and tetravalent conjugates (Figures S8 and S9). These experiments further exclude the possibility of side reactions occurring either from PS or 2′-MOE substitutions. Therefore, based on our experimental results, we hypothesized that the product from the first norbornene–tetrazine addition remains reactive toward tetrazine to result in higher-valent products. To the best of our knowledge, this is the first report of 1:2 stoichiometry during the norbornene–tetrazine iEDDA reaction.

We considered that the double bonds of an unsaturated intermediate such as 2, 3, or 4 (Scheme 1) might react with a second equivalent of tetrazine to afford higher-valent structures. We tested this hypothesis by repeating the reaction using a strained alkyne (bicyclo[6.1.0.]nonyne, BCN) instead of norbornene. When the triple bond of the alkyne undergoes a cycloaddition reaction with tetrazine, it yields an aromatic product directly (Scheme 2). Since this reaction does not proceed via intermediates 2, 3, or 4, it cannot undergo additional reactivity during the alkyne–tetrazine conjugation reaction and should not react if our hypothesis is true. Therefore, we designed a divalent BCN core, Heg-BCN2, to react with tet-PS-ON. Gratifyingly, the reaction yielded exclusively a 1:1 divalent product, even at higher concentrations of tet-PS-ON and longer incubation time (7 d) (Figure S10). Furthermore, the reaction between tet-PS-ONs and a small-molecule derivative, which contains a single BCN unit showed only 1:1 stoichiometry in the resulting product (Figure S11). To further confirm whether the 1:2 stoichiometry is specific only to the norbornene substrate, we carried out a reaction between 3-methyl-6-phenyl-1,2,4,5-tetrazine and trans-cyclooctenol (TCO–OH, another alkene dienophile commonly used in the iEDDA reaction with tetrazines). As confirmed by HRMS data, only the monomeric adduct was observed for this reaction with no 1:2 product seen (Scheme S3 and Figure S12). Therefore, the unexpected reactivity appears specific to norbornene among the commonly used tetrazine reaction partners.

Scheme 2. iEDDA Reaction between Tetrazine and Bicyclo[6.1.0.]nonyne Yields an Aromatic Product Directly.

To examine the concentration dependence of the unexpected reactivity, we tested the possibility of dimer formation between 3-methyl-6-phenyl-1,2,4,5-tetrazine and 5-norbornene-2-methanol at equimolar ratios but at different reaction concentrations (Table S3). Dimer formation was observed in all cases, irrespective of the final reaction concentrations (Figures S13–S16).

To establish the structure of the 1:2 product, we carried out a reaction between 5-norbornene-2,2-dimethanol 7 and 3-methyl-6-phenyl-1,2,4,5-tetrazine 6 (Scheme 3). Reverse-phase HPLC demonstrated the existence of multiple products with the same molecular weight (consistent with the condensation of two tetrazine units and one norbornene unit with loss of two N₂ molecules), including one major product, which we isolated (Figure 2). We conducted thorough one-dimensional (1D) and two-dimensional (2D) NMR analyses of this major product (Table S4 and Figures S17–S24). In ¹H NMR, the integration of aromatic protons attached to the tetrazine derivative relative to norbornene protons in the aliphatic region clearly confirmed the presence of two tetrazine-derived moieties and one norbornene-derived moiety in the final product (Figure S17). The two −OH groups of norbornene-2,2-dimethanol at 4.38 and 4.14 ppm and a singlet (2H) at 5.93 ppm disappeared upon D₂O exchange in DMSO-d₆, indicative of four exchangeable protons (Figure S18). The exchangeable 2H singlet at 5.93 ppm could be accounted for by two protons attached to the same nitrogen atom, consistent with the presence of an NH₂ in the molecule. Most strikingly, we were intrigued by both the appearance of one extra proton in the aromatic region at 7.20 ppm and the disappearance of the signal corresponding to the methyl group attached to one of the two tetrazine units. Based on this detailed NMR analysis, we assigned the structure as hydrazone-containing compound 12.

Scheme 3. Plausible Mechanistic Pathway of 1:2 Reactivity between 5-Norbornene-2,2-dimethanol (7) and 3-Methyl-6-phenyl-1,2,4,5-tetrazine (6).

Figure 2.

HPLC profile of the reaction mixture showing the predominant isomer (compound 12) and the two additional product peaks. As shown, all three of these major product peaks had the same observed mass (M + H⁺ = 443.1 Da).

We carried out density functional theory (DFT) calculations to explore whether a reasonable mechanistic pathway exists to compound 12. Gratifyingly, the energetics of this series of reactions and rearrangements were predicted to be favorable (Tables S5 and S6). As such, we propose that compound 8 undergoes rearrangement to compound 9, which contains an exocyclic alkene that undergoes a consecutive cycloaddition reaction with the second tetrazine molecule 6, to result in spiro compound 10 which isomerizes to form 11 and undergoes ring-opening to 12 (Scheme S4). The regiochemistry of the cycloaddition between 6 and 9 was predicted to be strongly favored by the DFT calculations and is consistent with the literature precedent;^26,27 this regiochemistry is also confirmed by the presence of two strong heteronuclear multiple bond correlations (HMBCs) between H11 (7.20 ppm) and C14 (22.0 ppm) and CH₃ protons (2.54 ppm) and C11 (126.7 ppm) (Figures S24 and S25).

DFT calculations, both in vacuum and in a simulated solvent, showed that the second cycloaddition has substantially lower activation energy than the first and thus explains the facile formation of the bis-adduct (described in detail in supporting Tables S5 and S6 and Figure S25). The isomerization of 10 to 11 and the ring-opening of 11 to 12 are thermodynamically favored. Finally, we reasoned that if compound 12 was indeed a hydrazone-containing molecule, it would rapidly react with aldehydes to form an azine adduct. Indeed, the addition of two different aldehydes (propionaldehyde and valeraldehyde) to the condensation product rapidly led to the formation of products with the mass of the corresponding azine, independently confirming the existence of a hydrazone in structure 12 (Table S7 and Figures S26 and S27).

Given the importance of the tetrazine-derived methyl group in this mechanism, we tested the reactivity of two additional (non-methyl) tetrazine substrates with three norbornene substrates. The reaction between 3,6-diphenyl-1,2,4,5-tetrazine and norbornenes, even in the presence of excess tetrazine, produced only monomeric addition products in all three cases (Table S8 and Figures S28–S36). However, with the same substrate ratios, we found that 3-phenyl-1,2,4,5-tetrazine produced both monomeric and dimeric adducts with norbornenes based on HRMS data. The dimeric products of this reaction were not stable enough to be isolated and characterized. This and the existence of other minor dimeric products derived from methyltetrazine 6 (Figure 2, additional isomeric product peaks) suggests that while the exocyclic alkene of structure 9 is particularly susceptible to additional reactivity with tetrazine, the endocyclic double bonds such as those in intermediates 2, 3, and 4 may also be susceptible to reaction with additional equivalents of tetrazine.

Conclusions

We have identified an unprecedented 1:2 stoichiometry during the norbornene–tetrazine iEDDA reaction. While synthesizing multimeric oligonucleotides, we observed multimeric structures with higher valency than expected. Oligonucleotides with various chemical modifications and small-molecule tetrazines reacted with the norbornene substrate in a similar manner, thus confirming that the reaction is general. Thus, we propose that unsaturated intermediate 8, formed during the reaction between norbornene and the first tetrazine unit, undergoes a consecutive cycloaddition reaction to accommodate the second tetrazine molecule to afford higher-valent products. The electrospray ionization–mass spectrometry (ESI–MS) data for multimeric oligonucleotide products and the detailed 1D, 2D nuclear magnetic resonance (NMR) and LC–MS/HRMS data for the small-molecule derivatives provide insight into this reaction pathway, further supporting the 1:2 reactivity of norbornene in this type of conjugation reactions.

For ligation or conjugation approaches where base pairing directs 1:1 stoichiometry, tetrazine–norbornene chemistry can be a useful approach.²⁸ For other applications, especially where 1:1 stoichiometry is essential, researchers should choose a dienophile (such as TCO or BCN, Scheme 2) that circumvents the production of intermediates with additional reactive potential.

Experimental Section

General

Reagents and anhydrous solvents were purchased from commercial sources and were used without any further purification. Reaction progress for small-molecule work was monitored by thin-layer chromatography (TLC) using aluminum sheet silica gel 60 F254 (Merck) plates. Compounds were purified by column chromatography and/or RP-HPLC as stated. Unless otherwise noted, mixtures of ethyl acetate and hexane were used as the eluent for column chromatography. ¹H, D₂O exchange, ¹³C{1H}, ¹³⁵DEPT, and 2D NMR spectra were acquired on a Bruker Avance III HD 500 MHz NMR instrument. Chemical shifts are reported in ppm (δ scale), and coupling constant (J) values are reported in hertz (Hz). Data are represented as singlet (s), broad singlet (brs), doublet (d), doublet of doublet (dd), broad doublet (brd) and multiplet (m). Structural assignments were made with additional information from gCOSY, gHSQC, and gHMBC experiments.

Oligonucleotides and the dimeric core structures were synthesized on a Applied Biosystems 394 DNA synthesizer and were characterized by HPLC-MS (Table S1 and Figure S2). All conjugation reactions were performed in 20 mM Tris·HCl, 200 mM NaCl, pH 7.5 buffer at 25 °C. Reaction progress for oligonucleotide work was monitored by HPLC-MS spectroscopy.

Synthesis of 5-Norbornene-2-methanol Phosphoramidite

5-Norbornene-2-methanol phosphoramidite was synthesized using a previously reported method²⁸ (Scheme S2). Briefly, DIPEA (3 equiv) was added to a stirred mixture of 5-norbornene-2-methanol (1 equiv) in DCM. 2-Cyanoethyl N,N-diisopropylchlorophosphoramidite (1.5 equiv) was added slowly to the reaction mixture and stirred for 45 min at rt. The reaction mixture was diluted with 50 mL of DCM, and the organic layer was washed with aqueous saturated NaHCO₃ (25 mL × 2), followed by saturated NaCl solution (25 mL × 1) and dried over Na₂SO₄. The organic layer was then evaporated to dryness and purified by column chromatography using ethyl acetate and hexane as the eluent. The final purified norbornene-methanol phosphoramidite was characterized by ³¹P NMR (151.9 MHz, CDCl₃) δ 147.47, 147.34, 147.32, 147.09 ppm.

Postsynthetic Oligonucleotide Conjugation to Methyltetrazine NHS Ester

Tetrazine conjugation to oligonucleotides was carried out postsynthetically as previously reported,²⁸ via coupling between 5′-amino-modified ONs and methyltetrazine NHS ester (Scheme S1). Briefly, methyltetrazine NHS ester dissolved in DMSO (35 equiv) was added to a 1.5–2.0 mM ON solution in 200 mM HEPES buffer (pH 8.3), maintaining the ratio of 200 mM HEPES buffer (pH 8.3) and DMSO as 1:1 to ensure optimal solubility of the tetrazine substrate. The mixture was incubated at 40 °C for 48 h. Then, the reaction mixture was desalted using a Glen Pak desalting column to remove excess unreacted methyltetrazine NHS ester. Further, RP-HPLC purification was done using a C18 semipreparative column with a flow rate of 2.5 mL/min using a linear gradient of 5–60% B (100% acetonitrile). TEAA (0.1 M) was used as buffer A. Reaction success was confirmed by HPLC-MS (Table S1).

Synthesis of Multivalent Oligonucleotide Conjugates

The dimeric norbornene core (Heg-Nor2, 100 μM) and tetrazine oligonucleotide (200, 400, and 800 μM) were combined in a conical tube. The reaction mixture was dried with a vacuum centrifuge. The dried residue was redissolved in reaction buffer (20 mM Tris·HCl, 200 mM NaCl, pH 7.4), and the reaction mixture was left at room temperature for the desired period of time. Conjugated oligonucleotides were analyzed by HPLC-MS at specific time intervals as mentioned in the main text.

Synthesis of 12

5-Norbornene-2,2-dimethanol (100 mg, 6.48 × 10^–4 mol, 1.0 equiv) and 3-methyl-6-phenyl-1,2,4,5-tetrazine (224 mg, 1.30 × 10^–3 mol, 2.0 equiv) were dissolved in N,N-dimethylformamide (300 μL) and stirred overnight at rt. The reaction mixture was extracted with ethyl acetate (10 mL × 3) and washed with water (×2) and saturated NaCl (×1). The excess tetrazine and other impurities were removed by silica gel chromatography using ethyl acetate and hexane as the eluent. All dimeric isomers, which were inseparable with this eluting system, were collected as a mixture (colorless gum, 118 mg, 41% (based upon a molecular weight of 442.2)). This mixture of dimeric isomers of 12 was then purified by multiple injections on a preparative HPLC (Agilent 1260 Infinity HPLC, Agilent PLRP-S C18, 8 μm, 300 Å LC column (150 mm × 25 mm)) by eluting over 29 min using an isocratic gradient of 3–16% acetonitrile (0–2 min) and then 16–43% acetonitrile (2–29 min) in water at 40 °C, using a flow rate of 15 mL/min. Absorbance was recorded at 260 nm, and peaks were fractionated and collected. Peaks were analyzed by electrospray ionization–mass spectrometry (ESI–MS) to determine their molecular weight using an Agilent 6130 Quadrupole LC–MS coupled with an Agilent 1100 HPLC. Samples were eluted with 100% methanol containing 0.2% formic acid (v/v) at a flow rate of 0.5 mL/min. Multiple peaks were found to contain species with a molecular weight matching that of a dimeric compound (see Figure 2).

ESI–MS confirmed that the major peak from HPLC purification contained a dimeric compound (MW 442.2). Fractions of this peak were pooled and after lyophilization gave product 12 as a white fluffy solid (44 mg, 15%). The major peak collected after preparative HPLC was lyophilized to dryness and resuspended in the deuterated solvent for detailed NMR analysis (see Supporting Methods and Data). ¹H NMR (CD₃CN, 500 MHz) δ 7.50–7.45 (m, 5H), 7.29–7.28 (m, 3H), 7.20 (s, 1H), 7.17–7.15 (m, 2H), 5.30 (s, 2H), 3.89–3.87 (d, 1H), 3.41–3.39 (d, 2H), 3.33–3.31 (m, 2H), 3.17 (brs, 1H) 2.98–2.97 (d, 1H), 2.82 (brs, 1H), 2.78 (brd, 1H), 2.54 (s, 3H), 2.17 (s, 1H), 1.79–1.76 (brd, 1H), 1.58–1.55 (brd, 1H), 1.24–1.20 (dd, 1H), 0.99–0.96 (dd, 1H) ppm. ¹³C{¹H} NMR (CD₃CN, 125 MHz) δ 161.7, 158.3, 150.4, 141.6, 141.1, 139.1, 130.7, 129.4, 129.2, 129.1, 129.6, 128.2, 126.66, 67.8, 66.0, 49.0 48.7, 48.0, 42.8, 40.3, 38.2, 35.7, 22.0 ppm. HRMS (ESI) m/z calcd. for C₂₇H₃₀N₄O₂ [M + H]⁺ 443.2442, found: 443.2441.

Additional methods are contained in the Supporting Information.

Data Availability Statement

The data underlying this study are available in the published article and its Supporting Information.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.joc.2c02861.

• Additional experimental details; NMR (¹H, deuterium exchange, 13C, ¹³⁵DEPT, HSQC, HMBC), MS, and LC–MS spectra; and DFT calculations (methods, computed energies, cartesian coordinates) (PDF)

The authors declare no competing financial interest.