Synthesis of pyrimido[1,6-a]quinoxalines via intermolecular trapping of thermally generated acyl(quinoxalin-2-yl)ketenes by Schiff bases

Acyl(quinoxalin-2-yl)ketenes generated by thermal decarbonylation of 3-acylpyrrolo[1,2-a]quinoxaline-1,2,4(5H)-triones react regioselectively with Schiff bases under solvent-free conditions to form pyrimido[1,6-a]quinoxaline derivatives in good yields.


Introduction
Quinoxaline is a 4-aza isostere of quinoline, which rarely occurs in structures of natural products. Its derivatives are gaining popularity in medicinal chemistry and pharmacology because many of them exhibit various biological activities [1,2].
To develop a new synthetic approach towards pyrimido[1,6-a]quinoxalines we looked through the procedures to their closest  analogues -pyrido[1,2-a]quinoxalines, the synthesis of which has been explored more frequently [3,4,. The analysis helped us to disclose a tempting but challenging methodology, which has the potential to be extended for the synthesis of the desired heterocyclic system, via intermolecular trapping of thermally generated acyl(quinoxalin-2-yl)ketenes [20,21,23,24,28,29,38] (Figure 2).
Syntheses utilizing acylketenes are of practical and theoretical interest due to the high reactivity of acylketenes and the structural diversity of the reaction products [47][48][49][50][51][52][53][54]. The introduction of the quinoxalin-2-yl substituent into acylketenes results in the formation of a peculiar system of conjugated double bonds, which can potentially act as either oxo-diene or aza-diene ( Figure 3). To the best of our knowledge, there is no example of the involvement of the aza-diene fragment of acyl(quinoxalin-2yl)ketenes into intermolecular trapping by hetero-dienophiles published so far. In this article we report a synthetic protocol towards pyrimido[1,6-a]quinoxalines via the intermolecular trapping of acyl(quinoxalin-2-yl)ketenes by Schiff bases.
According to the literature data, precursors I and II are unsuitable for achieving the proposed goal as the generated ketene IV reacts only at its oxo-diene fragment in intermolecular trapping reactions with various dienophiles [57][58][59][60][61][62]. Under these circumstances precursors III generating ketenes V seemed to be the only suitable candidates for the development of a strategy towards pyrimido[1,6-a]quinoxalines.
First, we studied the decarbonylation of precursors III -3-acylpyrrolo[1,2-a]quinoxaline-1,2,4(5H)-triones (PQTs, 1a-h) by simultaneous thermal analysis (STA, Table 1). According to the data obtained, PQTs 1a-h underwent thermal decomposition with a mass loss accompanied by an endothermic effect and CO evolution ( Figure 4). The values of the mass loss corresponded to the elimination of a CO molecule from a PQT. Having taken into account the results of the thermal analysis, we examined the feasibility and conditions of the intermolecular reaction of the ketene generated from PQT 1a with benzal-  Table 2.
The reaction mixtures contained only three types of products, and we succeeded to identify each of them. The structures of the reaction products were elucidated as the desired pyrimido[1,6a]quinoxaline 3a, quinoxalinone 4a [29] and pyrido[1,2a]quinoxaline 5a [29] (Scheme 2). Product IV of an alternative intermolecular trapping reaction (Table 1) was not detected.
The most likely way of the formation of quinoxalinone 4a is hydration of the ketene with subsequent decarboxylation (Scheme 2); more careful drying the reaction vials and solvents easily reduced the amount of compound 4a.
The formation of pyrido[1,2-a]quinoxaline 5a can be explained by a concurrent process of ketene dimerization (Scheme 2) [29] in comparison to the intermolecular trapping of it by benzalaniline (2a). Since the yields of the target product 3a decreased and the yields of compound 5a increased at prolonged time of reaction, the formation of the target compounds deemed to be reversible.
Performing the reaction under solvent-free conditions at the onset decarbonylation temperature (Table 1) exceeded our expectations and gave the best yields of the target compound 3a ( Table 2, entry 6).
Being inspired by the optimization results obtained, we examined the scope of the reaction applying the developed methodology with PQTs 1a-h and Schiff bases 2a-d. The results are shown in Figure 5.
Unfortunately, our attempts to involve Schiff bases synthesized from aliphatic aldehydes and ketones did not give any satisfactory results because of various nucleophilic side-reactions.
We found that the intermolecular trapping worked perfectly in case of N 5 -substituted PQTs 1a-f and did not work at all with N 5 -unsubstituted PQT 1g and 1h. The failure to obtain products 3o and 3p from PQTs 1g and 1h can be explained by the occurrence of intramolecular cyclization in these ketenes resulting in the formation of furoquinoxalines 6a,b [56,63,64] which were confirmed by UPLC-MS data as the sole products of the reaction (Scheme 3).

Conclusion
We have developed a facile synthesis of pyrimido[1,6a]quinoxaline derivatives via the intermolecular trapping of thermally generated acyl(quinoxalin-2-yl)ketenes by Schiff bases. The reaction proceeds under solvent-free conditions without any additives and catalysts. The elaborated method might be applicable to the syntheses of pharmaceutically important substances.

Supporting Information
Supporting Information File 1 Experimental details, copies of 1 H and 13 C NMR spectra of pyrimido[1,6-a]quinoxalines 3a-n, STA plots of PQT 1a-h and X-ray crystal structure details of compounds 3g,j.