Functionalization of Oxazolo[4,5-b]pyrazines by Deprotometallation

: Different 2-arylated oxazolo[4,5- b ]pyrazines, obtained by palladium(II)-catalysed domino reaction from 2,3-dichloropyrazine and the corresponding carboxamides, were functionalized by deprotometallation. Employing lithium 2,2,6,6-tetramethylpiperidine, in order to form a lithio derivative, and then trapping it by iodolysis, proved to be inefficient. However, the presence of a zinc-based in situ trap allowed most substrates to be functionalized. Deprotonation of the pyrazine ring was observed in the presence of tolyl and anisyl groups at the oxazole 2-position. In contrast, with chlorophenyl and thienyl groups in this 2-position, deprotonation rather occurred on these groups either competitively or exclusively. The regioselectivities were discussed in the light of calculated p K a values of the substrates in THF. Finally, in the case of 2-phenyloxazolo[4,5-b ]pyrazine, we converted the mixture of 5- and 6-iodinated products into the corresponding 5,6-diiodide which was further functionalized by a double Suzuki coupling. Functionalizations of 2-arylated oxazolo[4,5- b ]pyrazines are reported. Access to the corresponding iodides was investigated, and deprotolithiation in the presence of a zinc-based in situ trap proved to be the most efficient approach. The observed regioselectivities were analysed in the light of the calculated p K a values. Triarylated oxazolo[4,5- b ]pyrazines were synthesized from a diiodo derivative obtained by iterative deprotonation-iodolysis. analyzed, and in particular their regioselectivity in the light of the calculated pKa values. The various halogenated ketones were next involved in copper-catalyzed two-fold C-N bond formation in order to obtain fused systems based on 2-aminopyrimidines. Besides potential antibacterial effect, reached 2-aminobenzothiopyrano[4,3,2-de]quinazoline proved to inhibit PIM1 (IC50: 0.61 μM) and CDK2/cyclin A (IC50: 2.0 μM) kinases.


Introduction
The heterocycle oxazolo [4,5-b]pyrazine was synthesized for the first time in 2014 by Demmer, Bunch and co-workers. [1] This planar heterocycle is structurally close to benzoxazole and oxazolopyridine, which are key scaffolds of both natural- [2] and synthetic products, [3] often endowed with attractive biological activities. [4] Thus, there is an interest in the medicinal chemistry field for methodologies which can give access to derivatives starting of 2-arylated oxazolo [4,5-b]pyrazines (Fig. 1). Among the methods used to functionalize aromatic compounds and heterocycles, deprotonative metallation is an efficient strategy. [5] Nevertheless, due to their low LUMO level, diazines are sensitive to nucleophilic attacks (e.g. to competitively give dimers), and thus hardly compatible with polar organometallics such as aryllithiums. [6] Consequently, in the absence of a substituent able to stabilize generated diazinyllithiums, by-products are generally observed. [7] Within the last two decades, methods have been developed to tackle the chemoselectivity issue of deprotometallation. Lithium-ate bases (e.g. zincates) and 'Turbo' bases (e.g. zinc amides containing LiCl) have been identified as suitable reagents to this date, [8] but few examples concern diazines without directing/acidifying groups. [9] In situ 'trans-metal trapping' is an alternative way to address this issue by intercepting the generated aryllithiums with an electrophilic trap as soon as it is formed. [10] For the functionalization of bare diazines, methods employing different base-in situ trap pairs have been reported. In this vein, the tandem LiTMP-Zn(TMP)2 (TMP = 2,2,6,6-tetramethylpiperidino) allowed Mongin and co-workers to deprotometallate pyrazine at room temperature in THF (THF = tetrahydrofuran), a result evidenced by subsequent iodolysis. [11] Mg(TMP)2·2LiCl-ZnCl2 also worked for the same purpose under similar conditions, [12] as well as LiTMP-Ga(CH2SiMe3)3 in hexane at room temperature. [13] In the frame of this paper, we report our efforts to functionalize the 2-arylated oxazolo [4,5-b]pyrazines 1a-12a by deprotonative metallation strategies.

Scheme 1. Formation of 2-phenyloxazolo
Substrates possessing numerous heteroatoms can be reluctant to LiTMP-mediated deprotometallation. [14] It proved to be the case from 1a since the reactions using 1 or 2 equiv of LiTMP in THF at -75 °C for 1 h showed after iodolysis low conversions to 1b and 1c (conversions below 25%; Table 1,  entries 1 and 2).
We thus turned to the use of LiTMP-Zn(TMP)2, generated in situ from ZnCl2·TMEDA (TMEDA = N,N,N',N'tetramethylethylenediamine) and LiTMP in a 1:3 ratio. [15] Surprisingly, almost no conversion was detected when using 0.5 equiv LiTMP-Zn(TMP)2 in THF at 0 °C for 2 h. The amount of base was therefore increased to 1 equiv LiTMP-Zn(TMP)2 (Table  1, entry 3). Unlike previously observed for diazines and their benzo analogues under similar conditions, [11] no dimer formation was noticed. Thus, the reaction temperature was raised to room temperature to afford a mixture of the 5-and 6-iodo derivatives 1b and 1c in 34% and 7% isolated yield, respectively (Table 1, entry 4). Nor were dimers formed at room temperature, but the presence of two diiodinated products increased (possessing a second iodo group on the phenyl group, at the ortho position).
We thus decided to attempt the use of LiTMP (1.5 equiv) in the presence of ZnCl2·TMEDA (1 equiv; more soluble than ZnCl2) as in situ trap, this pair being efficient to ensure accelerated and chemoselective reactions. [16] When employed at -30 °C, products from dimetallation were not observed, and the iodides 1b and 1c were isolated in 50% and 34% yield, respectively (Table 1,    We next studied the effect of a substituent on the phenyl ring by employing the 2-arylated oxazolo [4,5-b]pyrazines 2a-7a in the deprotometallation-iodolysis sequence. Under the conditions optimized from 1a, the monoiodides 2b-7b and 2c-7c were obtained with similar b/c ratios (Table 2, Fig. 3). No competitive reactions were observed on the aryl group.
However, product distribution became more complex when the reaction conditions were applied to the 2-(chlorophenyl)oxazolo [4,5-b]pyrazines 8a-10a (Table 3). From 8a and 9a, analysis of the crudes showed the presence of diiodides containing an additional halogen on the chlorophenyl ring (entries 1 and 2). The reactions were carried out in the presence of an in situ trap (ZnCl2·TMEDA), which can impede the isomerization of the initially formed lithio compound to a more stable one. However, such isomerizations are extremely fast, [17] and the experimental procedure favours thermodynamic control of anion formation (slow addition of base to the substrate). Thus, to try to understand this result, and why it differs from 10a (entry 3; Fig. 3), the pKa values in THF (Fig. 4) were calculated within DFT framework by using earlier elaborated approach. [18]

Accepted Manuscript
European Journal of Organic Chemistry There is no data on the acidity of oxazolo [4,5-b]pyrazines in literature. The most related experimental studies deal with equilibrium CH acidities of benzoxazole, benzofuran, benzothiazole, N-methylindole in THF, [19] and benzoxazole, benzothiazole in DMSO. [20] The deprotonation of 1a-12a can proceed at both the pyrazine ring or the aryl group. The by far most acidic position of bare oxazolo [4,5-b]pyrazine is blocked in 1a-12a with an aryl moiety. When generalizing our previous data on CH acidities of azoles and azines, [18,21] one should firmly expect to find the most acidic site in pyrazine ring in case of 1a-7a. But when turned to electron-withdrawing substituents (8a-12a), competitive pKa values could be found within them (Fig. 4) and deprotonation can proceed at the aryl group.
The results point to the aryl hydrogens of 8a and 9a being more acidic, when compared with those of 1a and 10a. We believe this explains the observed iodination of the phenyl ring.
The pKa calculations also help understand why mixtures of 5-and 6-iodinated oxazolo [4,5-b]pyrazines are systematically formed by deprotometallation. To understand why deprotonation at the pyrazine ring could not be avoided in favour of more acidified sites at the aryl ring (e.g. in the case of 9a), we decided to consider possible coordination of the pyrazine heteroatoms onto lithium (Fig. 5). Coordination of pyrazine nitrogens to metals presumably takes place in premetallation complexes and transition structures, and will increase the acidity of the neighbouring hydrogens considerably, making this proton abstraction competitive with that of activated aryl groups. In addition, whereas coordination of N7 to lithium (complex 1a·LiTMP-N7) seems to contribute more efficiency to acidification in favour of a deprotonation at C6, 1a·LiTMP-N4 was predicted to be slightly more stable in favour of a deprotonation at C5. In view of the low pKa values of the thienyl groups in 11a and 12a (Fig. 4), reactions are expected to occur at this fivemembered ring. Indeed, when 11a was submitted to LiTMP (1 equiv) in the presence of ZnCl2·TMEDA (1 equiv) in THF at -30 °C, subsequent iodolysis led to the 5-iodo derivative 11b; however, it could not be separated from the diiodide 11b' also formed in the reaction. An inseparable mixture of two monoiodides (12b and 12c) and one diiodide (12bc) was obtained when 12a was treated similarly (Scheme 2).
Because we could not regioselectively functionalize the arylsubstituted oxazolo[4,5-b]pyrazines, we considered the introduction of a second iodine from the mixture of 1b and 1c, generated by deprotometallation-iodolysis of 1a, in order to converge towards a single product. It has been shown in literature that iodopyrazine can be deprotolithiated at -78 °C upon treatment by LiTMP in THF. [22] Using ZnCl2·TMEDA as in situ trap, we could avoid products arising from diazyne formation, [23] and isolate the 5,6-diiodo derivative 1bc in 92% yield. Subsequent functionalization by double Suzuki coupling led to the triarylated derivatives 1d and 1d' (Scheme 3). Direct formation of 1bc by dideprotometallation of 1a using an excess of base [18b,21] did not work in this case. Scheme 2. Deprotometallation-iodolysis sequence from 11a and 12a. Scheme 3. Deprotometallation-iodolysis sequence from 1b-1c mixture and double Suzuki coupling from 1bc; ORTEP diagrams (50% probability) of 1bc.

Conclusions
Probably because they possess numerous heteroatoms, deprotometallation of oxazolo [4,5-b]pyrazines by classical hindered lithium amides are sluggish reactions. Recourse to in situ trapping was efficient in pushing the formation of arylmetals to completion, but a poor regioselectivity was noticed due to the high chemical similarity of the 5 and 6 pyrazine positions. It proved possible to convert the mixture of 5-and 6-iodo derivatives 1b and 1c into the corresponding 5,6-diiodide 1bc, thereby opening the door to subsequent functionalizations of this heterocycle.

Experimental Section
General synthetic and analytical details. All reactions were performed under an argon atmosphere. THF was distilled over sodiumbenzophenone. Column chromatography separations were achieved on silica gel (40-63 μm). Melting points were measured on a Kofler apparatus. IR spectra were collected on a Perkin-Elmer Spectrum 100 spectrometer. 1 H and 13 C Nuclear Magnetic Resonance (NMR) spectra were recorded on a Bruker Avance III spectrometer at 300 MHz and 75 MHz, respectively. 1 H chemical shifts (δ) are given in ppm relative to the solvent residual peak and 13 C chemical shifts are relative to the central peak of the solvent signal. [24] Chemical shifts are expressed in ppm. Microanalyses were performed on a Flash 1112 Thermo Fisher elemental analyser. Compounds 1a-12a were obtained as reported before. 1 ZnCl2·TMEDA was prepared as described previously. [15c] General crystallographic details. The X-ray diffraction data were collected either using an APEXII Bruker-AXS diffractometer (graphite monochromatized Mo-K radiation ( = 0.71073 Å)) for the compounds 1a, 6b and 9b', or using a D8 VENTURE Bruker AXS diffractometer (multilayer monochromatized Mo-K radiation ( = 0.71073 Å)) for 2a, 1b, 1c, 2b, 3b, 3c, 4b, 6c, 7c, 10c and 1bc. Except for 1c, for which the 10.1002/ejoc.201800481

Accepted Manuscript
European Journal of Organic Chemistry structure was solved by direct methods using SIR97, [25] all the structures were solved by dual-space algorithm using the SHELXT program. [26] Structural refinements were performed with full-matrix least-square methods based on F 2 (SHELXL-2014). [27] All non-hydrogen atoms were refined with anisotropic atomic displacement parameters and finally, H atoms were included in their calculated positions. The molecular diagrams were generated by ORTEP-3 (version 2.02). [28] Computational details. All calculations were performed within the DFT framework by using an approach elaborated earlier. [18b] Geometries were fully optimized at the B3LYP/6-31G(d) level of theory without any symmetry constraints implied. Vibrational frequencies were calculated at the same level of theory in order to characterize stationary points and to calculate zero-point vibrational energies (ZPVE) and thermal corrections. The single point energies were obtained using the B3LYP/6-311+G(d,p) level and tight convergence criteria. The solvent effects were treated using the polarized continuum model (PCM) with the default parameters for THF. The pKa values were obtained from the Gibbs energy of the homodesmic reaction between the studied (RH) and a probe (HetH) aromatic substrates: where N-methylindole with pKa(THF) = 38.1 [19] was used as the probe. [4,5-b]pyrazine (1a). Pd(OAc)2 (0.14 g, 5 mol%), XantPhos (0.71 g, 10 mol%), Cs2CO3 (9.6 g, 2.4 equiv) and benzamide (1.5 g, 12.4 mmol) were placed in a dried flask with a large magnetic stir bar. 1,4-Dioxane (40 mL, 0.3 M) and 2,3-dichloropyrazine (1.4 mL, 1.1 equiv) were added to the flask through a septum. The septum was replaced by a condenser flushed with argon, connected to a source of positive argon pressure. Under vigorous stirring, the reaction mixture was heated to reflux for 4 h. The reaction mixture was diluted with EtOAc (60 mL) and filtered, before evaporating the solvent. The crude was purified by flash column chromatography (heptane-EtOAc 3:1) to yield the pure compound (1.91 g, 81%). 1

Synthesis of the iodides from 1a-10a.
General procedure 1. A solution of LiTMP (prepared by adding BuLi (about 1.6 M hexanes solution, 1.5 mmol) to a stirred, cooled (0 °C) solution of 2,2,6,6-tetramethylpiperidine (0.25 mL, 1.5 mmol) in THF (3 mL) and stirring for 5 min) cooled at -30 °C was slowly transferred to a solution of the required heterocycle (1.0 mmol) and ZnCl2·TMEDA (0.25 g, 1.0 mmol) in THF (3 mL) at the same temperature. The resulting mixture was stirred for 15 min at -30 °C and then cooled to -70 °C. A solution of I2 (0.38 g, 1.5 mmol) in THF (5 mL) was introduced to the reaction mixture and it was stirred for 1.5 h at room temperature. The reaction was quenched with sat. aq. Na2S2O3 (10 mL) and extracted with EtOAc (3 x 10 mL). The combined organic layers were washed with brine (10 mL), dried over Na2SO4, and the solvent was removed under reduced pressure. The crude was purified by flash column chromatography on silica gel (eluent: heptane-Et2O 9:1) to yield the product.