Modification of Purine and Pyrimidine Nucleosides by Direct C-H Bond Activation

Transition metal-catalyzed modifications of the activated heterocyclic bases of nucleosides as well as DNA or RNA fragments employing traditional cross-coupling methods have been well-established in nucleic acid chemistry. This review covers advances in the area of cross-coupling reactions in which nucleosides are functionalized via direct activation of the C8-H bond in purine and the C5-H or C6-H bond in uracil bases. The review focuses on Pd/Cu-catalyzed couplings between unactivated nucleoside bases with aryl halides. It also discusses cross-dehydrogenative arylations and alkenylations as well as other reactions used for modification of nucleoside bases that avoid the use of organometallic precursors and involve direct C-H bond activation in at least one substrate. The scope and efficiency of these coupling reactions along with some mechanistic considerations are discussed.


Introduction
Transition metal catalyzed traditional cross-coupling reactions have contributed significantly to the formation of new carbon-carbon bonds and to the synthesis of biaryl compounds. With few exceptions, the traditional Pd-catalyzed coupling reactions require two activated substrates, one is the OPEN ACCESS organometallic, alkene (Heck reaction), or terminal alkyne (Sonogashira reaction) and the other is the halide or triflate [1,2]. The most often used Stille, Suzuki, Negishi, Kumada and Hiyama reactions need an organometallic (Sn, B, Zn, Mg, and Si) component and a halide or pseudohalide. Owing to the high impact of these reactions in organic synthesis, natural product synthesis and pharmaceutical applications, the 2010 Nobel Prize in Chemistry was awarded jointly to Richard F. Heck, Ei-ichi Negishi and Akira Suzuki [3]. Pd-catalyzed cross-coupling reactions are carried out under mild conditions and can be performed in the presence of most functional groups. The mechanisms in most cases follow three major steps of: (i) oxidative addition, (ii) transmetallation, and (iii) and reductive elimination [1,2].
Transition metal-catalyzed cross-coupling reactions which are based on direct C-H functionalization have been recently developed [4][5][6][7][8][9]. These methodologies, which eliminate the use of organometallic substrates, compete with traditional Pd-catalyzed cross-couplings in the development of new strategies for the formation of carbon-carbon bonds. These reactions require only one activated substrate (C-H activation) and sometimes even no activation is required for either substrate (double C-H activation). They are atom efficient and avoid the synthesis of often unstable activated substrates. Major challenges associated with C-H functionalization reactions include: (i) the need for developing regioselective activation of specific C-H bonds in the presence of other C-H bonds; (ii) low chemoselectivity which means it is necessary to protect sensitive functional groups before performing the coupling; and (iii) the necessity to work at high temperature needed to activate C-H bonds with intrinsic low activity, which often causes decomposition of the substrates. Pd and Cu are two of the most common transition metal catalyst used for the C-H functionalization.
Transition metal-catalyzed approaches towards the synthesis of base-modified nucleosides can be divided into five major categories as depicted in Figure 1. The first two approaches are based on cross-couplings between two activated components. One involves reactions between metal-activated nucleoside bases and halides ( Figure 1, Path a) while the second employs couplings between halo (or triflate) modified nucleoside bases and organometallics (Path b). These approaches were extensively reviewed [10][11][12] and are not discussed in this account. The next two approaches are based on cross-couplings between only one activated component and require C-H activation at the second substrate. One involves reactions between C-H activated bond in nucleoside bases and halides (Path c), while the second employs couplings between halo-modified nucleoside bases and arenes, which, in turn, require selective C-H activation (Path d). The last approach involves cross-couplings between two inactivated substrates [cross-dehydrogenative coupling (CDC) reactions; Path e]. Direct C-H functionalization approaches (Paths c-e) alleviate some drawbacks associated with the synthesis of modified nucleosides employed in traditional Pd-catalyzed cross-coupling reactions (Paths a-b). They also avoid usage of the toxic organotin components, which are problematic during biological studies, or the sometimes unstable organoboronic substrates. Numerous C5 or C6 modified pyrimidine nucleosides and C2 or C8 modified purine nucleosides have been synthesized in last 40 years employing the transition-metal assisted cross-coupling reactions [10]. Some of them show potent biological activity and/or are utilized as mechanistic or labelling probes ( Figure 2). For example, the (E)-5-(2-bromovinyl)-2′-deoxyuridine (1, BVDU) has been found to be a highly potent and selective anti-herpes agent [13]. The bicyclic furanopyrimidine-2-one nucleoside analogues bearing an aryl side chain 2 display remarkable antiviral potency against the Varicella-Zoster virus [14]. The 5-thienyl-3 or 5-furyluridine 4 were used as molecular beacons for oligonucleotide labeling [15][16][17][18]. The 8-pyrenyl-2'-deoxyguanosine 5 serves as a probe for the spectroscopic study of the reductive electron transfer through DNA [19,20]. Furthermore, the 8-vinyl and 8-ethynyladenosines 6 show cytotoxic activity against tumor cell lines [21], while oligodeoxynucleotides modified with the 8-alkynyl-dG possess thrombin inhibitory activity [22].

Cross-Coupling of Adenine Nucleosides with Aryl Halides
Hocek and coworkers reported the first example of direct arylation of adenosine 7 with aryl halides by selective activation of the C8-H bond which gave access to 8-arylated adenosine analogues 9. The cross-coupling occurred in the presence of a stoichiometric amount of CuI (3 equiv.) and a catalytic load of Pd(OAc)2 (5 mol %) in DMF at elevated temperature (100 °C/22 h or 150 °C/5 h) to produce 9 in 50%-68% yields (Scheme 1, Table 1 entry 1) [23]. The authors were able to improve the coupling conditions (e.g., shortening reaction time and lowering the reaction temperature), as compared to their earlier work on C8-H arylation of purines and adenines [24,25] (vide infra), by addition of piperidine to the reaction mixture. They assumed [23] that formation of dimethylamine, as a side product of the prolonged heating of the DMF solvent during the C8-H arylation of purines, favorable influenced the rate of the arylation reaction, which is consistent with Fairlamb's findings [26,27]. Consequently, they found that the addition of higher boiling secondary amine such as piperidine (4 equiv.) was beneficial to the coupling reactions. Couplings of 7 with aryl iodines also produced N6,8-diarylated byproducts 11 in 12%-18% yield, whereas only 8-arylated products 9 were isolated when less reactive aryl bromine were employed. Scheme 1. Pd-catalyzed direct C8-H arylation of adenosine 7 and 2'-deoxyadenosine 8 with aryl halides. When 2'-deoxyadenosine 8 was subjected to this direct arylation protocol desired 8-arylated products 10 were produced only after the temperature was lowered to 125 °C (31% after 5 h; entry 2). It is worth noting that this protocol was applicable to unprotected nucleosides and allowed for the first time the single-step introduction of the aryl group at the C8 position without the need to (i) halogenate nucleoside substrates, or (ii) use expensive arylboronic acids or toxic arylstannanes [10].
Fairlamb and coworkers also demonstrated that a combination of a stoichiometric amount of Cs2CO3 with a substoichiometric amount of piperidine provided the best yield for Pd/Cu-mediated C8 arylation of 2'-deoxyadenosine 8 with various aryl halides (80 °C, 15 h) to give 10 in 32%-95% yields ( Table 4 entry 5) [26]. They also noted that sequential direct arylation of 8 with iodo(bromo)benzene followed by Suzuki-Miyaura cross-coupling of the resulting 8-bromophenyl-2'-deoxyadenosine gave convenient access to the new class of rigid organofluorescent nucleosides (RONs) analogues [29]. The arylation conditions were also extended to the adenosine analogues modified at either the ribose or the adenine moieties. Thus, 2'-deoxy-2'-fluoroadenosine 13 gave the 8-arylated product 14 almost quantitatively (94% isolated yield) under similar conditions; probably because the 2'-fluoro substituent is known to increase the stability of the N-glycosylic bond and to favor the syn conformation (Scheme 2). Coupling of 2-fluoro-2'-deoxyadenosine 15 with iodobenzene also effected 8-arylation concomitant with the displacement of fluorine by piperidine to give the 2,8-disubstituted 2'-deoxyadenosine 16 (Scheme 3) [26]. The chemistry of these couplings has been discussed in recent reviews [11,30].  These Pd-catalyzed/Cu-mediated methodologies were successfully applied to the synthesis of numerous 8-arylated purines and adenines. In 2006, Hocek and coworkers elaborated the original protocol for the efficient direct C8-H arylation of 6-phenylpurine analogues 17 using aryl iodides in the presence of Cs2CO3 and CuI to give 18 (Scheme 4) [24]. This route required prolonged heating at high temperature (160 °C/60 h) in DMF. Furthermore, it was essential to perform the coupling with strict exclusion of air to avoid formation of two byproducts, sometimes, in substantial yields (6%-54%). One byproduct was 19, which was formed by double arylation at C8 position and the ortho position of the phenyl ring at C6. The other byproduct was the 8,8'-bispurine dimer. Various 6,8,9-trisubstituted and 2,6,8,9-tetrasubstituted purine analogues were synthesized using this approach in combination with Suzuki cross-coupling reaction and Cu-catalyzed N-arylation at 9 position [24,25]. These conditions (Cs2CO3 or piperidine) were successfully employed for the synthesis of 8,9-disubstituted adenines but 6-N-(di)arylated byproducts were observed [24,25]. This protocol was also applied for direct C8-H arylation of adenines anchored to solid phase via 6-N amino group (in the presence of piperidine as base) [31]. Alami and coworkers developed a microwave-assisted direct C8 arylation of free-(NH2) 9-N-protected adenine 20 with aryl halides catalyzed by Pd(OH)2/C (Pearlman's catalyst) in the presence of CuI (Scheme 5) [32]. Reaction took only 15 min at 160 °C in NMP solvent when Cs2CO3 was used as base to give 21 in up to 90% yield. The application of Pd(OH)2/C catalyst allowed coupling with aryl bromides and even less reactive aryl chlorides [33]. Sequential combination of C8-arylation with ArCl and 6-N-arylation with ArBr or ArI (using Xantphos [34] instead of CuI) provides access to disubstituted adenines 22 [33].
Fairlamb and coworkers reported a detailed mechanism for the direct C8-arylation of adenine ring with aryl halides mediated by Pd and Cu in the presence of Cs2CO3 [26,27]. The authors noted that the use of a stoichiometric amount of Cu(I) is key to the direct arylation of the adenine ring and that the process parallels the arylation of imidazole ring at the 2 position [35]. As depicted in Scheme 6, Cu(I) was proposed to assist the C-H functionalization process by an initial coordination to the adenine N7 atom. The subsequent base-assisted deprotonation leads to the formation of 8-cuprioadenine intermediate A or N-heterocyclic carbene like cuprates, which can then undergo a standard Pd(0) catalytic cycle for cross-coupling with aryl halides. This process resembles Sonogashira's reaction between alkynylcuprates and halides [26,27]. The requirement for excess of CuI was attributed to the high binding affinity of Cu(I) for both the substrate and presumably the 8-arylated product(s). The dinucleoside copper(I) complex between 7-N and 6-NH atoms of the adenine have been identified as important intermediate [26].

Cross-Coupling of Inosine and Guanine Nucleosides with Aryl Halides
Guanosine 23 was found to be a poor substrate for the direct C8 arylation as indicated by the low yield (15%) of 8-phenylated product 25 under the conditions (120 °C) which were effective for adenosine analogues (Scheme 7) [26]. Analogous arylation of 2'-deoxyguanosine 24 at lower temperature (80 °C) yielded product 26 but only in 6% yield. The authors hypothesized that in the case of guanine substrates, Cu I -coordination most probably occurs at sites distal to C8 hampering efficient arylation. A similar inhibitory effect, associated with the ionizable protons in the guanine moiety was observed during the Suzuki couplings with 8-haloguanine nucleosides [36]. The authors also suggested that guanine-type nucleosides are poor substrates for direct C8-H arylation due to the lack of the "templating" role of exocyclic 6-amino group present in adenine nucleosides. The Pd-catalyzed/Cu-mediated direct C8-H arylation of inosine 27 proceeded proficiently to afford 8-phenylated product 29 in good yield (60%) at 120 °C (Scheme 8) [26]. The analogous functionalization of 2'-deoxyinosine 28, due to the stability of the glycosylic bond, had to be carried out at lower temperature to give product 30 but in only 19% yield. Recently, Pé rez and coworkers synthesized the 8-arylated inosine analogues via a microwave-assisted Pd/Cu-catalyzed direct C8-H arylation [37]. In order to increase the solubility of the nucleoside substrate, 2',3'-O-isopropylideneinosine 29 was employed to couple with iodopyridines or aryl iodides, by adopting Fairlamb's protocol [26,29], to produce 30 in only 1 h at 120 °C (Scheme 9).
The five-, six-or seven-membered e-fused purines 39 have been prepared by the intramolecular double C-H activation (at C8 and ortho position of the phenyl ring) of 9-N-phenylalkylpurines 38 in the presence of Pd catalyst and silver salt oxidant in high to excellent yields (Scheme 12) [42]. The seven-, eight-, or nine-membered e-fused purines of type 40 were prepared by the one-pot coupling of 38 with iodobenzene [42]. The reaction sequence was believed to be initiated by direct intermolecular C8-H arylation of 38 with iodobenzene followed the intramolecular cross-dehydrogenative-arylation between two phenyl rings to give product 40. Scheme 12. Pd-catalyzed intramolecular cross-dehydrogenative arylation of 9-N-substituted purines via C8-H bond activation.

Miscellaneous Direct C8-H Functionalizations
You and co-workers reported intermolecular Pd/Cu-catalyzed regioselective C8-H cross-coupling of 1,3-diethyl xanthine 41 (R 1 = R 2 = Et, R 3 = H) with electron-rich furans 42a and thiophenes 42b [43]. For example coupling of 41 with 2-methylthiophene in the presence of catalytic amount of the copper salt gave diheteroarene product of type 43 in 96%, indicating the tolerance of the free NH group at the 9 position to the reaction conditions (Scheme 13). The differences in the electron density of two heteroarene components was believed to facilitate the reactivity and selectivity in the two metalation steps [44] of the catalytic cycle of this cross-dehydrogenative arylation reaction. Scheme 13. Pd/Cu-catalyzed cross-dehydrogenative arylation of purines with heteroarenes.

N1-Directed Modifications of C6-Substituted Purine Nucleosides via ortho C-H Bond Activation
The Cu-catalyzed direct C-H activation/intramolecular amination reaction of the 2',3',5'-tri-O-acetyl-6-N-aryladenosines 47 were employed for the synthesis of fluorescent polycyclic purine and purine nucleosides of type 48 (Scheme 16) [52]. It was found that addition of Ac2O significantly improved the reaction rate (2 h at 80 °C) when Cu(OTf)2 (5 mol %) was used as copper source and PhI(OAc)2 was used as the oxidant. The 6-N-aryladenosines substrates 47 containing electron-withdrawing groups in the benzene ring gave better yields (~85%-92%) than those bearing electron donating groups (~45%-62%). The proposed catalytic cycle involves initial coordination of Cu Qu and coworkers developed a Pd-catalyzed strategy for the regioselective ortho monophenylation of 6-arylpurine nucleosides (e.g., 49) via (N1 purine nitrogen atom)-directed C−H activation using a large excess (30 equiv.) of iodobenzene (Scheme 17) [53]. It was essential to perform the reaction under an inert N2 atmosphere at 120 °C in the presence of AcOH in order to synthesize 50 in excellent yield (85%). It was believed that AcOH might help (i) to overcome the poisoning of the catalyst attributed to the multiple nitrogens present in purine ring, and (ii) to facilitate the reductive elimination step. However, the phenylation reaction was not successful when PhI was replaced with PhBr or PhCl.
It is noteworthy that the conditions described by Qu and coworkers (Pd(OAc)2/AgOAc/AcOH) are selective for the arylation at the ortho site of the C6-phenyl ring, while the CuI-catalyzed coupling of 49 with aryl halides in the presence of Pd(OAc)2/piperidine produced exclusively C8-arylated product 51 [23], or a mixture of double arylated products 50/51 in addition to 8,8'-dimers when analogous purine substrates 49 and Pd(OAc)2/Cs2CO3 were used [24,25]. Lakshman and coworkers reported direct arylation of 6-arylpurine nucleosides of type 52 by the Ru-catalyzed C-H bond activation (Scheme 18) [54]. The coupling usually required only 2 equiv. of aryl halides and K2CO3 or Cs2CO3 as a base to give mixture of the mono and diarylated ortho products 53 and 54 in ratio of approximately of 2-7 to 1 with no 8-arylated byproducts detected. It is noteworthy that these conditions were applicable to acid-sensitive substrates such as 2'-deoxynucleosides as opposed to the previously described Pd-catalyzed arylation promoted by AcOH [53]. Both aryl iodides and bromides gave arylated products in good yield, but aryl iodides proceeded with higher yields and better product ratio. Also, aryl halides bearing EWG gave higher yields compared to the ones bearing EDG.
A possible mechanism for this direct ortho-arylation was proposed involving purinyl N1-directed electrophilic attack by the aryl/Ru IV complex A on the ortho position of the C6-aryl ring atom of the substrate 52 (Scheme 19) [54]. Subsequent reductive-elimination of the five-membered Ru complex B gave product 53. The 2-amino-6-arylpurine nucleosides were unreactive, indicating that the presence of C2-amino group is critical and often inhibits the reactivity of purine nucleoside towards C-H activation reactions.
The Pd-catalyzed C-H bond activation and oxidation of the silyl protected C6-aryl ribonucleosides of type 55, as well as the 2'-deoxy counterparts (e.g., 56), in the presence of PhI(OAc)2 as a stoichiometric oxidant in MeCN provided access to monoacetoxylated products 57 or 58 in good yields (Scheme 20) [55]. Increasing the loading of PhI(OAc)2 to 3 equiv gave mainly the diacetoxylated products 59 or 60. The involvement of the N1-purinyl atom in this N-directed C-H bond activation was demonstrated by isolation and crystallographic characterization of the dimeric Pd II -containing cyclopalladated C6 naphthylpurine derivative. The latter complex, together with PhI(OAc)2, was shown to be effective in catalyzing oxidation of substrate 56 to 58. Scheme 20. Pd-Catalyzed N1-directed ortho C-H acetoxylation of 6-arylpurine nucleosides.
Chang and coworkers reported Rh-catalyzed intermolecular amidation of 6-phenylpurine nucleosides with sulfonyl azides via purinyl N1-assisted C-H activation (Scheme 21) [56]. Amidation of 6-arylpurine nucleoside 61 proceeded smoothly and with excellent ortho-selectivity to afford product 63 in 70% yield. The presence of a free amino group at the C2 position of purine substrate 62 inhibits once more the coupling efficiency to give amidation product 64 in 45% yield. No glycosylic bond cleavage was noted under the optimized conditions. It is worth noting that the coupling conditions require no additional oxidant and release N2 as the only byproduct.

Cross-Coupling with Aryl Halides
Direct C-H arylation of the pyrimidine nucleosides is currently limited to uracil bases. The main challenge which needed to be overcome was a regioselective activation of the C5-H or C6-H bond of the uracil ring. Also lacking are efficient conditions which could be applicable to the natural uridine and 2'-deoxyuridine analogues. Hocek and coworkers reported the regioselective C-H arylations of 1,3-dimethyluracil (65, DMU, R = Me) with aryl halides. Thus, arylation in the presence of Pd(OAc)2 and Cs2CO3 mainly formed the 5-arylated uracil analogues 66, while coupling in the presence of Pd catalyst and CuI (3 equiv.) preferentially formed the 6-arylated derivatives 67 (Scheme 22) [57,58]. Interestingly, Cu-mediated arylation in the absence of Pd catalyst gave exclusively 6-aryluracils 67 albeit in lower yields. These couplings required high temperature (160 °C) and long time (48 h) and were not applicable to the synthesis of unsubstituted uracils. Also the electron-deficient aryl bromides were poor substrates. In order to prepare the unprotected uracil derivatives, these protocols were also applied to N-benzyl protected uracil derivatives. Subsequent debenzylation (e.g., 67, R = Bn) afforded efficiently 5-or 6-aryled uracil derivatives [57,58]. The experimental results indicated that different mechanisms are involved in these diverse arylation reactions [57,58].
Kim and coworkers reported the direct arylation of 1,3-dimetyluracil 65 with aryl bromides (including electron-deficient ones) in the presence of Pd(OAc)2, K2CO3 and PivOH (130 °C/12 h/DMF to give predominantly C5 arylated products 66 in up to 79% yield. A small amount of 6-arylated isomers 67 were also observed [59]. However, application of this condition to 1-(tetrahydrofuran-2-yl)-3benzyluracil 68 (a substrate having a glycosylic bond) resulted in severe decomposition. Nonetheless lowering the temperature to 100 °C (12 h) allowed the isolation of the 5-phenyluracil analogue 69 as the sole product in 55% yield (Scheme 23) [59]. The intramolecular C6-H arylation of the uracil derivatives bearing a Morita-Baylis-Hillman adduct at the N1 position in the presence of TBAB and Pd(OAc)2 provided a convenient access to the azepine scaffold [60]. Based on the literature reports, the mechanism for the direct arylation of uracil analogues (e.g., DMU, 65) with aryl halides at either the C5 or C6 position in the presence of Pd or Cu/(Pd) are summarized in Scheme 25. The coupling with only Pd(OAc)2 was proposed to proceed via an electrophilic metalation-deprotonation (EMD) mechanism [5,9] (path a, complex A), and thus follow the regioselectivity of substitution at the C5 position to give 66 [59]. The formation of the C5 arylated products was also suggested [57] to proceed via the concerted metalation-deprotonation (CMD) mechanism [9,62] (complex B). On the other hand, direct C-H arylation in the presence of CuI or in combination with Pd(OAc)2 was hypothesized to proceed via an Ullmann-type mechanism [27,63], which involves the formation of a carbanion by the abstraction of the most acidic C6-H [64][65][66] in the uracil ring (e.g., 65) with base (path b). Subsequent cupration, transmetallation and reductive elimination leading to the formation of C6 arylated uracil 67 [57,61]. Alternatively, addition of the copper center of the arylcopper(III) complex C to the more nucleophilic C5-position in 65 followed by either base-promoted anti-elimination or Heck-type syn β-hydrogen elimination has been also proposed for C6-arylation [61]. Moreover, Heck-type carbopalladation of the 5,6-double bond of 65 with ArPdX species has been also considered for the arylation of uracil ring with aryl halides [59].
The Pd-catalyzed cross-dehydrogenative heteroarylation between 1,3-dialkyluracils 77 and pyridine-N-oxides 78 (3 equiv.) substrates in the presence of Ag2CO3 at 140 °C for 12 h gave 5-(2-pyridyl-N-oxide)uracils 79 in good-to-high yields (Scheme 28) [70]. As expected, the 3-substituted pyridine-N-oxides gave products 83 with excellent regioselectivity at the less bulky site. The coupling, however, was not compatible with either 1,3-dibenzoyluracil or unprotected uracil substrates. Reduction of N-oxides 79 with PCl3 in toluene yielded the corresponding 5-uracil derivatives substituted with a 2-pyridyl ring. The electrophilic palladation at the C5 of uracil and the coordination of the palladium atom to N-oxide was believed to control the regioselectivity (at C5 of the uracil ring and C2 of pyridine oxide) of these double C-H activation cross-couplings.

Cross-Dehydrogenative Alkenylation at C5 Position
The discovery of the potent antiviral activity of E-5-(2-bromovinyl)-2'-deoxyuridine (1, BVDU) led to the exploration of synthetic routes based on the oxidative coupling of uridine nucleosides with alkenes. Such routes avoided the use of mercury that was central to Walker's synthetic approach involving the condensation of 5-mercurated 2'-deoxyuridine with methyl acrylate and radical decarboxylation-bromination sequence [71]. They also seem advantageous to other coupling protocols which employ coupling between 5-halouracil nucleosides and organometallics or methyl acrylate [10,72]. In 1987, Itahara reported the oxidative coupling of uracil nucleosides 80 with maleimides 81 which gave 5-substituted coupling products of type 82 (Scheme 29) [73]. Using stoichiometric amounts of Pd(OAc)2 was however necessary. The yields for uridine and 2'-deoxyuridine substrates were lower (4%-22%) than those of the DMU substrate (~20%-50%). The yields for DMU substrate slightly increased in the presence reoxidants such as AgOAc, Na2S2O8 or Cu(OAc)2.
Also in 1987, Hirota and coworkers reported the oxidative coupling of uridine 80b and 2'-deoxyuridine 80c with methyl acrylate or styrene using either stoichiometric amounts of Pd(OAc)2 in MeCN at ambient temperature or catalytic loading of Pd(OAc)2 in the presence of tert-butyl perbenzoate as the reoxidant to generate 5-vinyl uridine analogues 84 (Scheme 30) [74]. The couplings proceeded stereoselectively to give the trans isomers. The reaction conditions were compatible with unprotected and protected uridine substrates, though, coupling of 2',3'-di-O-isopropylideneuridine with methyl acrylate produced also the 5',6-cyclouridine byproduct in 23% yield. Yun and Georg recently reported Pd-catalyzed cross-dehydrogenative coupling of 1,3-disubstituted uracils as well as protected uridine 85 and 2'-deoxyuridine derivatives 88 with tert-butyl acrylate in the presence of AgOAc and PivOH in DMF at 60 °C/24 h to give 5-alkenyl products 87 and 89 in 66% and 75% yield, respectively (Scheme 31) [75]. The coupling occurred with the regio-(C5) and stereoselectivity (E-isomer). However, 3-N-methyl protection at the uracil ring was necessary. The coupling was postulated to occur via the electrophilic palladation pathway [76] at the C5 position of the uracil ring followed by deprotonation with the pivalate anion to give palladated intermediate. Coordination with the alkenes via transmetallation, and subsequent β-elimination provided 5-alkenyluracil derivatives.

Miscellaneous Direct C-H Functionalizations
Pd-catalyzed direct C-H acetoxylation at the electron-rich C5 position of uracil nucleosides with PhI(OAc)2 under reasonably mild conditions (60 °C, 3-5 h) have been also developed (Scheme 32) [77]. The acyl protected uridine 80a and silyl protected 2'-deoxyuridine 80d were compatible with these conditions to give the corresponding 5-acetoxy products 90 in 55% and 25% yields, respectively. The reaction was proposed to proceed via oxidative electrophilic palladation at the electron rich C5 position to give the 5-palladauracil intermediate followed by oxidation to give Pd(IV) intermediate, which yielded 5-acetoxyluridine via the reductive elimination of Pd(II).
The Cu-catalyzed intermolecular C5-H amination of DMU 65 with 4-bromoaniline in the presence of [hydroxy(tosyloxy)iodo]mesitylene has been recently developed (Scheme 33) [50]. The regioselectivity of this C5 amination was proposed to be controlled via the formation of iodonium salt intermediate 91, which is consistent with C5 electrophilic aromatic substitution that is typical for uracil ring. Subsequent Cu-catalyzed amination gave the 5-amino product 92 in 65% overall yield.  [78]. This coupling between Csp2-H (in the uracil ring) and Csp3-H (in the side chain) bonds was not, however, successful when R 2 = H. The authors suggested that coupling most probably involved single electron transfer (SET) processes and might require a more stable tertiary radical on the side chain to proceed. Scheme 34. Intramolecular cross-dehydrogenative cyclization at C6 of uracil ring.
Recently, C5-H trifluoromethylation of DMU 65 by means of electrophilic, nucleophilic or radical "CF3" species in an effort to synthesize 6-aryl-5-(trifluoromethyl)uracils (e.g, 96) by direct activations of both C5-H and C6-H bonds in consecutive manner has been attempted (Scheme 35) [79]. Thus, reactions of 65 with electrophilic (Umemoto's or Togni's) or nucleophilic (Rupert's) reagents in combination with Pd or Cu catalyst either failed or led to the formation of 5,5-or 5,6-dimeric products or 5-CF3 product in low yields. However, radical trifluoromethylation of 65 with sodium trifluoromethanesulfinate in the presence of tert-butyl hydroperoxide provided 1,3-dimethyl-5-(trifluoromethyl)uracil 95 in 67% yield [79]. The subsequent direct arylation at C6-H of 95 with 4-iodotoluene in the presence of Pd(OAc)2 and CuI/CsF afforded desired 5,6-disubstituted uracils 96 albeit in low yield (25%); probably because of the electron-withdrawing effect of the CF3 group at the 5 position on the pyrimidine ring. This C6-H arylation was not applicable to other aryl halides and often was accompanied by cleavage of the CF3 group due to hydrolysis followed by decarboxylation, especially when Cs2CO3 was used as the base.

Coupling of 5-Halouracil Nucleosides with Arenes and Heteroarenes
The lack of regioselectivity in direct activation of uracil derivatives (C5-H vs. C6-H) during cross-couplings with aryl halides, and fact that coupling conditions are usually unsuitable for unprotected uracils and natural nucleosides, were recently overcome by switching the halide substituents from aryl halides to uracil ring and allowing to react of 5-halouridines with arenes instead. Wnuk and coworkers found that the 5-iodouracil nucleosides 97 coupled with simple arenes or heteroaromatics 98 in the presence of Pd2(dba)3 and TBAF in DMF under milder condition (100 °C/1-2 h) to give the 5-arylated uracil nucleosides 99 in high yields (Scheme 36) [80]. This TBAF-promoted protocol, which proceeded without the necessity of adding ligands and/or additives, worked efficiently with the natural uracils and uracil nucleosides, and was compatible with the stability of the glycosylic bond for 2'-deoxyuridine substrates (e.g., 97c). The 5-(2-furyl, or 2-thienyl, or 2-pyrrolyl)uridine derivatives 99, that are important RNA and DNA fluorescent probes [15][16][17][18], were synthesized in up to 98% yields without the necessity of using organometallic substrates. The arylation proceeded also when TBAF was replaced with Cs2CO3 base with or without the presence of PivOH [80]. The analogous coupling of 5-iodocytidines with arenes failed to afford 5-arylated products.
The fact that 3-N-methyl-5-iodouracil substrates, which lack the ability to tautomerize to the enol form, did not undergo these couplings with arenes indicates that the C4-alkoxide (enol form of uracil) may participate in the intramolecular processes of hydrogen abstraction as depicted in Figure 3. The mechanism pathway, after the initial oxidative-addition of palladium to C5-halogen bond, might involve electrophilic aromatic palladation assisted by C4-alkoxide (e.g., A) or direct proton abstraction assisted by C4-alkoxide (e.g., B) [80]. This would be in agreement with earlier finding [7,81] that direct arylation is facilitated by a Pd-coordinated carboxylate group, which also can assist in intramolecular proton abstraction.

Conclusions
The use of the Pd(cat.)/Cu(stoich) system in the presence of bases such as Cs2CO3 and/or piperidine in DMF effects direct regioselective arylation of purine nucleosides at the 8 position. Protected and deprotected adenosine, 2'-deoxyadenosine as well as inosine or guanosine derivatives coupled efficiently under these conditions with aryl halides to give access to 8-arylated products in up to 99% yields. The C8-H functionalization process is proposed to occur via 8-cupriopurine intermediates or N-heterocyclic carbene like cuprates, which subsequently can undergo a standard Pd(0)-catalytic cycle for cross-coupling with aryl halides. The cross-dehydrogenative arylation protocols which involve C8-H bond activation have been also developed for the synthesis of purine fused ring systems. Intramolecular N1-purinyl nitrogen directed C-H bond activation was utilized for direct ortho modification (arylation, amination or acetoxylation) of the aryl rings at C6 position of purine nucleosides.
In the case of pyrimidine nucleosides, direct arylation protocols have only been developed for uracil bases, which, in turn, usually require protection at the 1-N and 3-N positions. The biggest challenge how to overcome regioselective C-H activation at the C5 or C6 position of uracil ring have been accomplished by employing different catalysts and ligands. For the cross-coupling with aryl halides, it was found that Pd-based catalytic systems were effective to promote C5 arylation, whereas Pd/Cu or Cu-mediated systems affected C6 arylation with good selectivity. The C5 arylation is proposed to proceed via the electrophilic or concerted metalation-deprotonation mechanisms, while the C6 arylation likely to occur via a cuprate intermediates. Pd-catalyzed cross-dehydrogenative coupling of uracil nucleosides with alkenes or arenes, in the presence of oxidants, were developed to give convenient access to 5-alkenyl/aryl derivatives. The 5-acetoxy and 5-amino derivatives were also synthesized by direct activation of C5-H bond in uracil ring. No examples of couplings involving direct C-H activation of cytosine ring have been reported.
The transition-metal catalyzed syntheses of the modified nucleobases by direct C-H bond activations have been substantially improved in the past decade. However, despite improved coupling efficiency and availability of milder reaction conditions applicable to the less stable deoxynucleosides, application of the direct C-H functionalization approaches to nucleotides and/or short (deoxy)oligonucleotides fragments still will require developing the conditions compatible both with solvent requirements for water-soluble nucleotides and stability of phosphate esters.