A CuAAC–Hydrazone–CuAAC Trifunctional Scaffold for the Solid-Phase Synthesis of Trimodal Compounds: Possibilities and Limitations

We present a trifunctional scaffold designed for the solid-phase synthesis of trimodal compounds. This scaffold holds two alkyne arms in a free and TIPS-protected form for consecutive CuAAC (copper(I)-catalyzed azide–alkyne cycloaddition), one Fmoc-protected hydrazide arm for reaction with aldehydes, and one carboxylic acid arm with CF2 groups for attachment to the resin and 19F-NMR quantification. This scaffold was attached to a resin and derivatized with model azides and aliphatic, electron-rich or electron-poor aromatic aldehydes. We identified several limitations of the scaffold caused by the instability of hydrazones in acidic conditions, in the presence of copper during CuAAC, and when copper accumulated in the resin. We successfully overcame these drawbacks by optimizing synthetic conditions for the derivatization of the scaffold with aromatic aldehydes. Overall, the new trifunctional scaffold combines CuAAC and hydrazone chemistries, offering a broader chemical space for the development of bioactive compounds.


3-Phenylbenzylazide 9
A suspension of 3-phenylbenzylbromide (5 g; 20.2 mmol) and NaN3 (2.6 g; 40.4 mmol) in DMSO (50 mL) was stirred overnight at 80 °C. After cooling to r.t., water (50 mL) was added and the resulting mixture was extracted with diethyl ether (4 × 50 mL). The organic layers were combined, washed with water (2 × 50 mL) and brine (2 × 50 mL), dried over anhydrous sodium sulfate and filtered. The diethyl ether was carefully evaporated under reduced pressure (azide is partially volatile and could be explosive) to give bright brown crude product, which was purified by flash chromatography on silica gel using a linear gradient of diethyl ether in petroleum ether to give 9 (3.8 g, 90%, colorless liquid).

Copper-Promoted Hydrolysis of Hydrazide in Solution
We investigated the ability of copper ions to promote the apparent hydrolysis of substituted hydrazides in solution using the model N′-butylbenzohydrazide (14). The reaction conditions (R1-R3) are shown in Scheme S2. Three different aqueous solutions of substituted hydrazide 14 (10 μM) without copper (R1), with Cu(II) ions (R2) and with Cu(I) ions (after the reduction with sodium ascorbate, R3) were prepared (Scheme S2). The hydrolysis of compound 14 was followed by HPLC; an aliquot of each solution (10 μL) was taken (at 30 min, 60 min, 210 min and 360 min) and directly injected into HPLC. The hydrolysis of compound 14 is characterized by the appearance of the peak of benzoic acid at 19.272 min ( Figure S1). The peaks of benzoic acid and compound 14 were identified by the injections of benzoic acid and compound 14 alone. The peak at 11.576 min could not be identified but may correspond to a complex formed by compound 14 and copper ions. In Figure S1, the HPLC chromatogram of the solution containing Cu(II) ions (CuSO4·5H2O alone, R2) after 30 min of reaction is shown. The rate of appearance of benzoic acid in the solutions containing copper ions is also shown in Figure S1.
Compound 14 in the solution without Cu ions (R1) remained unchanged after 6 h of the reaction. In contrary, the formation of benzoic acid is observed after 30 min of reaction in the solution with Cu(II) ions (R2). Thus, the presence of Cu(II) indeed promotes the hydrolysis of compound 14 in aqueous media. In the case of the solution with Cu(I) ions (CuSO4·5H2O and sodium ascorbate, R3), the hydrolysis is much slower (11% of benzoic acid after 6 h at r.t.) and may be attributed to the minor concentration of Cu(II) ions.

Condensation of Benzhydrazide and 3-(Trifluoromethyl)benzaldehyde in the Presence of Copper(II) Ions
We investigated if copper(II) ions interfere with the condensation between benzhydrazide and 3-(trifluoromethyl)benzaldehyde. The reaction conditions (R1-R3) are shown in Scheme S3.
The condensation was followed by HPLC; an aliquot of each solution (5 μL) was taken after 210 min of reaction (time necessary for the completion of the reaction without copper) and directly injected into HPLC. As CuSO4·5H2O has a limited solubility in ACN, 18-crown-6 ether was added in a third reaction (condition R3). The condensation is characterized by the appearance of the peak of 15 at 23.874 min and the disappearance of the peak of benzhydrazide at 7.896 min ( Figure S2). As shown in Table S1, the presence of one equivalent of copper has a negligible impact on the reaction advancement. Figure S2. RP-HPLC chromatogram of the reaction R3 after 90 min. The green line reflects the composition of a gradient (% of solvent B). RP-HPLC were carried out using the same system, analytical column and eluents as described in the main manuscript. Table S1. Advancement of the reactions R1, R2 and R3 after 90 min. The completion of the reaction is calculated by integration of the signals of 15 and benzaldehyde at 218 nm.
The reactions were followed by HPLC; after 1h of reaction, an aliquot of each solution (10 μL) was directly injected into HPLC. The proportions of compounds 12, 13 and 16 in the starting crude 13 and after 1 h of reaction (condition R1 or R2) are given in Table S2. Table S2. Proportion of 12, 13 and 16 in the starting crude 13 and in reaction mixtures R1 and R2 (after 1 h). The proportions were calculated by integration of signals (peaks) in HPLC chromatograms at 218 nm. The real proportion of 16 in the crude compound 13 is lower than observed by HPLC as hydrolysis of 13 occurred during the HPLC analysis (16 was not detected in the mass spectrum of the crude compound 13). From these results we concluded that the presence of copper ions involves an additional hydrolysis pathway to the classical [3,4] hydrazone hydrolysis. Based on the literature [5][6][7] we proposed the pathway for the formation of 12 shown in Scheme S5: after hydrolysis of the hydrazone moiety in aqueous acidic medium, the copper-promoted oxidation of the hydrazide 16 would lead to the benzoic acid 12 (after the addition of a molecule of water).

S7
Scheme S5. Proposed pathway for the hydrolysis of 13 in the presence of copper(II) ions.

Discussion on the Step Order of the Click Reactions
We determined the optimal step order for the two first click reactions (CuAAC and hydrazone formation) on solid-phase. The two possible pathways are shown in Scheme S6.

S8
In the pathway proposed in the main manuscript (steps a then b) the CuAAC went to completion after 16 h. In contrast, when the acylhydrazone formation was performed first (step b then step a), the CuAAC was not complete after two treatments of 16 h each: RP-HPLC analysis showed 27% of starting material 17 remaining after the second treatment ( Figure S3).
To explain these results, we speculate that the presence of the Fmoc group in the resin-bound compound 6s impacts the resin properties (lipophilicity, three-dimensional arrangement) and helps the diffusion of the bulky and lipophilic azide 9 into the resin.       (Table S2), the hydrazide 16 probably results from the hydrolysis of acylhydrazone 13 in the HPLC buffer. S13 19 F-NMR Spectrum of Compound 8 Figure S11. 19 F-NMR spectrum of compound 8 with the internal standard BrCF2COOEt (in DMSO-d6). The two 19 F signals of the CF2 group result from a slow interconversion of conformers. This has already been proven in our previous work by temperature dependence of the 19 F-NMR spectra and the coalescence of signals at ~100 °C [8].