Palladium‐Catalyzed Aryl Amination Reactions of 6‐Bromo‐ and 6‐Chloropurine Nucleosides

Abstract

Palladium‐catalyzed C—N bond forming reactions of 6‐bromo‐ as well as 6‐chloropurine ribonucleosides and the 2′‐deoxy analogues with arylamines are described. Efficient conversions were observed with palladium(II) acetate/Xantphos/cesium carbonate, in toluene at 100 °C. Reactions of the bromonucleoside derivatives could be conducted at a lowered catalytic loading [5 mol% Pd(OAc)₂/7.5 mol% Xantphos], whereas good product yields were obtained with a higher catalyst load [10 mol% Pd(OAc)₂/15 mol% Xantphos] when the chloro analogue was employed. Among the examples evaluated, silyl protection for the hydroxy groups appears better as compared to acetyl. The methodology has been evaluated via reactions with a variety of arylamines and by synthesis of biologically relevant deoxyadenosine and adenosine dimers. This is the first detailed analysis of aryl amination reactions of 6‐chloropurine nucleosides, and comparison of the two halogenated nucleoside substrates.

Introduction

Palladium‐catalyzed C—N bond forming reactions have become an important type of transformation in the synthetic repertoire.1 In this context, we have been gaining an understanding of and exploiting such catalyzed amination reactions of nucleosides.2 The high biochemical, physiological and medicinal value of compounds that can arise by alteration of the nucleoside scaffold adds significant importance to the development of methods for nucleoside modification.3,4

In 1999 we demonstrated, for the first time, that N ⁶‐aryl‐2′‐deoxyadenosine analogues could be efficiently synthesized via C—N bond‐formation using a suitable Pd‐ligand complex.5 Subsequently, we analyzed a series of ligands for C—C and C—N bond formation.6 This led to the identification of specific ligands that were effective for aryl amination. These included 2‐(dicyclohexylphosphino)‐2′‐(N,N‐dimethylamino)‐1,1′‐biphenyl and (±)‐BINAP. In subsequent aryl amination work the same two ligands also proved to be useful in gaining access to N ⁶‐(6‐benzo[a]pyrenyl)‐2′‐deoxyadenosine.7

In all these cases, the halopurine nucleoside tested and used was 6‐bromo‐9‐[2‐deoxy‐3,5‐di‐O‐(tert‐butyldimethylsilyl)‐β‐d‐ribofuranosyl]purine. In comparison, there are fewer analyses of the use of the 6‐chloropurine nucleoside analogues in Pd‐mediated amination reactions. Among these, reactions of 6‐chloro‐9‐[2,3,5‐tri‐O‐(tert‐butyldimethylsilyl)‐β‐d‐ribofuranosyl]purine using Pd₂(dba)₃/BINAP with either Cs₂CO₃ or t‐BuOK were evaluated.8 Although reactions of alkylamines proceeded well, comparable reactions of arylamines were not efficient. Using Pd(OAc)₂/(±)‐BINAP/Cs₂CO₃ we have studied the amination of silyl‐protected 6‐chloro‐9‐[2‐deoxy‐3,5‐di‐O‐(tert‐butyldimethylsilyl)‐β‐d‐ribofuranosyl]purine with a sterically‐demanding alkylamino derivative of a polycyclic aromatic hydrocarbon.9 Subsequently we have synthesized more elaborate carcinogen‐nucleoside adducts using this chloronucleoside.10 We have also studied C—N(sp ²) bond‐forming reactions of 6‐bromo‐ and 6‐chloropurine nucleosides with azoles.11

Despite these results, there has not been the development of effective Pd‐catalyzed aryl amination methods to involve 6‐chloropurine nucleosides. Also, no comparative performance evaluation of 6‐bromo‐ and 6‐chloropurine nucleosides under Pd‐catalyzed aryl amination conditions has been conducted. In this context it should be noted that S_NAr displacement reactions between a 6‐bromopurine nucleoside and two arylamines have been accomplished, but reactions of corresponding 6‐chloro compound are stated to be ineffective.12

Pd catalysis appears to offer a simple, unifying method for involving both 6‐bromo‐ and 6‐chloropurine nucleosides in aryl amination reactions. This paper reports our studies on such reactions of these nucleoside derivatives.

Results and Discussion

For the current work, 6‐bromo‐ and 6‐chloropurine ribonucleosides as well as the 2′‐deoxy analogues were chosen for analysis. We also wanted to assess silyl and acetyl protecting groups for the sugar moieties. Thus, six compounds 1a–2c (Figure 1) became the choices for the initial experiments. The overall goal was not the determination of the best conditions for reactions of each substrate, but rather to find generally optimal conditions for the aryl amination reactions that can be applied to the synthesis of a broad class of N ⁶‐aryladenosine analogues.

Figure 1.

Six nucleoside substrates selected for analysis.

Bromo nucleosides 1a–c were conveniently synthesized through a diazotization‐bromination protocol11,12 by appropriate modification of a literature route.13 Chloro nucleoside 2a 14 was prepared by silylation of 6‐chloro‐9‐(2‐deoxy‐β‐d‐ribofuranosyl)purine15 whereas 2b 11,16 and 2c were prepared by silylation and acetylation of commercially available 6‐chloro‐9‐(β‐d‐ribofuranosyl)purine. Alternatively, 2c can be prepared by chlorination of inosine triacetate.17

With these compounds, we began to evaluate conditions for aryl amination. We selected six ligands, 2‐(dicyclohexylphosphino)‐2′‐N,N‐(dimethylamino)‐1,1′‐biphenyl (L1), (±)‐BINAP (L2), Xantphos (L3), 2‐(dicyclohexylphosphino)biphenyl (L4), 2‐(di‐tert‐butylphosphino)biphenyl (L5) and 4‐(diphenylphosphino)‐9,9‐dimethylxanthene (L6) shown in Figure 2. The first three were chosen since these have previously yielded effective amination catalysts for nucleosides,2,5–11,18–24 whereas L4 and L5 were included due to their similarity to L1. Monophosphane L6 was included for comparison to L3. Pd(OAc)₂ and Pd₂(dba)₃ were the metal sources, K₃PO₄ and Cs₂CO₃ were chosen as bases.

Figure 2.

Six ligands selected for analysis.

Our initially reported work on 1a 5,6 was the starting point upon which the newer results were developed. Using p‐toluidine as a representative amine in the initial test reactions, the most significant results (product yields >50%) from several experiments are shown in Table 1.

Table 1.

The publisher did not receive permission from the copyright owner to include this object in this version of this product. Please refer either to the publisher's own online version of this product or the printed product where one exists.

Our first attempt was to further tune conditions in entry 1 using substrate 1a. Use of 10 mol% Pd(OAc)₂/15 mol% L1, under otherwise comparable conditions to entry 1, resulted in a poor yield (37%, 9 h). Use of 10 mol% Pd₂(dba)₃/30 mol% L1/1.5 molar equiv. Cs₂CO₃ in toluene at 100 °C, did not produce an improvement (41% yield in 4 h). This led us to consider L2 as well as L3. Initial reactions were conducted with 10 mol% Pd(OAc)₂/15 mol% L2 or L3/1.5 molar equiv. Cs₂CO₃, in toluene at 100 °C. Under these conditions, the reaction utilizing L2 gave a 64% yield of product, whereas that involving L3 gave 89% yield (entries 2 and 3). We next considered reducing the Pd/L3 loading by half, and this modification produced essentially no difference, resulting in an excellent 90% yield of the amination product (entry 4). Under these best conditions, replacing L3 with L4 or L5 resulted in poorer results (entries 5 and 6).

The foregoing results formed the basis for subsequent experiments, and a similar trend was observed with the ribose derivative 1b, where the lower Pd(OAc)₂/L3 loading gave excellent product recovery (compare entries 7 and 8). Whether both phosphane units in L3 are critical for an effective catalyst was next queried. Use of L6 in place of L3 (entry 9) clearly showed that L3 was superior.

Next, optimization of amination conditions with the chloronucleosides was conducted, starting with 2a. Use of 10 mol% Pd₂(dba)₃/30 mol% L1 with either K₃PO₄ or Cs₂CO₃ in DME gave ca. 50% product yields (entries 10 and 11). In contrast, this catalytic system with Cs₂CO₃ in PhMe gave a high 80% yield (entry 12), but replacing Pd₂(dba)₃ with 10 mol% Pd(OAc)₂ lowered the yield to 59% (entry 13). We next investigated the use of L2. At 10 mol% Pd(OAc)₂/15 mol% L2 an excellent 88% product yield was realized (entry 14), which decreased to 67% when the catalytic load was reduced by half (entry 15). When L2 was replaced with L3, the higher and lower catalyst load produced identical, good results (entries 16 and 17). On the basis of these experiments, substrate 2b was then investigated. At 10 mol% Pd(OAc)₂/15 mol% L2, again an excellent 88% yield of aminated product was obtained (entry 18), which remained unaltered when the catalyst load was decreased (entry 19). Upon replacement of L2 with L3, some other differences emerged. Use of the higher 10 mol% Pd(OAc)₂/15 mol% L3 resulted in an excellent 92% yield of product, but reduction of the catalytic load by half produced only a 74% yield (compare entries 20 and 21). Here again presence of 2 phosphane units in L3 proved essential as shown by the poorer result with L6 (entry 22)

These results indicated some differences in the reactions of 2a and 2b. With 2a, the higher and lower Pd(OAc)₂/L3 loads produced comparable results, whereas this was not the case with Pd(OAc)₂/L2. The exact opposite trend was observed with 2b, where the higher and lower loads of Pd(OAc)₂/L2 showed comparable results, but not with Pd(OAc)₂/L3. In an attempt to gain additional insight, three other reactions were conducted between 2b and the sterically hindered o‐toluidine (entries 23–25). Reaction of this amine and 2b catalyzed by 10 mol% Pd(OAc)₂/15 mol% L2 gave an 86% product yield, versus 73% with 5 mol% Pd(OAc)₂/7.5 mol% L2, both in 1 h. These yields did not change appreciably when the reaction was conducted over 2 h (entries 23 and 24, yields in parentheses). On the other hand use of 10 mol% Pd(OAc)₂/15 mol% L3 gave an excellent 92% yield (entry 25). A review of the data indicates that 10 mol% Pd(OAc)₂/15 mol% L3 gave consistently good results for reactions of both 2a and 2b.

Thus, the combination of Pd(OAc)₂/L3/Cs₂CO₃ in PhMe, at 100 °C is generally applicable for amination of 6‐bromo‐ and 6‐chloropurine nucleosides. However, a lowered catalytic load can be used for the former. We also compared reactions of acetate‐protected derivatives 1c and 2c using the best conditions identified for reactions of the bromo‐ and chloronucleosides (entries 26 and 27). In both cases, although reasonable product yields were attained, the results were not superior to those involving the silyl‐protected derivatives. Finally we tested 5 mol% Pd(PPh₃)₄ under otherwise identical conditions. This Pd(0) catalyst has been used for Suzuki–Miyaura cross‐coupling reactions of nucleosides.23,25 Both 1a and 2a reacted with p‐toluidine, but these reactions (entries 28 and 29) were inferior to the low catalyst loading experiments with L2 and L3. On the basis of these results, we then evaluated the amination reactions using the silylated nucleoside substrates and several amines with varying electronic and steric properties. These results are displayed in Table 2.

Table 2.

The publisher did not receive permission from the copyright owner to include this object in this version of this product. Please refer either to the publisher's own online version of this product or the printed product where one exists.

From the results in Table 2, it is evident that both bromo‐ (1a, b) and chloronucleosides (2a, b) are effective substrates for amination with a variety of arylamines. For reactions of the bromonucleosides 1a and 1b, conditions A [5 mol% Pd(OAc)₂/7.5 mol% L3] are reasonable for efficient reactions, and good‐to‐excellent yields were observed. Only the reaction of o‐toluidine with 1a gave <70% yield (entry 13). On the other hand, all reactions with o‐anisidine gave satisfactory results (entries 17–20). The electron‐deficient p‐acetylaniline reacted very well (entries 5–8), and good results were also obtained with m‐cyanoaniline (entries 9–12).

Reactions of the chloronucleosides 2a and 2b under conditions B [10 mol% Pd(OAc)₂/15 mol% L3] gave good‐to‐excellent results. These data clearly indicate that such chloronucleosides are excellent coupling partners in Pd‐catalyzed C—N bond‐forming reactions.

When comparing these catalyzed transformations to the S_NAr chemistry of 6‐halopurine nucleosides, the following are notable. The Pd‐catalyzed reactions of 1b produced comparable yields to displacement reactions on 1c,12 but require 3 times less arylamine. Also, Pd‐catalyzed reactions with the chloronucleosides resulted in amination products, whereas such chloronucleosides were reported to be unreactive under S_NAr conditions.12

Arylation of nucleosides by aryl bromides had been challenging until our recent report.24 Thus, reactivity properties of nucleosides often do not follow known trends of simpler aryl systems. With the reactions of the simple arylamines completed, we next analyzed the use of nucleosides themselves as the arylamine in coupling reactions with the halonucleosides. Nucleoside dimers are of interest since they can be formed in vitro by nitrous acid‐induced DNA cross‐linking. Such a process could have a physiological role due to the presence of nitrite in cured meats.26,27 Pd‐catalyzed approaches have been used for the non‐biomimetic synthesis of adducts produced by DNA cross‐linking.28,29 Among the various nucleoside dimers, deoxyguanosine‐deoxyguanosine (dG–dG), deoxyadenosine‐deoxyguanosine (dA–dG), and deoxyadenosine‐deoxyadenosine (dA–dA), the synthesis of dA–dA posed particular challenges. For instance, reaction of 1a and 5a using several Pd/L2 complexes were reported to be low yielding, and thus use of the iodo analogue of 1a was necessary.29 On the basis of our results, we wanted to reassess the reaction of 1a with 5a and analyze reactivity of the chloro analogue 2a for synthesis of dA–dA dimer. We also wanted to compare these reactions to those leading to the adenosine‐adenosine (A–A, ribose) dimers (Scheme 1). The results of these experiments are shown in Table 3.

Scheme 1.

Application of the Pd(OAc)₂/Xantphos system for the synthesis of the adenine nucleoside dimers.

Table 3.

The publisher did not receive permission from the copyright owner to include this object in this version of this product. Please refer either to the publisher's own online version of this product or the printed product where one exists.

From Table 3 it becomes clear that 6‐bromonucleosides 1a and 1b are good substrates for Pd‐catalyzed amination reactions with the more complex amines 5a and 5b. Good yields were obtained with 5 mol% Pd(OAc)₂/7.5 mol% L3 (entries 1 and 2), and some improvement was attained with the increased catalyst load of 10 mol% Pd(OAc)₂/15 mol% L3 (entries 3 and 4). More interestingly, chloropurine analogues 2a and 2b are also highly effective in these reactions and provide comparable product yields to the bromopurine nucleoside derivatives at the higher catalyst loading of 10 mol% Pd(OAc)₂/15 mol% L3. Thus, contrary to prior results,29 our data indicate that it is unnecessary to involve the iodo nucleosides in Pd‐catalyzed syntheses of nucleoside dimers.

Since modified nucleosides could possess interesting biological properties, the amination products were desilylated (see the Supporting Information for details, and Table S1). The deprotected compounds from this study as well as those obtained from previously synthesized compounds11,23,30 were subjected to a wide variety of antiviral assays. Unfortunately, none showed specifically interesting activity at subtoxic concentrations as shown in Tables S2–S5 of the Supporting Information. Instead, several compounds showed a mild cytostatic and/or cytotoxic activity against a variety of cell lines (CC₅₀ or MIC in the 10 to 100 μg mL⁻¹ range). Only the desilylated product from 4f was endowed with a marked cytostatic activity against CRFK cells (CC₅₀: 4.32 μM). This compound should be further explored for its potential antiproliferative activity against a broad variety of tumor cell lines. Pd‐catalyzed C—N and C—C bond‐forming reactions have led to this diverse set of compounds for antiviral and cytostatic analysis that would otherwise not be available.

Conclusions

In conclusion, Pd‐catalyzed C—N bond formation is an effective method for access to N ⁶‐aryl adenosine analogues. Herein we show that 5 mol% Pd(OAc)₂/7.5 mol% Xantphos is a generally effective catalyst in combination with Cs₂CO₃ in PhMe at 100 °C for reactions of 6‐bromopurine ribonucleosides as well as the 2′‐deoxy analogues. Remarkably, the corresponding 6‐chloropurine nucleosides are also excellent substrates for such reactions, requiring 10 mol% Pd(OAc)₂/15 mol% Xantphos, under otherwise comparable conditions. Finally, the method has been applied to the synthesis of deoxyadenosine‐deoxyadenosine and adenosine‐adenosine dimers. Bromo‐ as well as chloropurine ribo‐ and deoxyribonucleosides are effective substrates for reactions with protected 2′‐deoxyadenosine and adenosine. Therefore, the methodology described appears to present broad generality and scope for nucleoside modification with both 6‐bromo and 6‐chloropurine nucleosides.

Experimental Section

Thin layer chromatography was performed on 250 μm silica plates and column chromatographic purifications were performed on 200–300 mesh silica gel. Ligands L1–L5 and Pd(OAc)₂ were purchased from commercial suppliers. All other reagents were obtained from commercial sources and were used without further purification. Toluene was distilled over Na. Nucleoside substrates were prepared as described previously. Characterization data for all compounds are provided in the Supporting Information. ¹H NMR spectra were recorded at 500 MHz in the solvents indicated (when CDCl₃ was used, it was deacidified by percolating the solvent through a bed of solid NaHCO₃ and basic alumina). All proton spectra are referenced to residual protonated solvent resonance. No attempt has been made to ascertain the position of the arylamino proton (NHAr) in the products. Therefore, this and the aromatic protons are collectively assigned as Ar‐H. The sugar protons are numbered 1′–5′ beginning at the anomeric carbon and proceeding via the carbon chain to the primary carbinol carbon.

C—N Bond‐Formation under Conditions A

Pd(OAc)₂ (5 mol%), Xantphos (7.5 mol%) and Cs₂CO₃ (1.5 molar equiv.) were premixed in a small volume of toluene in an oven‐dried, screw‐cap vial equipped with a stirring bar. Bromonucleoside (1 molar equiv., 0.08 mmol of 1a or 1b), arylamine (2 molar equiv.) and toluene were then added to the vial so that the final reaction mixture was 0.1 M in nucleoside. The vial was flushed with nitrogen gas, sealed with a Teflon‐lined cap, and placed in a sand bath that was maintained at 101–102 °C. Reactions were monitored by TLC. Upon completion, the mixtures were diluted with Et₂O (7 mL) and washed with brine (5 mL). The organic phase was separated, dried (Na₂SO₄) and concentrated. Products were purified by column chromatography on silica gel using appropriate solvents (listed under individual headings), evaporated and finally dried under high vacuum to remove traces of solvent.

C—N Bond‐Formation under Conditions B

Pd(OAc)₂ (10 mol%), Xantphos (15 mol%) and Cs₂CO₃ (1.5 molar equiv.) were premixed in a small volume of toluene in an oven‐dried, screw‐cap vial equipped with a stirring bar. Chloronucleoside (1 molar equiv., 0.08 mmol of 2a or 0.085 mmol of 2b), amine (2 molar equiv.) and toluene were then added to the vial so that the final reaction mixture was 0.1 M in nucleoside. The vial was flushed with nitrogen gas, sealed with a Teflon‐lined cap, and placed in a sand bath that was maintained at 101–102 °C. Reactions were monitored by TLC. Upon completion, work‐up as above and column chromatography on silica gel using appropriate solvents yielded the amination products.

Biological Assays

The antiviral assays, other than the anti‐HIV assays, were based on inhibition of virus‐induced cytopathicity in HEL [herpes simplex virus type 1 (HSV‐1) (KOS), HSV‐2 (G), vaccinia virus, vesicular stomatitis virus, cytomegalovirus (HCMV) and varicella‐zoster virus (VZV)], Vero (parainfluenza‐3, reovirus‐1, Sindbis virus and Coxsackie B4), HeLa (vesicular stomatitis virus, Coxsackie virus B4, and respiratory syncytial virus), Crandel–Rees feline kidney (CRFK) [feline coronavirus (FIPV) and feline herpes virus] or MDCK [influenza A (H1N1; H3N2) and influenza B] cell cultures. Confluent cell cultures (or nearly confluent for MDCK cells) in microtiter 96‐well plates were inoculated with 100 CCID₅₀ of virus (1 CCID₅₀ being the virus dose to infect 50% of the cell cultures) in the presence of varying concentrations (100, 20, 4, ⬬ μmL^-1) of the test compounds. Viral cytopathicity was recorded as soon as it reached completion in the control virus‐infected cell cultures that were not treated with the test compounds. The minimal cytotoxic concentration (MCC) of the compounds was defined as the compound concentration that caused a microscopically visible alteration of cell morphology. The CC₅₀ was defined as the compound concentration that resulted in a 50% reduction of the MTT‐ or MTS‐directed (blue) viability staining of the (CRFK and MDCK) cell cultures. The methodology of the anti‐HIV assays was as follows: human lymphocyte MT‐4 cells (ca. 3×10⁵ cells/mL) were infected with 100 CCID₅₀ of HIV(IIIB) or HIV‐2(ROD)/mL and seeded in 200 μL wells of a microtiter plate containing appropriate dilutions of the test compounds. After 4 days of incubation at 37 °C, virus‐induced cytopathicity was recorded with the MTT dye staining method.

Supporting information

Detailed facts of importance to specialist readers are published as ”Supporting Information”. Such documents are peer‐reviewed, but not copy‐edited or typeset. They are made available as submitted by the authors.

miscellaneous_information