Synthesis of N–H Bearing Imidazolidinones and Dihydroimidazolones Using Aza-Heck Cyclizations

Abstract

The synthesis of unsaturated, unprotected imidazolidinones via an aza-Heck reaction is described. This palladium-catalyzed process allows for the cyclization of N-phenoxy ureas onto pendant alkenes. The reaction has broad functional group tolerance, can be applied to complex ring topologies, and can be used to directly prepare mono- and bis-unprotected imidazolidinones. By addition of Bu₄NI, dihydroimidazolones can be accessed from the same starting materials. Improved conditions for preparing unsaturated, unprotected lactams are also reported.

Graphic Abstract

Imidazolidinones are common motifs in both synthetic and natural bioactive compounds of medicinal relevance.^[1] Such compounds have shown activity against a range of conditions, including leukemia, lung cancer, metabolic disorders, and HIV. Particularly common amongst these bioactive compounds are imidazolidinones bearing unsaturation or other functionality vicinal to one of the nitrogen centers, or those bearing free N–H groups (Figure 1).^[2] Imidazolidinones can also serve as versatile synthetic intermediates as they can be converted into other functional groups such as diamines and guanidines.^{[1o, 3]}

Figure 1.

Representative Bioactive Imidazolidinones.

Because of their considerable value, various transition metal-catalyzed approaches to imidazolidinones have been described. These include Wacker-type cyclizations,^[4] carboaminations,^[5] alkene diamination,^[6] C–H functionalization,^[7] and other methods.^[8] However, with the exception of C–H functionalization, all require protection or acidification of the nitrogen centers, and result in N-protected or N-functionalized imidazolidinones; none of these methods can deliver imidazolidinones with a free N–H.^[9] Furthermore, none (including C–H functionalization) can prepare bis-unprotected imidazolidinones (Figure 2).

Figure 2.

State-of-the-Art in Imidazolidinone Synthesis.

Recently, Heck-type cyclizations of nitrogen electrophiles have emerged as a powerful and flexible strategy for preparing nitrogen-containing heterocycles,^[10],[11] and we have shown that unsaturated lactams bearing free N–H groups can be prepared via the palladium-catalyzed cyclization of O-phenyl hydroxamates onto pendant alkenes.^[12] We hypothesized that the synthesis of N–H bearing imidazolidinones from N-phenoxyureas might also be possible, which would allow for the direct synthesis of unprotected imidazolidinone scaffolds bearing unsaturation vicinal to one of the nitrogen centers. Herein, we describe the successful realization of that goal, and report conditions for the preparation of unsaturated, unprotected imidazolidinones using an aza-Heck strategy. Using easily accessed starting materials, and commercially available catalytic components, we show that the reaction enjoys broad functional group tolerance and is highly flexible with respect to accessible ring topologies. Further, by the addition of iodide additives, isomeric dihydroimidazolones can also be prepared from the same starting materials. These studies have also revealed catalytic conditions that are easier to implement than our earlier aza-Heck cyclizations, and we have demonstrated that these new conditions can also be used in unsaturated lactam preparation.

To start our investigation we required easy access to N-phenoxy ureas. Inspired by the work of Beauchemin,^[13] we designed a convenient two-pot, three component sequence that allows the assembly of the substrates from unsaturated carbonyls, amines, and 4-nitrophenyl 1-phenoxycarbamate in up to 85% yield over two steps (1, Scheme 1).^[14]

Scheme 1.

Access to N-Phenoxy Ureas.

Our study began by utilizing our previously optimal catalytic system of (COD)Pd(CH₂SiMe₃)₂ (2) and tris(2,2,2-trifluoroethyl)phosphite, and we were delighted to observe the desired cyclization of 3 in reasonable yield (Table 1, entry 1). We recognized, however, that 2 has limited commercial availability and is somewhat thermally sensitive.^[15] As such, we investigated the use of more accessible palladium precatalysts. Consistent with a Pd(0)/Pd(II) catalytic cycle, the use of (MeCN)₂PdCl₂ without an external reductant gave no desired product (entry 2). With the addition of metallic reducing agents, such as magnesium, small amounts of product were observed (entry 3). The use of soluble PhMgBr as the reducing agent proved more effective, even at half the catalyst loading (entry 4), and [(cinnamyl)PdCl]₂ proved to be superior as precatalyst (entry 5). At this point, the major byproducts in the reaction were due to alkene isomerization of both the product (to form dihydroimidazolone 5) and the starting material. Previous reports have shown that silver additives can suppress alkene isomerization in traditional Heck-cyclizations.^[16] Thus, we investigated the use of silver salts and found that the addition of substoichiometric AgOTs dramatically reduced alkene isomerization, allowing the desired product 4 to be formed in 86% yield with only minimal formation of byproduct 5 (entry 6). Subsequent studies, however, showed silver is not required in all reactions to prevent alkene isomerization (see below).

Table 1.

Reaction Optimization.

Entry	Pd Catalyst [mol %]	Additive [mol %]	Yield [%, 4/5] [a]
1	(COD)Pd(CH2SiMe3)2 [10]	none	79/4
2	(MeCN)2PdCl2 [10]	none	0/6
3	(MeCN)2PdCl2 [10]	Mg0 [300]	9/18
4	(MeCN)2PdCl2 [5]	PhMgBr [10]	29/31
5	[(cinnamyl)PdCl]2 [2.5]	PhMgBr [5]	62/24
6	[(cinnamyl)PdCl]2 [2.5]	PhMgBr [5] AgOTs [10]	86/6

^[a]Yield calculated by ¹H NMR with 1,3,5-trimethoxybenzene as internal standard.

The scope of reaction was next investigated (Scheme 2). The model substrate 4 was isolated in 84% yield, and closely related derivatives were prepared with similar efficiency (6-11). More substituted alkenes, including tetrasubstituted alkene, also participate well in the reaction (12-14). Notably, the latter two cases result in imidazolidinones bearing fully substituted carbon centers. In addition, in these cases (as well as others noted in the table) alkene isomerization was not an issue and silver additives were not required. Longer alkenyl derivatives were also tolerated (15-16). Excitingly, these transformations also demonstrate that bis-unprotected imidazolidinones can be prepared in good yield, allowing access to this motif for the first time via a metal-catalyzed strategy.

Scheme 2.

Substrate Scope.

A variety of ring topologies could also be prepared using this method. Examination of [6,5]-fused bicycle (17) gave a near quantitative yield, albeit in a 60:40 isomeric ratio of allylic-to-homoallylic urea. The closely related compound 18 was produced without alkene isomerization, suggesting a potential role of the Lewis basic pyridine in the isomerization of 17. [6,5]-spirocyclic ureas can also be prepared in good yield (19-21). Interestingly, both 19 and 20 were also obtained as alkene mixtures under the standard conditions. In contrast, with the use of 2 as precatalyst, this isomerization of the benzyl- (19) or para-methoxyphenyl-protected (PMP, 21) products could be avoided. Although these latter conditions do not allow access to the unprotected urea 20, the tolerance of the PMP-group provides a protecting group whose removal is compatible with alkenes, enabling efficient access to these unprotected ureas. One limitation that we have found is that tertiary substituents on the non-reacting nitrogen are not tolerated (22), presumably due to adverse steric interactions.

As with other aza-Heck cyclizations, 5-membered ring formation is considerably more efficient than for larger rings (23). Increased substitution on the 5-membered ring, however, is well tolerated (24), as is substitution on the alkyl group (25). A broad range of functional groups are also compatible with the cyclization, including ethers (6, 9, 21, 23, 29), esters (8, 33), isolated alkenes (35), aromatic halogens (10-11, 28), trifluoromethyl groups (7), protic functionalities (15-16, 20, 24, 26, 33), and a range of heterocycles (17-18, 26-27).

Although stereocenters exo to the forming rings do not provide stereocontrol (29), endocyclic stereocenters provide good to excellent levels of stereocontrol in ring formation (17-18, 30-33).^[17]

Urea-forming aza-Heck reactions can also be sequenced with our earlier aza-Heck protocol to rapidly build up complex nitrogen-containing polycycles. For example, aza-Heck cyclization of O-phenyl hydroxamate 36 using the previously reported aza-Heck strategy results in spirocyclic lactam 37 in good yield.^[12] Two steps covert 37 to N-phenoxy urea 38, which can then be converted to fused tricyclic product 39 in 93% isolated yield as single diastereomer using the urea-forming protocol (Scheme 3).

Scheme 3.

Access to Tricyclic Imidazolidinone Using Aza-Heck Strategies

As mentioned before, isomeric dihydroimidazolones were identified as the major byproduct in early optimization studies. Although we were able to avoid this byproduct under the conditions described above, we recognized that deliberate formation of the dihydroimidazolones would also be useful due to their prevalence in bioactive compounds.^[18] By adding iodide, in the form of Bu₄NI, selective formation of dihydroimidazolone is possible (Scheme 4). With the model substrate 3, 80% isolated yield of dihydroimidazolone 5 was obtained. A number of related examples are shown in Scheme 4, which demonstrates that highly substituted, and functionalized dihydroimidazolones can be obtained in good yields. Thus, simply by appropriate selection of additives, either imidazolidinones or dihydroimidazolones can be accessed from the same starting materials.

Scheme 4.

Preparation of Dihydroimidazolones.

To further demonstrate the utility of this transformation, we investigated the synthesis of factor Xa inhibitor 49, which has potential use as an anti-coagulant (Scheme 5).^[19] The aza-Heck cyclization of N-phenoxy urea 45 worked efficiently on gram scale to produce imidazolidinone 46. Alkylation with mesylate 47 yielded the cyanourea 48 in 72% yield, which can be converted into 49 using a known procedure.^[19]

Scheme 5.

Preparation of Factor Xa Inhibitor 49.

Finally, we have found that the catalytic conditions developed for this aza-Heck cyclization also work with cyclizations of O-phenyl hydroxamates. Compared to our previously published reaction conditions, which required the use of thermally sensitive 2,^[12],[15] the yields using these new conditions are either comparable or higher with much lower catalyst loading (Scheme 6). More importantly, these new conditions only require widely available precatalyst components, which considerably lowers the barrier for application of this method.

Scheme 6.

Examples of Lactam Substrates

In summary, we have developed a mild, catalytic method for the synthesis of unprotected, unsaturated imidazolidinones and dihydroimidazolones. Identification of PhMgBr as an external reductant allowed for the use of a cheaper, more accessible palladium(II) precatalyst in lower catalyst loadings compared to our earlier aza-Heck protocols. This method tolerates a wide array of functional groups, ring topologies, and functionalized alkenes, and can also be directly applied to the synthesis bioactive molecules.