Xanthate Based Radical Cascade Toward Multicomponent Formation of Pyrrolopyrimidines

A short sequential synthesis of pyrrolidino- pyridines and pyrimidines illustrates the potential of combining Ugi-Smiles couplings with radical tin-free processes.


Introduction
Among multi-component reactions (MCRs), the Ugi reaction has a privileged position due to its broad synthetic significance [1,2]. Its four points of diversity combined with possible further transformations of the resulting adducts have became very popular in heterocyclic synthesis. Indeed, the Ugi reaction has been associated with a wide range of post-condensation reactions in order to form even more complex scaffolds, including cycloadditions [3][4][5][6][7][8][9], metal-catalyzed processes [10][11][12][13][14][15] and cyclocondensations [16][17][18][19][20][21]. Among this very impressive bibliography, the combination of Ugi or Passerini reactions with radical chemistry has been scarcely documented [22][23][24][25][26][27]. Among the various radical systems explored in our group, we have mostly studied the potential of xanthate radical transfer resulting in various cyclizations of Ugi xanthate adducts bearing suitable alkene [22], alkyne [25] or aryl [26] moieties. In all these examples, α-chloroacetic acid was used as the acidic partner to introduce the xanthate moiety via a nucleophilic substitution. However, these intramolecular post-condensations did not exploit the full potential of the xanthate radical transfer. Indeed, compared OPEN ACCESS to tin hydride chemistry, the reversible nature of the addition of radicals onto the thiocarbonyl group is associated with high yielding intermolecular couplings between xanthates and alkenes [28,29]. Herein, we propose to explore this feature and to further increase the molecular diversity of the Ugi step through a xanthate triggered addition/cyclization cascade to form new heterocyclic systems (Scheme 1). Scheme 1. Synthetic strategy towards pyrrolidino fused systems.

Results and Discussion
Developing the radical chemistry of xanthates, the Zard group has reported various synthetic approaches towards fused cyclic systems such as indanes, indolines or azaindolines [30][31][32][33][34][35]. These syntheses feature an intermolecular addition of a xanthate on an alkene followed by an intramolecular trapping with a suitably positioned aryl group. As part of our ongoing interest in Ugi-type reactions, we naturally imagined to synthesize the radical acceptor through a Ugi-Smiles coupling. Indeed, we recently developed the use of electron-deficient phenols as acid surrogates in Ugi reactions to form N-arylamino carboxamides [36,37]. The use of a partner substituted with an alkene, such as allylamine, in this reaction would afford in one step an interesting substrate to test the tandem radical process (Scheme 1). Concerning the phenol, most reported Ugi-Smiles couplings involve a nitro group to perform the Smiles rearrangement. As nitroaryl derivatives are often associated with inhibition of radical chain processes, we prefer the use of hydroxy heterocycles (pyridines, pyrimidines) which also led to efficient multicomponent couplings [38]. N-allylaminopyridines and N-allylaminopyrimidines are known to react with xanthate under radical conditions, however the scope of these reactions seems to be limited by the nucleophilic behavior of the cyclic nitrogen atom. Indeed all reported radical tandem were performed on 2-or 5-halosubstituted heterocycles (chloro and fluoro) in order to lower the reactivity at the nitrogen center (Scheme 2) [31][32][33][34][35].

Four-Component Coupling Towards Radical Substrates
Various Ugi-Smiles adducts were thus prepared starting with allylamine as the amine partner. The latter was coupled with different hydroxy-pyridines and pyrimidines. The MCR coupling was performed under classical conditions using methanol as solvent. The starting radical substrates were obtained in modest to good yields. Table 1. Preparation of the radical substrates.

Radical Cascade
In previous studies on xanthate additions on N-allylamino-pyridines and pyrimidines, a first 1,2-addition product on the alkene could be obtained and isolated under heating with substoichiometric amounts of a radical initiator such as dilauroyl peroxide (DLP) in refluxing 1,2-dichloroethane (DCE) or ethyl acetate. Then the intermediate adducts could cyclize under treatment with further DLP or di-tert-butyl peroxide (more than 1 equivalent) at higher temperature allowing cleaner reactions (Scheme 3).
Even if the yields are expected to be lower, we preferred to perform the cascade under a single set of experimental conditions in order to keep the multicomponent sequence as short as possible. When adduct 1a was refluxed in DCE (0.2 M) with DLP added in 15 mol% portions every twenty minutes, we could isolate the desired cyclized pyrrolidine 2a as a 9:1 mixture of diastereomers [39] in 48% isolated yield, after overall addition of 1.5 equivalent of DLP (Scheme 3).

Scheme 3. Radical reaction conditions.
This sequence is initiated by decomposition of the xanthate into the electrophilic radical I, which is then trapped by the alkene moiety to give intermediate II. The latter is an electron-rich radical that can react with the starting xanthate to form intermediate IV. The latter being slightly less reactive than the starting xanthate towards radicals issued from DLP, accumulates in the medium and reacts when further amount of DLP is added without much 1 left to enter in the chain process. After addition of the radical onto the aryl ring, the intermediate species III further aromatize through interaction with the DLP (Scheme 4).

Scheme 4. Plausible mechanism.
To evaluate the scope of this cyclization, the reaction was performed with the Ugi-Smiles adducts formerly prepared and the results are displayed in Table 2. Pyridines and pyrimidines without any halogen substitutent at the 2 or 5 position can act as radical trapping agents in this sequence. In the study disclosed by Zard, initial attempts with 2-aminopyridine or 2-amino-5-methylpyridine as starting materials led to successful 1,2-addition product with the xanthate, but no indolizine could be formed in the following cyclization step [31]. Chloro or fluoro substituents are expected to lower the electron density in the aromatic ring leading to easier addition of the nucleophilic intermediate radical species, and this may explain the moderate yields observed. In our case, the reaction turns out to be poorly efficient with 4-aminopyridines, with yields not exceeding 30% ( Table 2, Entries 1-2), a nitro group at the 5-position of the 2-aminopyridine ring activates the cyclization as a higher 48% isolated yield could be obtained ( Table 2, Entry 6). Pyrimidines, in agreement with their higher electrophilicity, are better partners, giving yields over 50% with simple alkyl or aryl susbtituents on the heterocyclic core ( Table 2, Entry 3).

General
1 H-NMR spectra were recorded on a 400 MHz spectrometer (Bruker Avance 400), using CDCl 3 solvent as reference and/or internal deuterium lock. 13 C-NMR spectra were recorded on a 100.6 MHz spectrometer (Bruker Avance 400). Two-dimensional NMR spectroscopy [ 1 H -1 H COSY spectra, 1 H -13 C COSY spectra (HSQC) and long-range 1 H -13 C COSY spectra (HMBC)], were carried out to determine the correlation between 1 H and 13 C. The chemical shifts for all NMR spectra are expressed in parts per million to high frequency of TMS reference. Coupling constants (J) are quoted in Hz and are recorded to the nearest 0.1 Hz. The IR spectra were obtained using ATR accessories on a Bruker Tensor 27 instrument. High-resolution (HR) mass spectra were recorded on a Hewlett-Packard HP 5989B spectrometer. TLC was carried out using precoated plates of silica gel 60F254 (SDS-Carlo-Erba).

Conclusions
The reported two-step cascade constitutes a concise entry into pyrrolidino-pyridines and pyrimidines. Compared with previous results obtained on xanthate cascades performed on N-allylamino-pyridine and pyrimidines, these new couplings significantly widen the scope of these reactions. The modest efficiency of the whole process is counter-balanced by the simple experimental procedure and the straightforward access to important biologically relevant scaffolds [40][41][42][43]. This work is a new example of the potential of radical chemistry in Ugi post-condensations.