**4.2. Catalytic activities of synthetic G-quartets**

**Figure 3.** Chemical structures of the first TASQ developed by J. C. Sherman (**A**), [12] RAFT-G4 (**B**), [13] DOTASQ-C5 in equilibrium between its open (left) and closed conformations (right) (**C**), [14] PNADOTASQ (**D**), [15] and PyroTASQ (E).

In parallel, the group of E. Defrancq was focused on the functionalization of a cyclodecapeptide termed RAFT (for *regioselectivity addressable functionalized template*) by DNA strands.[127, 194– 196] This research led the team to synthesize a TASQ from a RAFT equipped with four guanosines.[13] Thanks to a collaboration with D. Monchaud's team, this molecule was studied as a DNAzyme (see section 4.2) and showed that the G-quartet intramolecular formation (confirmed by circular dichroism, and NMR studies) was able to catalyze the hemin oxidation/

However, the very first example of a water-soluble TASQ was proposed by L. Stefan *et al*. in 2011.[14] The key point was the choice of the *ad hoc* template, able to drive the G-quadruplex formation, and being fully soluble in water. It was why the cyclen macrocycle was chosen, for its high solubility, its ability of metal chelation, and its C4-type symmetry, identical to the one of a G-quartet.[198, 199] Even if this last criteria was not a *sine qua non* condition, it seemed to

[16]

reduction cyclic reaction.[197]

458 Advanced Catalytic Materials - Photocatalysis and Other Current Trends

The formation of a synthetic G-quartet was expected to have the same properties as a native one from a G-quadruplex. Indeed, this synthetic G-quartet is able to mimic the external (also termed accessible) G-quartet of the biological edifice.[12–16, 201] In this section, the main experimental data and results are highlighted to prove the feasibility of this new strategy of nature-mimicking catalysts. To begin with, the very first example of this approach, published by D. Monchaud's team, is presented and then the improvement with PNADOTASQ is ex‐ plained. A collaboration between this group and the E. Defrancq's one led to the study of the RAFT-G4, introduced in the previous section (see section 4.1), and detailed here.

In parallel, the work of H. O. Sintim's group, focused on intermolecular G-quartets (*i.e*., not TASQ) using c-di-GMP (for cyclic diguanylic acid) as a catalyst, will be presented.[141, 203, 204]

DOTASQs as a pre-catalyst: As explained before, the main step of the catalytic reaction is the activation of the hemin by interaction with the G-quartet. To verify this *sine qua non* condition, titrations of 1 μM hemin with DOTASQ were performed in cacodylic buffer.[205] The char‐ acteristic UV–Vis band of hemin from 350 nm to 400 nm was observed and, interestingly, the signal increased in all the cases with DOTASQ-C1, DOTASQ-C5, and 22AG, the telomeric Gquadruplex sequence d[AG3(T2AG3)3] used as a reference. Moreover, for the three systems, a stoichiometry of about 1:1 was found, and the dissociation constants were calculated using the following equation:

$$\mathbf{K}\_d = \frac{\left[\text{hemin}\,\,\text{[]}\,\text{cat.]}\right]}{\left[\text{hemin}\,\,+\,\text{cat.]}\right]} \tag{1}$$

Thus, the subsequent *K*<sup>d</sup> were found: 170 nM for DOTASQ-C1, 135 nM for DOTASQ-C5, and 235 nM for the native G-quadruplex 22AG. These results were duplicated with other concen‐ trations of hemin, and similar values were obtained. They show that both native and synthetic G-quartets are able to interact with hemin, forming the key step of the peroxidation reaction. Inspired by the optimal experimental conditions published by P. Travascio *et al*., and devel‐ oped for G-quadruplexes,[50] primitive experiments were carried out with 1 μM hemin, 2 mM ABTS, 600 μM H2O2, and from 0 μM to 50 μM DOTASQ, in Caco.KTD buffer. This buffer designated a cacodylic acid buffer composed of 10 mM lithium cacodylate (Caco), 10 mM KCl (letter K), 90 mM LiCl, with addition of 0.1 % (v/v) DMSO (letter D) favoring the solubilization of hemin, and 0.05 % (w/v) Triton X-100 (letter T), a nonionic surfactant promoting disaggre‐ gation of hemin.

**Figure 4.** Schematic representation of the peroxidase-like activity promoted by the use of a TASQ. [205, 206]

The catalytic activity was therefore evaluated by UV–Vis absorbance, in 1-mL-quartz cuvettes, measuring the formation of ABTS⋅+ (also proposed as ABTS⋅– in the scientific literature), the oxidized product of ABTS which absorbs at 420 nm.[80, 81] All the results were compared to a control experiment, strictly composed of all the same reagents, except the lack of the precatalyst (*i.e.*, the hemin is alone with ABTS and hydrogen peroxide, and not catalyzed by Gquadruplexes or TASQs). After less than 1 hour, all the absorbance signals were constant and proved the feasibility of the TASQ-catalyzed peroxidation reaction concept.

However, to confirm that the catalysis of hemin was due to the formation of the intramolecular G-quartet, a DOTASQ-C5 derivative, termed Prot.DOTASQ-C5, was synthesized. In fact, Prot.DOTASQ-C5 has "protected guanines" with 6-*O*-benzyl groups in the four guanine moieties. Thus, the formation of G-quartet is impossible because of the rupture of the H-bonds involving the carboxylic group in position 6. The catalytic activity of this compound is null, confirming that the formation of an intramolecular G-quartet inside the TASQ is mandatory to activate hemin and to catalyze the reaction.[205]

For the sake of comparison of the efficiency of the DOTASQ, apparent rate constants *k*cat were calculated, dividing the initial rate (*V*0) by the concentration of the catalyst ([cat.]), using Eq (2).

$$\mathbf{k}\_{\text{cat}} = \frac{\mathbf{V}\_0}{\boxed{\text{cat.}}} = \frac{\Delta \mathbf{A} \mathbf{b} \mathbf{s}\_i}{\mathcal{E}\_{\text{ABTS}^\*}} \mathbf{x} \frac{1}{\boxed{\text{cat.}}} \tag{2}$$

Thus, *k*cat of 0.36 h–1 and 0.29 h–1 were obtained for DOTASQ-C5 and DOTASQ-C1, respectively. For structural reasons, DOTASQ-C5 is able to form its intramolecular G-quartet easier than DOTASQ-C1. This better stability of the G-quartet favors the π-stacking of hemin and, consequently, the catalytic activity.

trations of hemin, and similar values were obtained. They show that both native and synthetic G-quartets are able to interact with hemin, forming the key step of the peroxidation reaction. Inspired by the optimal experimental conditions published by P. Travascio *et al*., and devel‐ oped for G-quadruplexes,[50] primitive experiments were carried out with 1 μM hemin, 2 mM ABTS, 600 μM H2O2, and from 0 μM to 50 μM DOTASQ, in Caco.KTD buffer. This buffer designated a cacodylic acid buffer composed of 10 mM lithium cacodylate (Caco), 10 mM KCl (letter K), 90 mM LiCl, with addition of 0.1 % (v/v) DMSO (letter D) favoring the solubilization of hemin, and 0.05 % (w/v) Triton X-100 (letter T), a nonionic surfactant promoting disaggre‐

460 Advanced Catalytic Materials - Photocatalysis and Other Current Trends

**Figure 4.** Schematic representation of the peroxidase-like activity promoted by the use of a TASQ. [205, 206]

proved the feasibility of the TASQ-catalyzed peroxidation reaction concept.

to activate hemin and to catalyze the reaction.[205]

*cat*

The catalytic activity was therefore evaluated by UV–Vis absorbance, in 1-mL-quartz cuvettes, measuring the formation of ABTS⋅+ (also proposed as ABTS⋅– in the scientific literature), the oxidized product of ABTS which absorbs at 420 nm.[80, 81] All the results were compared to a control experiment, strictly composed of all the same reagents, except the lack of the precatalyst (*i.e.*, the hemin is alone with ABTS and hydrogen peroxide, and not catalyzed by Gquadruplexes or TASQs). After less than 1 hour, all the absorbance signals were constant and

However, to confirm that the catalysis of hemin was due to the formation of the intramolecular G-quartet, a DOTASQ-C5 derivative, termed Prot.DOTASQ-C5, was synthesized. In fact, Prot.DOTASQ-C5 has "protected guanines" with 6-*O*-benzyl groups in the four guanine moieties. Thus, the formation of G-quartet is impossible because of the rupture of the H-bonds involving the carboxylic group in position 6. The catalytic activity of this compound is null, confirming that the formation of an intramolecular G-quartet inside the TASQ is mandatory

For the sake of comparison of the efficiency of the DOTASQ, apparent rate constants *k*cat were calculated, dividing the initial rate (*V*0) by the concentration of the catalyst ([cat.]), using Eq (2).

> *ABTS* e + <sup>D</sup> = = éù éù ëû ëû <sup>g</sup>

(2)

V Abs <sup>0</sup> <sup>1</sup> k x cat. cat. *i*

gation of hemin.

To compare TASQ with G-quadruplexes, a constant of *k*cat = 9.71 h–1 was got for 22AG, which is 25 times more efficient than DOTASQ-C5. This difference seems to be curious because the *K*d of DOTASQ and 22AG with hemin were close. A similar assessment was reported in the literature but with opposite results, with PS2.M and the RNA equivalent rPS2.M.[51] Indeed, their catalytic activities were similar, although the *K*<sup>d</sup> values were 27 nM and 900 nM, respec‐ tively.

It can be hypothesized that, even if the key step was defined as the hemin/G-quartet interaction, the environment around the G-quartet plays also a pivotal role for the catalytic efficiency. The presence of loops, the accessibility of the G-quartet to hemin, and other factors described before (see section 3.3) are critical for all the steps of the catalysis.[11, 56]

Finally, to eliminate all the doubts, another substrate was used instead of ABTS, and TMB was chosen. Its oxidation is a two-step reaction producing first a charge transfer complex, with absorbance at 370 nm and 652 nm, and a final diimine product absorbing at 450 nm.[82–84] Using different experimental conditions than before, the catalytic activity of 50 μM DOTASQ-C5 was evaluated in the presence of 1 μM hemin, 500 μM TMB, and 1.5 mM H2O2 in Caco.KTD buffer, and were compared with the same reaction with 2 μM 22AG. Like with ABTS, experi‐ ments with TMB approved the ability of TASQ to catalyze peroxidase-like catalysis. Obtained apparent catalytic constants were 0.02 h–1 for DOTASQ-C5 against 0.27 h–1 for 22AG.[205] Altogether, these results constitute the proof of concept of the use of native and synthetic Gquartets as a universal platform for the peroxidation reactions. However, efficiency had to be improved, and better results were obtained with the second generation of TASQ: the PNADOTASQ.

PNADOTASQ as a pre-catalyst: The main structural modifications made from DOTASQ to PNADOTASQ were the substitution of the original alkyl-arms by PNA guanine moieties. Thus, the new properties worn by this new TASQ are numerous and not detailed here (this molecule was also developed as a smart G-quadruplex ligand),[15] but the main point which must be highlighted here is the introduction of a total of four cationic charges, thanks to the presence of four pendant primary amine side chains. These positive charges (at physiological or acidic pH) are of the utmost importance to increase the interaction with hemin. Indeed, hemin is an Fe(III)–porphyrin holding two anionic charges due to two carboxylic groups. Electrostatic interactions were expected to facilitate the hemin approach and association with the G-quartet.

To verify this hypothesis, R. Haudecoeur *et al*. performed the catalytic reaction with TMB, using the same protocol described before, from 0 μM to 50 μM PNADOTASQ. For the sake of comparison, experiments with the same range of concentration were carried out with DO‐ TASQ-C5 (*i.e*., the most efficient DOTASQ) and also with 22AG at 2 μM as a reference.[206]

Interestingly, the results showed the far better catalytic ability of PNADOTASQ compared to the DOTASQ. Indeed, the initial rates were evaluated at 1.25 μM.min–1 and 0.05 μM.min–1 at 50 μM, respectively, corresponding to an average increase of efficiency of 2600 %. At this stage, PNADOTASQ is the more active TASQ catalyst for peroxidase-like reactions, and is closer to the activity of G-quadruplex, with a factor of 11 between it and 22AG (*i.e.*, in the experiments, 22 μM PNADOTASQ offers the same catalytic response than 2 μM of the natural G-quadruplex). PNADOTASQ confirms the role of intramolecular G-quartets to activate hemin and to perform DNAzyme-like catalysis, and also that improvement of the efficiency is just at the very beginning of the development. The modification of the arms is a positive point, but the role of the template, directly linked to the ability of synthetic G-quartet formation, is by consequence indirectly linked to the catalytic properties.

RAFT-G4 as a pre-catalyst: To evaluate the role of the template on the DNAzyme-like activity, the TASQ developed by E. Defrancq's team was tested.[197] Termed RAFT-G4, the molecule is composed of a cyclodecapeptide and four guanosine arms in which one triazole per arm is intercalated between the ribose and the peptide scaffold (playing a role in the stabilization of the intramolecular G-quartet).[13] As a first step, UV–Vis titrations of hemin by RAFT-G4 were performed to check the interaction, and compared with DOTASQ-C5. Interestingly, very similar signals shifted from the signal of hemin alone were obtained in the presence of both TASQ for the two peaks at 363 nm and 394 nm, corresponding to the aggregated and disag‐ gregated forms of the hemin (*i.e.*, catalytically active), respectively.[50]

Afterwards, catalytic experiments were carried out with 1 μM hemin, 2 mM ABTS, and 600 μM H2O2 in Caco.KTD buffer, with concentrations of DOTASQ-C5 and RAFT-G4 from 10 mM to 100 mM. Positively, the results revealed the ability of RAFT to be a catalyst, even if the efficiency was lower than for the DOTASQ-C5. To optimize the catalysis, the use of 10 mM ATP in Caco.KTD buffer at pH 4.8 (protocol based on previous studies with G-quadruplexes) was performed and showed the positive effect of these modifications of the experimental conditions. More precisely, this optimization led to an improvement of DOTASQ and RAFT-G4 efficiency by factors 1.8 and 5.1, respectively.[197]

These series of experiments triggered two main conclusions. First, the ability of a TASQ to be a good catalyst is dependent on the template, the nature of the guanine arms, and the stability of the G-quartet, and it seems to be for now difficult to rationalize their impacts. The multi‐ plication of examples of TASQ in the literature will be an invaluable chance to understand the role of each structural part of the molecules. Author wagers that new molecules like Pyro‐ TASQ[16] or NaphtoTASQ[200] will help to decipher a little more the reasons of these differences.

Another nice synthetic G-quartet system used as DNAzyme-like catalyst was developed by H. O. Sintim's team.[141] In this case, the molecule is not a TASQ forming an intramolecular G-quartet, but the cyclic diguanylic acid (also termed c-di-GMP), composed of two guanine residues. The authors proposed that c-di-GMP are able to form a discrete G-quadruplex formed by two intermolecular G-quartets (*i.e*., from four c-di-GMP molecules) at the micromolar level, but with the *sine qua non* presence of the intercalating proflavin molecule. In this case, it can be considered that proflavin assumes the role of a non-covalently linked template, unlike in TASQ. Experiments carried out with 0.5 μM hemin, 30 μM proflavin, 2 mM ABTS, and 2 mM H2O2 in Tris-HCl buffer (50 mM, pH 7.9) revealed that a catalytic activity can be distinguished for concentrations of c-di-GMP higher than 2 μM. Several control experiments were performed and clearly highlighted the fact that the formation of intermolecular G-quartets is required to observe catalytic activities.

All these data firmly confirm the role of synthetic G-quartets, both intramolecular and intermolecular, as efficient and promising nature-mimicking systems. In spite of the increasing excitation around this field, it is nevertheless honest to confess that the use of TASQ as catalysts for peroxidase-like experiments is, for the moment, far from the success of the DNAzyme, mainly in term of applications.[7, 23, 24] However, the very first ones were developed and are presented hereinafter.
