Direct Arylation of 6-Phenylpurine and 6-Arylpurine Nucleosides by Ruthenium-Catalyzed C–H Bond Activation

Abstract

One aryl or two? The title reaction predominantly gives monoarylation products with various amounts of diarylation product being observed in almost all cases (see scheme). Aryl iodides as well as aryl bromides were reactive under the optimized reaction conditions. The multiple nitrogen atoms in the purine, and oxygen atoms in the saccharide posed no problems in these transformations.

C–H bond activation represents an efficient approach to molecular functionalization.^[1,2] In the directed C–H bond activation, Lewis basic sites are exploited to draw the catalyst proximal to the reactive center. We have been interested in the C–H bond activation and arylation of nucleobases and nucleosides using the nitrogen atoms of the purine itself. In this context, 6-arylpurine has embedded 2-phenylpyridine and 4-phenylpyrimidine motifs. 2-Arylpyridines, and benzo[h]quinoline that contains a rigidified 2-phenylpyridine structure, have been the subject of C–H bond activation/arylation strategies using Pd, Ru, Rh and Fe.^[3–6] However, any metalcatalyzed conversion of purines and purine nucleosides is a challenging proposition owing to the presence of four nitrogen atoms in the nucleobases, plus additional oxygen atoms in the sugar unit; all of these heteroatoms could participate in metal sequestration and deactivation of catalytic processes.

As shown in Figure 1, the N1 nitrogen atom of the purine is well positioned to direct C–H bond activation. Alternatively, N7 can also function in a similar capacity. In order to gain preliminary insight, energy minimization was performed, using DFT at the B3LYP/6-311++G(2d,2p) level, on 2-phenylpyridine as well as 9-benzyl-6-phenyl-9H-purine. The distance between the metal-directing nitrogen atom and the ortho-hydrogen atom on the phenyl ring were calculated from the energy minimized structures, and are shown in Figure 2.

Figure 1.

N-directed C–H bond activation in 2-phenylpyridine, and two plausible modes of N-directed C–H bond activation in C6 arylpurines and nucleosides.

Figure 2.

Energy minimized structures of 2-phenylpyridine (left) and 9-benzyl-6-phenyl-9H-purine (right).

Consistent with known X-ray structures of 2-phenylpyridine derivatives,^[7] the aryl rings in 2-phenylpyridine are not coplanar. The pyridyl N to aryl ortho-H distance in this case is 2.49 Å. By comparison, in the energy-minimized structure of 9-benzyl-6-phenyl-9H-purine, the C6-aryl group is coplanar with the purine ring. We next compared the computed structure of 9-benzyl-6-phenyl-9H-purine to the crystallographic structures of 9-benzyl-6-phenyl-8-(p-tolyl)-9H-purine and 8,8’-bis(9-benzyl-6-phenyl)-9H-purine.^[8] In both compounds, the C6-phenyl group is coplanar with the purinyl system. We also evaluated the distance between N1 and the ortho-hydrogen atom of the C-6 aryl group for these cases. In 9-benzyl-6-phenyl-8-(p-tolyl)-9H-purine, this distance is 2.43 Å whereas the distance to the N7 is 2.34 Å. Similar distances of 2.43 Å and 2.36 Å, respectively, were obtained for 8,8’-bis(9-benzyl-6-phenyl)-9H-purine. These data closely match the DFT-derived distances for 9-benzyl-6-phenyl-9H-purine shown in Figure 2. Given the similar nitrogen atom to aryl hydrogen distances in 9-benzyl-6-phenyl-9H-purine and 2-phenylpyridine, we reasoned that C–H bond activation in purines and nucleosides should be feasible. In support of this hypothesis, serendipitously, undesired arylation of the C6-phenyl group has been observed in Pd/Cu-mediated C8 arylation of 9-benzyl-6-phenyl-9H-purine with iodobenzene or p-iodotoluene (2–10 equiv).^[8]

In initial experiments, [{RuCl₂(benzene)}₂] (A) and [{RuCl₂(p-cymene)}₂] (B) were selected as catalysts for assessing the C–H bond activation of 9-benzyl-6-phenyl-9H-purine (1), and 2 equivalents of iodobenzene were used to ensure complete consumption of 1. Results from a preliminary screen of conditions are shown in Table 1. Notably, 5 mol% of catalyst A in combination with 40 mol% of PPh₃ resulted in full conversion (Table 1, entries 1–4), and catalyst A appeared marginally superior to catalyst B (compare entries 4 and 5). Replacement of K₂CO₃ with Cs₂CO₃ led to a slower reaction (entries 6 and 7), and the mono/diarylation ratio was substantially altered. This indicates an unknown, but crucial role of base in these C–H bond activation processes.

Table 1.

Preliminary evaluation of the C–H bond activation process in 9-benzyl-6-phenyl-9H-purine (1).^[a]

Entry	Catalytic system[b]	Result[c]
1	A (2.5 mol%), PPh3 (20 mol%), K2CO3 (3 equiv)	1/2a/3a = 83:15: 2[d]
2	B (2.5 mol%), PPh3 (20 mol%), K2CO3 (3 equiv)	1/2a/3a = 83:15:2[d]
3	A (5 mol%), PPh3 (20 mol%), K2CO3 (3 equiv)	1/2a/3a = 63:34:3[d]
4	A (5 mol%), PPh3 (40 mol%), K2CO3 (3 equiv)	2a = 76% 3a = 17%
5	B (5 mol%), PPh3 (40 mol%), K2CO3 (3 equiv)	1/2a/3a = 3:84:13[d]
6	A (5 mol%), PPh3 (40 mol%), Cs2CO3 (3 equiv)	1:2a:3a = 7:52:41[d]
7[e]	A (5 mol%), PPh3 (20 mol%), K2CO3 (3 equiv)	2a = 56% 3a = 36%

^[a]Reaction conditions: 1 (0.2 M) in anhydrous NMP, iodobenzene (2 equiv), 120°C, 24 h.

^[b]A [{RuCl₂(benzene)}₂], B [{RuCl₂(p-cymene)}₂].

^[c]Yields are of isolated and purified products.

^[d]The reaction was incomplete. The ratio was determined from ¹H NMR spectra by the integration of the benzyl CH₂ resonances, which appear at δ = 5.49 ppm for 1, 5.41 ppm for 2a, and 5.32 ppm for 3a. Bn = benzyl, NMP = N-methylpyrrolidone.

On the basis of the foregoing, we evaluated the use of aryl iodides and aryl bromides for the direct arylation of 9-benzyl-6-phenyl-9H-purine (1) under the optimal conditions developed in Table 1. Results from these experiments are shown in Table 2.

Table 2.

C–H bond activation/arylation of 9-benzyl-6-phenyl-9H-purine (1) with aryl iodides and aryl bromides.^[a]

Entry	Aryl halide	Yield [%][b]	2/3
1		2a: 76	4.5:1
		3a: 17
2		2b: 81	4.5:1
		3b: 18
3[c]		2c: 79	5.6:1
		3c: 14
4		2d: 82	NA[d]
		3d: trace
5		2a: 72	6:1
		3a: 12
6		2b: 73	4.3:1
		3b: 17
7		2d: 75	NA[d]
		3d: trace

^[a]Reaction conditions: 1 (0.2 M) in anhydrous NMP, aryl halide (2 equiv), [{RuCl₂(benzene)}₂] (5 mol%), PPh₃ (40 mol%), K₂CO₃ (3 equiv), 120 °C.

^[b]Yields are of isolated and purified products.

^[c]The reaction was conducted using 10 mol% of [{RuCl₂(benzene)}₂] and 80 mol% of PPh₃.

^[d]Not applicable since only a trace of the diaryl product was detected.

Results in Table 2 indicate that reactions can be accomplished with both aryl iodides and aryl bromides, in yields of 75–99% (combined for mono and diarylated products). Yields of reactions with aryl bromides were respectable (Table 2, entries 5–7) but a little lower than those with aryl iodides.

We next explored arylations of the more complex 2’-deoxynucleoside substrates, which are quite labile and prone to facile deglycosylation.^[9] The requisite 6-arylpurine 2’-deoxyribonucleosides (6-aryl 2’-deoxynebularines) are readily available via our previously reported procedures.^[10] Results from the nucleoside arylation reactions are shown in Table 3.

Table 3.

C–H bond activation/arylation of 3’,5’-di-O-silyl 6-arylpurine 2’-deoxyribonucleosides (4) with aryl iodides and aryl bromides.

Entry	R =	Aryl halide	Yield [%][a]	5/6
1[b]	H		5a: 46	2:1
			6a: 23
2[c]	H		5a: 65	4.6:1
			6a: 14
3[b]	H		5b: 52	3.2:1
			6b: 16
4[c]	H		5b: 54	7.7:1
			6b: 7
5[c]	H		5c: 53[d]	7.6:1
			6c: 7
6[c]	H		5d: 62	4.1:1
			6d: 15
7[c]	F		5e: 31	1:1.5
			6e: 44
8[c]	F		5f: 42	2.1:1
			6f: 20
9[c]	OPh		5g: 29	1:1.8
			6g: 53
10[e]	OPh		5g: 24	1:3
			6g: 72
11[c]	OMe		5h: 39	2.4:1
			6h: 16
12[c]	OMe		5i: 46	2.7:1
			6i: 17
13[c]	H		5a: 36	2:1
			6a: 18
14[c]	H		5b: 57	3.2:1
			6b: 18

^[a]Yields are of isolated and purified products.

^[b]Reaction conditions: 4a (0.2 M) in anhydrous NMP, aryl halide (2 equiv), [{RuCl₂(benzene)}₂] (10 mol%), PPh₃ (40 mol%), K₂CO₃ (3 equiv), 120°C, 30 h.

^[c]Reaction conditions: 4a–d (0.2 M) in anhydrous NMP, aryl halide (2 equiv), [{RuCl₂(benzene)}₂] (10 mol%), PPh₃ (80 mol%), K₂CO₃ (3 equiv), 120°C.

^[d]Owing to the close R_f values of products, a trace of the diarylated product was present.

^[e]Reaction conditions were the same as [c] except that [{RuCl₂(p-cymene)}₂] was used in place of [{RuCl₂(benzene)}₂].

The results in Table 3 indicate that the procedure is readily applicable to the sensitive 2’-deoxyribonucleoside substrates. However, 10 mol% of the Ru catalyst was needed to obtain complete consumption of the precursor. In contrast to the reactions of purine 1, where a 1:8 ratio of Ru catalyst/PPh₃ was needed for complete reaction (Table 1 entry 4), use of a 1:4 or 1:8 ratio of the Ru catalyst/PPh₃ led to full conversion of 4a (entries 1–4 in Table 3), but there were some differences. With iodobenzene, increased PPh₃ resulted in a better yield as well as mono/diarylation ratio (entry 1 versus 2). With p-iodotoluene, again a better ratio of mono/diarylation was observed with increased PPh₃, although the yield of the monoarylation remained almost the same (entry 3 versus 4). Electron-rich p-iodoanisole and electron-deficient p-iodoacetophenone both yielded successful arylation reactions (entries 5 and 6). Interestingly, for reasons currently unknown, when R = F or OPh (4b, 4c), greater diarylation was observed in reactions with iodobenzene (entries 7 and 9). The same trend was also observed in the arylation of 4c with iodobenzene using catalyst B (entry 10). With 4d (R = OMe), greater monoarylation is again observed (entries 11 and 12). As in the case of purine 1, aryl bromides were also reactive (entries 13 and 14).

Using catalyst A, we then examined C–H bond activation/arylation of m-nitrophenyl nucleoside derivative 7^[10] and 2-amino-6-phenylpurine nucleoside 8^[11] (Figure 3) with iodobenzene. In the case of 7, no product formation was seen but some precursor decomposition occurred, thus indicating very electron-depleted aromatic rings, not surprisingly, are resistant to arylation (see mechanism in Scheme 1). Substrate 8 that is structurally similar to 4a, but with an additional C-2 amino group, remained largely intact and showed no product formation. It appears therefore that presence of the amino group abolishes any reactivity.

Figure 3.

Nucleoside substrates that did not yield arylation products. TBDMS = tert-butyldimethylsilyl.

Scheme 1.

Possible mechanism for the C–H bond activation/arylation of 6-arylpurine and 6-arylpurine 2’-deoxyribonucleosides.

Since in previous experiments 2 equivalents of aryl halide were used to achieve complete consumption of the purine or nucleoside precursor, we evaluated the arylation of 1 and 4a with a variety of stoichiometries of iodobenzene; all other reaction conditions remained the same. These results are shown in Table 4. With purine 1, incomplete reaction was observed with 1 equiv of iodobenzene, but substantially more monoarylated product 2a was formed (entry 1). However, with 120 mol% of iodobenzene the result was nearly the same as that obtained with 200 mol% (compare entry 2 in Table 4 to entry 1 in Table 2). Increasing iodobenzene to 400 mol% yielded a greater proportion of diarylated product 3a, but complete conversion to diarylated 3a did not occur (entry 3). Doubling the reaction time did not increase diarylation either (entry 4). Stoichiometry of iodobenzene had very little effect on the mono/diarylation ratio of nucleoside precursor 4a (entries 5–7 in Table 4 and entry 2 in Table 2), although a prolonged reaction time (60 h) was detrimental, and led to a lower yield (entry 7).

Table 4.

Reactions of 1 and 4a using various stoichiometries of iodobenzene.

Entry	Substrate	Mol% of Ph–I	t[h]	Yield [%][a]	Mono/diaryl products
1[b]	1	100	24	2a: 91%	2a/3a = 45:1[c]
				3a: 2%
2[b]	1	120	24	2a: 80%	2b/3b = 6.7:1
				3a: 12%
3[b]	1	400	24	2a: 60%	2b/3b = 2:1
				3a: 30%
4[b]	1	400	48	2a: 59%	2b/3b = 1.8:1
				3a: 32%
5[d]	4a	120	30	5a: 62%	5a/6a = 6.2:1
				6a: 10%
6[d]	4a	400	30	5a: 62%	5a/6a = 5.2:1
				6a: 12%
7[d]	4a	400	60	5a: 53%	5a/6a = 4.1:1
				6a: 13%

^[a]Yields are of isolated and purified products.

^[b]Reaction conditions: 1 (0.2 M) in anhydrous NMP, PhI (see Table), [{RuCl₂(benzene)}₂] (5 mol%), PPh₃ (40 mol%), K₂CO₃ (3 equiv), 120°C.

^[c]Reaction was incomplete, 4% of 1 was recovered.

^[d]Reaction conditions: 4a (0.2 M) in anhydrous NMP, PhI (see Table), [{RuCl₂(benzene)}₂] (10 mol%), PPh₃ (80 mol%), K₂CO₃ (3 equiv), 120°C.

While this work was in progress, complementary Pd-catalyzed C–H bond activation of 6-arylpurines and three acetate-protected 6-arylpurine ribonucleosides has been reported.^[12] It is therefore reasonable to offer a comparison of the two approaches. In contrast to the reactions reported here, those catalyzed by Pd required 30 equiv of aryl iodides, with reaction times from 48 h (nucleosides) to 60 h (purines), at 120 °C.^[12] Thus, the present conditions are substantially more conservative on the aryl halide and generally provide faster reactions. Pd catalysis produces only monoarylation,^[12] plausibly due to slow reactions, whereas under Ru catalysis both mono and diaryl products are formed. Use of AcOH as solvent at 120 °C and N₂ atmosphere have been stated as crucial for the Pd-catalyzed reactions.^[12] Exposure to AcOH at elevated temperature may be unsuitable for acid-sensitive substrates such as the deoxyribonucleosides described here. Also, an inert atmosphere is strictly necessary for Pd-catalyzed reactions. Plausibly, this is to suppress undesired dimerization of the purine, as dimerization has been observed previously when air was introduced into Pd-catalyzed arylation reactions.^[8] Most notably, as shown here, aryl bromides can be used in the Ru-catalyzed chemistry, whereas aryl bromides and chlorides were unreactive under Pd catalysis.^[12]

In analogy to reactions of 2-arylpyridines,^[4d] a possible mechanism involving the purinyl N1 nitrogen atom is depicted in Scheme 1. The monoarylated products could reenter the catalytic cycle resulting in diarylation. Consistent with an N-directed electrophilic attack by the aryl-Ru^IV complex onto the C-6 aryl ring, no reaction was observed with the m-nitrophenyl nucleoside derivative 7. This is because reaction would have to occur at the electron deficient para- or ortho-positions to the nitro group.

Due to interest in oxidative C–H cross-coupling processes,^[3c,13–17] we evaluated whether purinyl nitrogen atom-directed C–H bond activation could be used to promote C–H cross coupling.^[18] This reaction gave dimerized 8,8’-bis(9-benzyl-6-phenyl)-9H-purine^[8] in ca. 20% yield, indicating a facile dimerization of 1 under oxidative conditions (50% of 1 was recovered).

We have demonstrated that Ru-catalyzed C–H bond activation/arylation can be accomplished with purines and deoxyribonucleosides. Both aryl iodides and aryl bromides are reactive, leading to densely functionalized products in relatively few steps. Metal-catalyzed approaches, such as that described here, provide facile routes to novel entities that are otherwise not easily accessed. This is important in the context of the high biological value placed on purine and nucleoside derivatives. Further work is currently ongoing in our laboratories on these and related chemical processes, including understanding the role of the purinyl N7, if any.