Herein, we report the development of a novel route to tricyclic lactam products via a facile aza-[4 + 2] cycloaddition of catalytically generated acyliminium ions. Employing a Ca(NTf2)2/nBu4NPF6 catalyst system in low loadings, a range of diverse fused ring systems can be synthesised in predominantly good yields.
Graphical abstract: Aza-[4 + 2] cycloadditions employing catalytically derived N-acyliminium ions
The development of new methods to access small, fused ring systems bearing multiple functional groups and synthetic handles remains an important goal within synthetic organic chemistry. 3-Substitued isoindolones represent an important pharmacophore found in an increasing number of bioactive small molecules,1,2 with growing interest in tertiary substituted aza-cycles (Fig. 1).
Fig. 1 Exemplar bioactive small molecules.
Access to these fragments is typically through intramolecular cyclisation of pendant sp-rich functional groups,3 mediated by both stoichiometric4,5 and catalytic Lewis acids (Fig. 2).6 Further cyclisation reactions employing stoichiometric Brønsted acids have also been reported.7,8 Additional methods include aza-Navarov cyclisation cascades9 and tandem Aza-Prins/Friedel–Crafts reactions.10 Although these methods present elegant solutions, many of them take advantage of the inherent reactivity of a pendant functional group. This somewhat limits the scope of the reaction and builds in additional synthetic steps. Herein, we report our work in developing a [4 + 2] cycloaddition protocol to produce 6,5,6 fused tertiary aza-cycles in good to excellent yields.
Fig. 2 Recent methods to access isoindolone cores.
Our work began by taking advantage of our previously reported methodology to access N-acyliminium ions via catalytic dehydration.11–14 We therefore began our investigation using these conditions, employing hydroxyisoindolinone 1 and dimethylbutadiene (2a) as model substrates (Table 1). As shown, optimisation of temperature, solvent, and catalyst loading led to conditions that afforded the [4 + 2] product in high isolated yield. In essence, the reaction proceeded in a range of solvents, with temperature being the variable that had the biggest impact. Furthermore, running the reaction in the absence of any part of the catalyst system was unsuccessful.
Table 1 Optimisation studies
image file: d2ob01663j-u1.tif
Entry Catalyst Additive Loading Temp. Solvent Time (min) Yield
1 Ca(NTf2)2 nBu4NPF6 10 mol% 65 °C DCM 30 73%
2 Ca(NTf2)2 nBu4NPF6 10 mol% 65 °C EtOAc 30 81%
3 Ca(NTf2)2 nBu4NPF6 10 mol% 65 °C DCE 30 82%
4 Ca(NTf2)2 nBu4NPF6 10 mol% 65 °C HFIP 30 57%
5 Ca(NTf2)2 nBu4NPF6 10 mol% 65 °C Toluene 30 71%
6 Ca(NTf2)2 nBu4NPF6 10 mol% 40 °C DCE 30 n.r.
7 Ca(NTf2)2 nBu4NPF6 10 mol% 50 °C HFIP 60 57%
8 Ca(NTf2)2 nBu4NPF6 10 mol% 80 °C EtOAc 90 70%
9 Ca(NTf2)2 nBu4NPF6 10 mol% 80 °C DCE 90 82%
10 Ca(NTf2)2 nBu4NPF6 10 mol% 80 °C Toluene 90 76%
11 Ca(NTf2)2 nBu4NPF6 5 mol% 80 °C DCE 60 82%
12 Ca(NTf2)2 nBu4NPF6 5 mol% 80 °C DCE 30 88%a
13 Ca(NTf2)2 nBu4NPF6 1 mol% 80 °C DCE 90 32%
14 Ca(NTf2)2 nBu4NPF6 1 mol% 80 °C DCE 120 74%a
15 Ca(NTf2)2 — 5 mol% 80 °C DCE 30 n.r.
16 — nBu4NPF6 5 mol% 80 °C DCE 30 n.r.
17 — — — 80 °C DCE 30 n.r.
a 1.5 equiv. diene used.
With these conditions now optimised, we wanted to explore the substrate tolerance of the reaction, with particular emphasis on differing electronics. As shown (Fig. 3), para-electron donating (3b) and withdrawing (3c) groups afforded the desired product in good yield, with a small reduction in yield observed in the trifluoromethyl substrate. meta-Electron withdrawing (3d) and ortho, para- (3e) electron donating groups both worked well, as did benzodioxole (3f). Heterocycles were also tolerated, with pyridyl (3g, 3h) and thiazole (3i) derivatives being synthesised in moderate yields.
Fig. 3 Dimethyl butadiene derived products.
We also explored bis-phenyl diene 2b, which was also well tolerated in most cases (Fig. 4). The trifluromethyl group (3l) retarded the reaction, with prolonged reaction times and higher temperatures having little effect on the overall conversion.
Fig. 4 Further substrate scope.
We next turned our attention to unsymmetrical dienes, as up to this point, we have employed dienes bearing the same group at each position. To this end, diene 4 was synthesised15 and subjected to the above optimised conditions (Fig. 5).
Fig. 5 Unsymmetrical diene substrate scope.
Once again, the reaction was tolerant to a variety of different functional groups including electron donating (5b) and withdrawing groups (5c, 5d), meta (5e) and acid sensitive (5f) functionalities. Furthermore, sulfur (5g) and nitrogen (5h, 5i) containing heterocycles provided the desired products in decent to good yields. In all cases, only a single isomer was observed and isolated, with no evidence of other isomers present, as determined by 1H-NMR of the crude reaction mixture.
Finally, we wanted to explore the synthetic utility of these compounds (Fig. 6). To this end, 3a was synthesised on gram scale in excellent yield. Subjecting 3a to Upjohn dihydroxylation conditions (OsO4 (cat), NMO) provided diol 6, as a single diastereomer in good yield, while Prilezhaev epoxidation (mCPBA) gave epoxide 7, once again as single diastereomer in excellent yield. Finally, lithium aluminium hydride reduction afforded the pyridoisoindole 8 in decent yield.