Bioorthogonal Chemistry

One of the main approaches used in this project is the application of bioorthogonal chemistry to react modified ADP-ribose units with their fluorophore-containing partners, which were previously introduced by reporter-tagged NAD⁺ molecules into PAR chains. Thus, the concept of bioorthogonality and the most commonly used reactions are described in this chapter.

1.4.1 General Remarks

Any chemical reaction that can occur inside of living systems without interfering with native biological processes is considered a bioorthogonal reaction.^[74] That means that a reaction must proceed selectively between the reporter and its counterpart without the formation of side products and without reacting with other biological function present in its surrounding, such as free thiols. Moreover, the formed conjugation bond should be stable and chemically inert towards the attack of nearby biological nucleophiles. Due to the low concentration of the species of interest and possible quick turnover, these reactions need to have fast reaction kinetics that ideally do not rely on further additives. In addition, the reaction needs to proceed under physiological conditions (37 °C and pH 6-8) and all reaction partners and product should be active, stable and non-toxic under these conditions.^[75]

Huge progress has been made to develop and apply bioorthogonal chemistry for the study of biological processes. Thus, different biomolecules have been labelled with suitable reporter groups, such as amino acids, carbohydrates, nucleosides and nucleotides, as well as lipids, and other small molecules. Moreover, the spectrum and scope of bioorthogonal reactions is steadily increased, covered by numerous reviews.^[74-79]

In the following the most-commonly used reactions, namely copper(I)-catalysed azide-alkyne cycloadditions (CuAAC), strain promoted azide-azide-alkyne cycloadditions (SPAAC) and tetrazine ligations (DAinv) are introduced.

1.4.2 Cu^I-Catalysed Azide-Alkyne Cycloaddition

In 1963, Huisgen initially reported the 1,3-dipolar cycloaddition between azides and alkynes.^[80] In 2002, Sharpless^[81] and Meldal^[82] independently discovered, that the addition of catalytic amounts of copper(I) (Cu^I) salts results in a dramatic rate enhancement. These reactions proceed readily under physiological conditions and in a biological environment to provide 1,4-disubstituted triazoles. This reaction, widely known as ‘click reaction’, is the most popular bioorthogonal reaction and alkynes as well as azides can serve as bioorthogonal reporter groups due to their small size and easy synthetic accessibility (Figure 5A).^[76]

In the current model^[83] of the reaction mechanism, Cu^I forms first a π-complex with the triple bond of the alkyne and then a second Cu^I atom forms a σ-bound copper-acetylide under proton release. This intermediate now coordinates the azide and forms an unusual six-membered metallacycle. This ring is contracted to a triazolyl-copper-derivative and consecutive hydrolysis yields the regioselectively formed 1,4-disubstituted triazole (Figure 5B).

To further accelerate the reaction and to stabilise the +I oxidation state, several ligands such as water soluble tris(3-hydroxypropyltriazolyl methyl)-amine (THPTA) or chelate-assisting ligands such as 2-picolyl-azide have been applied (Figure 5C). Moreover, some protocols also use Cu^II salts that are reduced in situ by additives such as ascorbic acid.

Using these protocols, reaction rates with second-order rate constants of 10-200 M^-1·s^-1 in the presence of 20-500 µM Cu^I are achieved.^[78] The primary disadvantage of the CuAAC reaction is the cellular toxicity of the metal catalyst. It was found that Cu^I forms reactive oxygen species (ROS) responsible for damage and degradation of its biological environment.

Moreover, dehydroascorbate was found to be responsible for protein crosslinking and resulting in their precipitation.^[78] Although more bio-friendly metal/ligand combinations were discovered and individual optimisation of the respective reaction can reduce its harmfulness,^[84] the reaction is not ideal for labelling biomolecules in living cells.

Figure 5. Copper(I)-catalysed azide-alkyne cycloaddition (CuAAC). (A) CuAAC of either alkyne-tagged reporter and azide-functionalised dye or azide-tagged reporter and alkyne-functionalised dye. (B) Current model of CuAAC mechanism.^[83] (C) Structures of selected ligands and additives.

1.4.3 Strain Promoted Azide-Alkyne Cycloaddition

To circumvent the need of Cu^I and other additives, a new concept relying on rate enhancement of the azide-alkyne cycloaddition through ring-strain was introduced in 2004 by Bertozzi and co-workers and is now termed strain-promoted or copper-free click reaction (SPAAC).^[85] Here, the terminal alkyne is replaced by the cyclooctyne OCT, which releases around 18 kJ·mol^-1 of ring strain, while reacting with azides. However, computational analysis revealed that cyclooctynes resemble the transition state of the 1,3-dipolar cycloaddition more likely than terminal alkynes, which leads to a decrease in distortion energy. This distortion energy, rather than strain relief, is responsible for the fast coupling of these alkynes to azides.^[86-87]

During this reaction, both triazole regioisomers are formed within a concerted [4+2]

cycloaddition (Figure 6A and B). Since this discovery, a number of cyclooctyne derivatives with different improvements were reported and second order rate constants range between 10^-2 and 1 M^-1·s^-1 (Figure 6C).^[77] Rate enhancement was achieved by either attaching electron withdrawing groups at the propargylic position (MFCO, DIFO) or by further increasing the ring strain through aryl rings (DIBO, DIBAC, BARAC) or a cyclopropyl ring (BCN).^[78] The most recent improvement resulted in the currently fastest SPAAC reagent:

3,3,6,6-tetramethyl-thiacyloheptyne (TMTH). Here, the ring is contracted to a seven

Theoretical Background

membered ring with sulphur as heteroatom.^[88] Cycloheptynes with other substitutions and heteroatoms are too unstable to be isolated and must be prepared as a photoreleasable precursor, which forms in situ a cycloheptyne after irradiation that is useful as SPAAC reagents.^[89-90]

In general, an optimal balance between reactivity and stability must be obtained for applications in living systems, and unfortunately, the more reactive the cyclooctyne becomes the less stable it is. Cyclooctynes were found to undergo several side reactions such as homotrimerisation or reactions with intracellular thiols,^[91] and cysteine sulfenic acids were found to react even 100 times faster with cyclooctynes then azides (25M^-1·s^-1, Figure 6D).^[78]

As a conclusion, the azide/cyclooctyne pair should be thoroughly selected or optimised in accordance to its application. A recent report investigating intracellular staining applications thus suggests to use cyclooctyne as the reporter group and label it with fluorophore azides in order to circumvent unspecific signal.^[92] As cyclooctynes are much bigger than azides and less water-soluble, the success of this strategy depends on the system to be investigated.

N₃

Figure 6. Strain promoted azide-alkyne cycloaddition (SPAAC). (A) SPAAC of either azide-tagged reporter and cycloalkyne-functionalised dye or cycloalkyne-tagged reporter and azide-functionalised dye. (B) Concerted [4+2] cycloaddition. (C) Development of cycloalkynes for SPAAC with rate constants and years of discovery.^{[76, 90]} (D) Possible side reactions of involved cyclooctynes.^[78]

1.4.4 Tetrazine Ligation

The tetrazine ligation has been introduced concurrently by Fox et al.^[93], Hilderbrand et al.^[94]

and Wießler et al.^[95] in 2008. It is a Diels-Alder reaction with inverse electron demand (DAinv), where tetrazines react with different alkenes in a [4+2] cycloaddition to an intermediate, that immediately undergoes a Retro-Diels-Alder reaction under the release of nitrogen gas

(Figure 7A and B). If oxidants are around, the yielded dihydropyridazines can furthermore reobtain aromaticity forming a stable pyridazine cycloadduct.

Depending on the alkene applied in these reaction, very high rates ranging between 1 and 10⁴ M^-1·s^-1 can be observed.^[77] Common alkenes are norbornenes, trans-cyclooctenes and cyclopropenes (Figure 7C). But also smaller, not activated, terminal alkenes were found to react with tetrazines in a bioorthogonal manner, but with reduced reaction constants (10^-3 to 10^-1 M^-1·s^-1).^[96] Also the type of tetrazines applied for the reaction influences the reaction rate (Figure 7D). Important for Diels-Alder reaction is the overlap of involved orbitals, but also steric considerations, solubility and stability define the applicability of the respective tetrazines. As a general tendency, substituted tetrazines are more stable under physiological conditions, but are less reactive due to electronic and steric reasons.^[97]

However, tetrazines were also found to produce unwanted background staining due to hydrophobic binding and biological cross-reactivity and thus reaction kinetics should be thoroughly controlled.^{[78, 92]}

Figure 7. Tetrazine ligation (DAinv). (A) DAinv of either alkene-tagged reporter and tetrazine-functionalised dye or tetrazine-tagged reporter and alkene-functionalised dye. (B) Reaction mechanism consisting of [4+2] cycloaddition, retro-Diels-Alder reaction and oxidation and regioisomers. (C) Selected tetrazines, their rate constant with trans-cyclooctene and stability.^[97] (D) Selected alkenes and their rate constants with PyTz.^[96-99]

Theoretical Background