Synthesis of aminonaphthoquinone derivatives with Ceric Ammonium Nitrate (CAN) as a catalyst: NMR characterization and in silico reaction mechanism

Three different aminonaphthoquinones of great interest in medicinal chemistry due to their diverse biological activities were more efficiently synthesized and characterized starting from naphthoquinones with hydrazoic acid in the presence of ceric ammonium nitrate (CAN). We have previously reported a highly time demand synthesis of 2amino-1,4-naphthoquinone and 2-amino-3-methyl-1,4-naphthoquinone in the absence of the CAN catalyst. In the current study, we have also obtained 3-amino-5-hydroxy2 1,4-naphthoquinone and reduced all reaction times in the presence of CAN as a catalyst. Reaction rates have been increased to circa three times their original times. All aminonaphthoquinones have been characterized by NMR, vibrational, and chromatographic techniques. Additionally, we have proposed a reaction mechanism for the amination of naphthoquinone derivatives in an acid medium, based on in-depth DFT calculations.


Introduction
Naphthoquinones are aromatic compounds very abundant in nature, distributed in higher plants, bacteria, fungi, arthropods, and individual representatives of the animal kingdom [1 -3]. Aminonaphthoquinones are of great interest in medicinal chemistry due to the variety of biological activities they present as cytotoxic, antimalarial, antifungal, antibacterial, anti-inflammatory, and antitumor agents [4][5][6][7][8][9][10]. Since naphthoquinones mainly present nucleophilic addition and substitution reactions, their amination reactions occur in an acid medium where the oxygen of a ketone group of the molecule acts as a Lewis base causing the delocalization of π electrons present in the quinone nucleus, where nucleophilic agents like amino groups can attack [11] (Scheme 1).
The amination of 1,4-naphthoquinone derivatives is an essential step in the synthetic pathway of more complex derivatives such as indolequinones [12][13][14][15]. The synthesis of aminonaphthoquinones is possible by reacting 1,4-naphthoquinone derivatives with nucleophiles agents such as hydroxylamines, gaseous ammonia, sodium nitrite with subsequent reduction, primary amines, and hydrazoic acid [11]. The reaction between quinone derivatives and hydrazoic acid have been widely studied in terms of reaction conditions and reaction mechanism [16,17]. However, concerning to aminonaphthoquinones, the evaluation of more efficient routes of synthesis, as well as an in-depth study of the formation mechanism, to the best of our knowledge has not been reported.
Additionally, we carried out an in silico study of the more favorable reaction mechanism of the amination of naphthoquinone derivatives in acidic medium. With the here reported results we want to contribute to the study of the amination reactions of 4 quinone compounds that lead to more efficient reaction conditions in energy consumption and reaction time.

Chemistry
In this study, all amination reactions were carried out in ethanol/water (5:3, v/v) at pH 4. Hydrazoic acid was prepared in situ, maintaining a 1:6 molar ratio of starting material/azide of sodium. The reaction temperature was established by starting the experiments at room temperature and 50 ºC if no reaction was observed (Scheme 2).
The reaction time was determined by monitoring with thin layer chromatography. Initially, to observe the effect of the amount of catalyst in the amination reaction, the synthesis of 4 was carried out by varying the molar proportions of CAN. Table 1 shows the proportions of CAN used and the yields in each reaction (I-IV). As a result, an inversely proportional relationship between the amount of catalyst and the reaction yield is evident. Cerium has been proposed as a reaction mechanism to act as a Lewis base, binding to one of the carbonyl oxygens of naphthoquinones, triggering a delocalization of electrons that would activate the quinone ring for a nucleophilic attack [19,20]. It is likely that an excess of catalyst in the reaction acts as an oxidizing agent and not as a Lewis base, causing the appearance of unwanted products and a lower yield. This hypothesis is supported by observing the chromatographic profiles made after each reaction, where higher proportions of catalyst generate more complex profiles with higher numbers of spots ( Figure S1). Table 2 shows a comparison of the reaction times and temperatures and the reaction yield in synthesizing aminonaphthoquinones derivatives with/without a catalyst. All amination reactions with CAN showed a reduction in reaction times versus reactions without the catalyst. The coordination bond between the carbonyl oxygen and cerium (IV) may be more stable than that established with the proton of the acidic medium, generating a more favorable chemical balance for the nucleophilic attack of the azide, increasing the kinetics of the reaction. On the other hand, it was possible to obtain compounds 4 and 6 at room temperature, indicating a decrease in activation energy when using CAN in the amination reactions.
For compound 5, no change in reaction temperature is observed when CAN is used in the reaction; there is no decrease in the activation energy of this reaction due to the steric effect exerted by the methyl group near the azide attack site. Finally, in the amination reaction of 3, we expected to obtain a mixture of two isomers, 6 and 7, but it was only possible to identify compound 6. The amino group position in compound 6 was assigned by analyzing the HSQC and HMBC spectra ( Figure 1). Seemingly the hydrogen bond formed between the hydroxyl group in the C5 position and the C4 carbonyl group limits the interaction of cerium (IV) with the latter, increasing the probability of union with the oxygen of the C1 carbonyl that is free, driving the azide attack at position C3 (Scheme 3). This regioselective nucleophilic attack at the C3 position of compound 3 is observed when amination is carried out with hydroxylamine, obtaining both isomers with a higher yield of compound 6 [21]. On the contrary, when the reaction is accomplished by direct amination with ammonia in methanol, compound 7 is the only product. However, If a methoxy group replaces the hydroxyl at C5, the amination occurs at C3 [22].

Reaction energies and spontaneity
Geometry optimization calculations have been performed on the naphthoquinones (compounds 1-3 in Figure 2), as well as on the aminonaphthoquinones of the products (compounds 4-6 in Figure 2), in order to estimates the reaction energies by considering ∆E=Eprod -Ereac. In this expression, Ereac and Eprod correspond to the sum of the absolute energies of reactants and products. It is worth mentioning that the CAN catalyst effect has not been explored because its action has been demonstrated to be 8 mostly kinetic. Compounds 1-6 have a singlet ground state with an electronic structure characterized by frontier molecular orbitals of π nature mainly located on the naphthalene fragment. The calculated reaction energies (∆E), including enthalpies (∆H) and Gibbs free energies (∆G), are reported in Table 3. The great amplitude and negative sign of the energies (∆E < -104 kcal·mol -1 ) indicate the considerable stability of products over to reactants. Similarly, the great exothermicity and spontaneity of the reaction is also confirmed by ∆H < -106 kcal·mol -1 and ∆G < -105 kcal·mol -1 . These values reveal that the overall amination reaction is thermodynamically irreversible and is in line with a high reaction yield, as observed in the experiments.  We have also carried out analytical frequency calculations to perform the theoretical IR spectra of the reactants 1-3 and products 4-6. It is worth noting that no negative frequencies were found on the above optimized geometries, indicating that each structure is a global minimum. Table 4 shows the correspondence between some theoretical and experimental characteristic IR bands of the products. In general, the bands are well matched, which supports the prediction ability of the computational method. However, it can be seen that the lower energy, the better match of the bands.
The IR spectra of compounds 4-6 are comparable to each other, as shown in Figure   S2.

Activation energies and reaction profile
We have also explored different pathways to establish the reaction mechanism of the amination process of compounds 1-3. Geometry optimizations of the stationary points within each reaction have been performed so that, according to reported mechanisms of similar quinones [17], the minimum energy path able to connect 1-3 with 4-6 would be assessed, respectively. The whole reaction can be understood as a nucleophilic substitution; however, the solved mechanism concern series of reaction steps of nucleophilic addition and elimination that lead to the formation of different reaction intermediates.
Let us explain the reaction mechanism of compounds 1-3 as follows: the first stage of the reaction consists of nucleophilic addition of the azido anion that results in the formation of the keto form 2-azidohydroquinone (Int1 in Scheme 4). Some experimental data suggest that this step is the slower, and hence rate limiting step of the reaction [17]. Accordingly, we have searched the corresponding Transition State   In Figure 3 are depicted the reaction profile of amination of compounds 1-3. Except for Int1, all intermediates are strongly stable with respect to reactants (the energy differences are larger than 21 kcal·mol -1 ), in agreement with a highly favorable reaction and the fact that the nucleophilic addition of N3is the rate determining step of the reaction. Besides, the formation of six intermediates is in line with the extensive time demand of the reaction in the absence of a catalyst.

Conclusion
In The mass spectra were obtained from an Agilent HP 6890 gas chromatograph equipped with an HP 5973 selective mass detector (Agilent Technologies, Santa Clara, CA, USA). TLC, MS, NMR, and HPLC analytical data confirmed that the purity of test compounds was ≥96%.

Synthesis of 2-amino-1,4-naphthoquinone (4).
To a solution of 1 g of 1 in 50 mL of ethanol was added 2.37 g of sodium azide in 30 mL of distilled water, then was added 3.5 g of CAN, the pH adjusted to 4 with HCl and allowed to stir for 6 h at room temperature. After completing the reaction time, extraction was performed with ethyl acetate (3 X 50 mL). The organic extracts were washed with saline solution, dried over anhydrous Na2SO4, and concentrated. The residue was crystallized from methanol to afford an orange needles; 47.6 % yield; m.p.
The geometry optimization of compounds 1-6 was performed at the B3LYP level of theory [24] with the Ahlrichs def2-TZVP basis set (the resolution of identity approach has been used with the def2/J and def2-TZVP/C auxiliary basis sets for Coulomb and correlation integral calculations, respectively) [25][26][27], as implemented in the ORCA 4.2.0 package [28,29]. It has also been included the dispersion correction developed by Grimme (included in ORCA by the D3BJ approximation) and the implicit solvent effects by the Conductor-like Polarizable Continuum Model (CPCM) [30]. The dielectric constants and refractive index used for ethanol as a solvent is 24.30 and 1.361, respectively. The threshold for the energy convergence in the self-consistent field procedure was 1x10 -8 a.u. No negative normal modes were obtained by analytical frequency calculations on the optimized geometries. The orca_mapspc module [31] and Avogadro visualization tools have been used for plotting IR spectra and geometries, respectively [32].

Supporting Information
Supporting information features a chromatogram photo of the reaction product of compounds 4, comparative of theoretical and experimental IR spectra of compounds 4-6, copies of 1 H and 13 C NMR spectra of compounds 4-6, HMQC and HMBC spectra of compound 6,