Synthesis of Densely Functionalized Pyrimidouracils by Nickel(II)-Catalyzed Isocyanide Insertion.

A robust nickel-catalyzed oxidative isocyanide insertion/C-H amination by reaction of readily available N-uracil-amidines with isocyanides affording polysubstituted pyrimidouracils has been reported. The reaction proceeds in moderate to quantitative yield, under mild conditions (i.e., green solvent, air atmosphere, moderate temperature). The broad range of structurally diverse isocyanides and N-uracil-amidines that are tolerated make this method an interesting alternative to the currently available procedures toward pyrimidouracils.

Transition-metal-catalyzed isocyanide insertions have been dominated by palladium catalysis,[1] but the last couple of years have seen a surge in row IV base metal catalysis.[2] In addition to their higher natural abundancy and lower cost as advantages, which render these base metals more interesting from an economic and sustainability standpoint, row IV metals can also participate in one-electron transfer processes.[3] This reactivity is not observed in palladium catalysis, and opens up new, mechanistically distinct reaction pathways for the synthesis of heterocyclic scaffolds. For example, fused pyrimidine scaffolds commonly occur in nature, and related heterocycles have been reported to demonstrate a variety of biological activities, like adenosine potentiating (coronary dilatators),[4] antioxidant,[5] antifungal,[6] antiviral,[7] and antibacterial[8] activities. Additionally, similar structures have been reported as phosphoribosyl-1-pyrophosphate synthetase inhibitors[9] and dihydrofolate reductase inhibitors.[10] Although these pyrimidouracils show promise as biologically active scaffolds, the current syntheses have a very narrow scope, utilize highly specific reactants, or require harsh conditions. Recently, we reported the Cu(I)-catalyzed oxidative amination of N-uracil-amidines 1 to afford substituted xanthines (Scheme ).[11] We envisioned that combining this with isocyanide insertion could provide efficient access to pyrimido[4,5-d]pyrimidine-2,4-diones (pyrimidouracils) (Scheme ). The starting materials, the N-uracil-amidines 1, are readily available from 6-chlorouracils.[11] We started our studies to develop an intramolecular imidoylative amination toward functionalized pyrimidouracils with the model reaction of N-(1,3-dibenzyluracil)benzamidine 1a with t-BuNC (2a). When applying the reaction conditions developed to obtain 1,3-dibenzyl-8-phenylxanthine (Scheme ) in combination with 2a, no formation of the desired compound 3a was observed, irrespective of the oxidant (see Supporting Information (SI)).[11] Optimization of the transition metal catalyst led to Ni(OAc)2·4H2O as the optimal catalyst (see SI). Nickel catalysis has been extensively investigated in polymerization of isocyanides.[13] Although imidoylative nickel catalysis in small molecule synthesis was first reported in 1993,[14] it has remained underinvestigated, and only a handful of redox-neutral[15] and oxidative[16] examples have been reported since. Satisfyingly, 15 mol % Ni(OAc)2·4H2O in DMSO at 120 °C under an oxygen atmosphere gave 93% pyrimidouracil 3a after 5 h (Table , entry 1). Subsequently, the influence of the temperature, reaction time, and solvent was studied. The temperature could be lowered to 50 °C without a significant drop in yield of 3a (entries 1–6). At this temperature, the reaction is effective in both polar and apolar aprotic solvents. The efficacy of recommended solvents with respect to green chemistry[12] was investigated, most of which afforded 3a in good yield (entries 9–13), although protic ethylene glycol did not lead to product formation (entry 8). Anisole proved to be the optimal solvent in terms of both efficiency and green credentials, affording 3a in 99% yield after 16 h at 50 °C (entry 11). When we performed the reaction under air rather than under an atmosphere of molecular oxygen, no discernible decrease in yield of 3a (entries 11 and 14) was observed. In the model reaction, combining 1a and 2a, the catalyst loading could be lowered to 1% (entry 15). Unfortunately, this low loading was later found to give diminished yields when isocyanides other than t-BuNC (2a) were employed. Thus, we selected a 5 mol % catalyst as the optimal conditions for this imidoylative amination (entry 14). Curiously, when the reaction was performed under an argon atmosphere under otherwise identical conditions, the corresponding pyrimidouracil 3a was still formed in high yield (entry 16).[17] With these conditions in hand, we set out to investigate the scope of this transformation. First we investigated the reaction of different isocyanides 2 with benzimidamide 1a, affording pyrimidouracils 3 (Scheme A). Tertiary aliphatic isocyanides smoothly couple with 1a to afford the corresponding pyrimidouracils 3a and 3f under the optimized conditions. Utilization of aliphatic secondary isocyanides afforded pyrimidouracils 3b and 3c, in good yields, and even the use of functionalized N,N-diethyl-4-isocyanopentan-1-amine furnished the product 3g in acceptable yield. Both cyclic (3c) and acyclic (3b, 3g) secondary isocyanides are tolerated. Primary and benzylic isocyanides were also readily inserted, leading to the corresponding pyrimidouracils 3d, 3e, 3h, and 3i in high yield. Even aromatic isocyanides appear compatible with the optimized conditions. For example pyrimidouracil 3j was formed in good yield using our methodology. Even the electron-deficient methyl 2-isocyanobenzoate could be converted into pyrimidouracil 3m, although in diminished yield. Employing the notoriously unstable 2-naphtyl isocyanide did not furnish pyrimidouracil 3k, but led to immediate and full decomposition. Next, we turned our attention to chart the scope of the N-(1,3-dibenzyluracil)amidines 1 in this nickel-catalyzed cross-dehydrogenative imidoylative amination process. First, we studied substitutions on the N-(1,3-dibenzyluracil)benzimidamides (1b–d) as input for our reaction (Scheme B). The use of N-(1,3-dibenzyluracil)-4-chlorobenzimidamide 1b with differently substituted isocyanides 2 led to the isolation of the corresponding functionalized pyrimidouracils 4a–c in excellent yields. With a stronger electron-withdrawing trifluoromethyl group on the benzimidamide moiety (1c), our protocol generally afforded pyrimidouracils 4d–f in slightly lower yields. Similar observations were observed with the more electron-rich N-(1,3-dibenzyluracil)-4-methoxybenzimid-amide 1d affording pyrimidouracils 4g–i (Scheme B). These results confirm the above-described finding that our reaction not only is quite generally compatible with tertiary isocyanides but also tolerates primary, secondary aliphatic and aromatic isocyanides. Hereafter, N-(1,3-dibenzyluracil)alkimidamides 1e–g featuring an aliphatic rather than an aromatic R2-functionality were also investigated (Scheme ). Gratifyingly, N-(1,3-dibenzyluracil) acetimidamide 1e was readily converted to the corresponding pyrimidouracils 5a–h under the previously optimized conditions using a range of diversely functionalized isocyanides 2.

Scheme 1

Direct Oxidative and Oxidative Imidoylative Amination of N-Uracil-amidines

Table 1

Optimization of Conditions for Direct Oxidative Imidoylative Amination towards 3aa

Selected examples, full optimization study in the SI. Reaction conditions: N-(1,3-dibenzyluracil)benzimidamide (1a, 0.5 mmol, 1 equiv), tert-butyl isocyanide (2a, 1.5 mmol, 3.0 equiv), and Ni(OAc)2·4H2O (0.025 mmol, 5 mol %) were stirred at indicated temperature for indicated time.

Colored according to Chem21 solvent guide[12] (red = hazardous, yellow = problematic, green = recommended).

Yield determined by 1H NMR analysis using 1,3,5-trimethoxybenzene as an internal standard.

Ni(OAc)2.4H2O (15 mol %) and t-BuNC (2a, 1.0 mmol, 2.0 equiv).

Performed under an air atmosphere.

Ni(OAc)2·4H2O (1 mol %).

Performed under an Ar atmosphere. EG = ethylene glycol, MIBK = methyl isobutyl ketone, DMC = dimethyl carbonate, PC = propylene carbonate.

Scheme 2

Isocyanide Scope in Combination with Substituted N-(1,3-Dibenzyluracil)-benzimidamides 1a–d

Reaction conditions: 1 (0.5 mmol), isocyanide 2 (1.25 mmol), Ni(OAc)2·4H2O (0.025 mmol) in anisole (2 mL), run under air at 50 °C.

Performed on 1 mmol scale.

Scheme 3

Isocyanide Scope in Combination with N-(1,3-dibenzyluracil)alkanimidamides 1e–g

Reaction conditions: 1 (0.5 mmol), isocyanide 2 (1.25 mmol), Ni(OAc)2·4H2O (0.025 mmol) in anisole (2 mL), run under air at 50 °C.

Pyrimidouracil 5b was isolated in moderate yield (30%), presumably due to the volatility of isopropyl isocyanide. Compared to the 2- arylated pyrimidouracils 4, the 2-methylpyrimidouracils 5 are generally produced in somewhat lower yields. The relatively low yield of 5g can be explained by the promiscuous reactivity of N-(2-isocyanoethyl)morpholine. These β-amino isocyanides are known to intramolecularly form internal imidoyl species as a side reaction.[18] Aromatic isocyanides such as 2,6-dimethylphenylisocyanide are compatible with the developed conditions (5h), although naphthyl isocyanide again did not furnish isolable quantities of 5i. Increasing the size of the amidine substituent (R2) leads to a significant and consistent increase in the yield of pyrimidouracils 5. Thus, when isopropyl-substituted amidine 1f was reacted with tert-butyl-, cyclohexyl-, and 2,6-dimethylphenyl isocyanide, the corresponding pyrimidouracils 5j–m could be isolated in good to quantitative yield. Unfortunately, immediate polymerization was observed in the synthesis of 5n, as the reaction mixture turned black and turbid upon adding 2-bromo-4-fluorophenyl isocyanide. Pyrimidouracil 5n was not observed. Satisfyingly, our catalytic system has a high tolerance for the amidine substrate, as illustrated by the use of N-(1,3-dibenzyluracil)pivalimidamide 1g (Scheme ). The corresponding 7-tert-butyl pyrimidouracils 5o–s could be obtained in good to excellent yields under the optimized conditions. The substrate 1g showed excellent compatibility with secondary and tertiary aliphatic isocyanides to afford 5o and 5p, but also with benzyl isocyanide, allowing for the formation of 5q. Our system proves to be compatible with commercially available aromatic isocyanides as well, as evidenced by the isolation of 5r and 5s in good yields. Even the use of the α-acidic methyl isocyanoacetate led to the formation of 5t, albeit in low yield (14%). To directly access pyrimidouracils of type 6, featuring N-methyl substituents, we also performed several reactions with N-(1,3-dimethyluracil)benzimidamide 1h (Scheme ). The combination of 1h with primary, secondary, and tertiary aliphatic isocyanides gave the corresponding dimethylated pyrimidouracils 6a–c in good to excellent yields. Pyrimidouracils 6 are more prone to tailing, rendering column chromatography more cumbersome. Still, functionalized isocyanides perform rather well in combination with 1h. For example, the N5-functionalized pyrimidouracil 6d was isolated in a respectable 51% yield while 49% of the substrate 1h could be recovered. Apparently this reaction proceeds less readily, similar to the formation of 5g (Scheme ). Satisfyingly, the reaction with aromatic 4-methoxyphenyl isocyanide afforded the corresponding 1,3-dimethylated-5-(4-methoxy-phenyl)-pyrimidouracil 6e, albeit in only 21% yield.

Scheme 4

Isocyanide Scope in Combination with N-(1,3-Dimethyluracil)benzimidamide 1h

Reaction conditions: 1h (0.5 mmol), isocyanide 2 (1.25 mmol), Ni(OAc)2·4H2O (0.025 mmol) in anisole (2 mL), run under air at 50 °C.

It is noteworthy that our methodology can be used in combination with a deprotection step to liberate 1,3-unsubstituted pyrimidouracils 7 or 5-aminopyrimidouracils 8 allowing postfunctionalization (Scheme A). Palladium-catalyzed hydrogenolysis afforded the 1,3-debenzylated pyrimidouracil 7 with excellent conversion. However, purification proved challenging, providing the target compound 7 in only 50% isolated yield. Similarly, when pyrimidouracil 1a was treated with triflic acid the amine 8 was obtained quantitatively without further purification (Scheme B).

Scheme 5

Post-functionalization of Pyrimidouracils

In order to elucidate the mechanism, several control experiments were performed (Scheme ). A radical mechanism[19] involving homolytic aromatic substitution is not deemed likely, as addition of TEMPO (2.0 equiv) does not hamper the formation of pyrimidouracil 3a under standard reaction conditions (Scheme a).[19a,20] Such a mechanism would imply a heterolytic aromatic substitution, which is also deemed unlikely due to the fact this reactivity is not observed in the less electron-rich benzamidine analogue 9, even under elevated temperatures. Additionally, we investigated the possibility of β-hydride elimination of proposed intermediate II (Scheme ). Replacing the amidine proton with a benzyl functionality (11) completely inhibits the conversion to 5-(tert-butylimino)-5,6-dihydropyrimido[4,5-d]pyrimidine-2,4(1H,3H)-dione 12. While this lends some credibility to a mechanism including a β-hydride elimination mediated formation of the corresponding carbodiimide V, it does not fully exclude the Ni(II)/Ni(III) catalyzed mechanism (Scheme ).[16a,16b,21] The reaction is initiated by the formation of N-amidinonickel intermediate I. Subsequent insertion of the isocyanide affords C-amidinonickel intermediate II. This intermediate undergoes C–H functionalization to give nickelacycle intermediate III. After one-electron oxidation to cyclic Ni(III) intermediate IV, reductive elimination furnishes the observed pyrimidouracils and a Ni(I) species, which undergoes a second one-electron oxidation to regenerate the Ni(II) catalyst. Alternatively, β-hydride elimination from the likely intermediate II affords the carbodiimide V, which may cyclize to the product 3.

Scheme 6

Control Experiments

Scheme 7

Proposed Mechanism

In conclusion, we have developed a highly effective and robust nickel-catalyzed cross-dehydrogenative imidoylative amination of N-uracil-amidines (1) with isocyanides affording underexplored fused pyrimidouracils. The transformation proceeds with high efficiency, regardless of the steric and electronic nature of the N-uracil-amidine substrate. Additionally, this imidoylative C–H functionalization is compatible with a broad range of isocyanides, including primary, secondary, and tertiary aliphatic, benzylic, and aromatic isocyanides. Finally, we were able to liberate the corresponding deprotected products 7 and 8 using standard procedures. Current efforts in our laboratories are directed toward a better understanding of the mechanism of this type of transformation as well as further scaffold variations to access a broader range of nitrogen heterocycles.