Synthesis of N⁶ ,N⁶-Dialkyl Adenine Nucleosides With In Situ Formed Hexaalkylphosphorus Triamides

Abstract

Reactions between secondary amines and phosphorus trichloride (PCl₃) leads to the formation of the corresponding tris(dialkylamino)phosphines or hexaalkylphosphorus triamides [HAPT: (R₂N)₃P]. Reaction of silyl-protected 2′-deoxyinosine and acetyl-protected inosine with the in situ formed HAPT and iodine (I₂) leads to the formation of N⁶,N⁶-dialkyl adenosine and 2′-deoxyadenosine. In some cases the stoichiometry of the amine is important as also the use of a tertiary amine base. The effect of amine stoichiometry on the reaction of HAPT with I₂ has been studied by ³¹P{¹H} NMR.

Introduction

The physiological importance of adenine and modified adenine nucleosides cannot be overstated. Because of this high importance, the adenosine core has been the subject of structural modifications. The emergent compounds have been studied for their wide ranging effects such as modulation of A₁, A₂ and A₃ receptors that control important biofunctions,^[1] as antiviral,^[2] anticancer^[3] and antimalarial^[4] pharmacophores. The classical method for introducing modifications at the exocyclic amino group of adenine nucleosides is the S_NAr displacement, in which a leaving group at the 6-position of the purine is displaced with suitable amines.^[5]

One important limitation to the use of S_NAr displacement is the availability of suitable nucleoside derivatives, many of which involve non-trivial synthesis. The most significant problem in these processes is the cleavage of the labile glycosidic bond in these processes. Although C6 sulfonates can be prepared from hypoxanthine nucleosides, competing N1 and O6 sulfonylation occurs.^[6] In order to circumvent some of the problems associated with access to electrophilic nucleoside precursors, non-aqueous diazotization methods have been developed.^[7] Nevertheless, newer and simpler methods are needed for the synthesis of N-modified adenine nucleosides.

In this context, the combination of PPh₃/I₂/amine has been reported for the introduction of morpholinyl, piperidinyl and imidazolyl groups into the C6 position of purine nucleosides.^[8] This reaction presumably proceeds by formation of [(Ph₃P⁺I)I^-], which reacts at the amide carbonyl group of the hypoxanthine residue to afford a nucleoside phosphonium salt(Scheme 1). Expulsion of Ph₃PO from this intermediate by the amine then results in the N-modified adenine nucleoside. Along the same lines, the reaction of hypoxanthine nucleosides with 1H-benzotriazol-1-yloxy-tris(dimethylamino)phosphonium hexafluorophosphate (BOP) and amines has been used to synthesize N-substituted adenine nucleosides (Scheme 1). ^[9] An O-6 phosphonium salt was proposed as the intermediate in these transformations, which undergoes reaction with the amine by displacement of hexamethylphosphoramide (tris(dimethylamino)phosphine oxide or HMPA).^[9] We have recently analyzed the mechanism of reaction of the hypoxanthine core with BOP and have confirmed that in the absence of an amine, a nucleoside phosphonium salt is formed en route to the O⁶-(benzotriazol-1-yl) nucleoside derivatives.^[10] These new nucleoside derivatives are excellent substrates for modification at the C-6. The chemical understanding gained subsequently led to the development of a second generation synthesis of O⁶-(benzotriazol-1-yl)inosine and 2′-deoxyinosine using PPh₃/I₂/HOBt,^[11] as well as polymer-supported reagents for nucleoside modification.^[12] As a result of our interest in understanding whether other phosphines can be utilized for modification at the C-6 of purine nucleosides, we became involved in the work described herein.

Results and Discussion

In our initial experimentation using 3′,5′-bis-O-(t-butyldimethylsilyl)-2′-deoxyinosine and the PPh₃/I₂ combination, we quickly discovered that the products were contaminated by the Ph₃P=O and that separation was difficult. Scrutiny of the literature (Supporting Information to ref. 8b) indicated a similar problem. We expected that amelioration of this problem could be attained through the use of a polymer-supported phosphine (Figure 1). To our surprise polymer-supported triphenylphosphine from two different sources^[13] proved ineffective for the reaction. Therefore, we focussed attention on the family of the proazaphosphatranes which have proved to be interesting in other reactions.^[14] These compounds (Figure 1) also proved ineffective for the conversions with significant amounts of starting materials left in many instances.

Figure 1.

Reagents that were initially tested.

We then reasoned that increasing the ease of removal of the resultant phosphine oxide might allow circumvention of the problem. Tris(dimethylamino)phosphine (HMPT) was expected to not only display enhanced nucleophilicity in its reaction with I₂, but the formed HMPA should potentially be removable by washing with water. It was with this notion that we embarked on subsequent experimentation, the results of which are reported herein.

On the basis of the foregoing rationale, we initially subjected 3′,5′-bis-O-(t-butyldimethylsilyl)-2′-deoxyinosine (1a) to reaction with HMPT/I₂/morpholine/(iPr)₂NEt (DIPEA), in toluene at 90 °C (Scheme 2). This reaction proceeded smoothly and appeared clean by TLC. However, the ¹H NMR spectrum of the product from this reaction was interesting. The signals for the purine, saccharide and morpholinyl resonances were clearly visible. However, there was an additional broad resonance at δ 3.52 ppm. It was the presence of this signal that led us to doubt the purity of the product obtained.

Scheme 2.

Initial attempts at converting 1a to the C-6 morpholinyl and dimethylamino derivatives.

Since N,N-dimethyladenosine is formed in the reaction of 2′,3′,5′-tri-O-acetylinosine with HMPT and CX₄ (X = Br or Cl),^[15] we reasoned that the byproduct obtained in the reaction of 1a may be the N,N-dimethyl 2′-deoxyadenosine derivative 2a (Scheme 2). Competing formation of dimethylamides in HMPT mediated conversions of 2,2,2-trihaloethyl esters to amides has been reported in the literature.^[16] Hence, our next effort was the synthesis of 2a by simply replacing morpholine with Me₂NH. This reaction proceeded smoothly affording the N,N-dimethyl-2′-deoxyadenosine derivative 2a in good yield. Using this authentic 2a we were able to identify it the byproduct formed along with the morpholinyl derivative 2b in the initial reaction with morpholine (pure 2a displays a broad resonance at δ 3.52 ppm that sharpens upon heating to 45 °C).

The experiments above indicated that the combination of HMPT/I₂ could be used to produce effective activation of the C-6 amide linkage with no problems associated with product isolation and purification. However, a problem that remained to be solved was the competing formation of 2a. This reaction could potentially be more significant if amines with nucleophilicities lower than that of dimethylamine were to be used.

At this juncture we reasoned that specific tris(dialkylamino)phosphines or hexaalkylphosphorus triamides (HAPT) could potentially be prepared from individual amines and used in combination with I₂ for activation of the C6 amide carbonyl. The ensuing nucleoside phosphonium salts could then be subjected to subsequent displacement reactions with the amines used for forming the specific HAPT. In such an event, there would be only one possible amine nucleophile in the reaction.

In this context, HAPT derivatives such as tris(pyrrolidino)- and tris(piperidino)phosphines have been prepared by reaction of PCl₃ with pyrrolidine and piperidine, respectively.¹⁷ In our experiments, we decided to conduct the entire operation consisting of: (a) formation of the HAPT, (b) activation of the C-6 amide, and (c) the amination reaction, as a one-pot process without isolation of the individual tris(dialkylamino)phosphines. We chose secondary amines for the current work for two reasons; primary amines could potentially form polymeric products with PCl₃ and reactions of HAPT with primary amines could lead to iminophosphoranes [(R₂N)₃P=NR].^[18] With secondary amines these potentially complicating problems can be avoided. Finally, in initial experiments, we opted to use the secondary amine itself as base. Our overall logic is represented in Scheme 3. As can be seen from this scheme, a total 9 molar equiv of the amine would be needed per molar equiv of PCl₃.

Scheme 3.

Plausible synthesis of adenine nucleosides via in situ formation of HAPT [(R₂N)₃P].

On the basis of these considerations, the initial conditions we opted for were 2.6-3.0 molar equiv PCl₃/20-30 molar equiv 2° amine/2.2 molar equiv I₂ with toluene (PhMe) as solvent. Using these conditions, 3′,5′-bis-O-(t-butyldimethylsilyl)-2′-deoxyinosine (1a) and 2′,3′,5′-tri-O-acetylinosine (1b) were converted to a series of N,N-disubstituted adenine derivatives (Table 1).

Table 1.

Initial reactions of hydroxyl-protected 2′-deoxyinosine and inosine with in situ formed HAPT/I₂ and 2° amine.

Entry	Substrate	Amine	Substrate/PCl3/amine/I2, time	Product, yield[a]
1	1a		1:3:30:2.2 1 h	2a: 92%[b]
2	1b		1:2.6:22:2.2 2.5 h	3a: 81%[b]
3	1a		1:3:30:2.2 6 h	2b: 69%
4	1b		1:2.6:20:2.2 2 h	3b: 71%
5	1a		1:3:30:2.2 2 h	2c: 65%
6	1b		1:2.6:30:2.2 5 h at r.t. then 15 min at 90 °C	3c: 76%
7	1a		1:3:30:2.2 4 h	2d: 73%

^[a]Yield is of isolated, purified products.

^[b]Reaction using a 2 M solution of Me₂NH in THF.

Despite the simplicity of the reaction and the overall reasonable results obtained (Table 1), difficulties were faced in some reactions. For example, reaction involving pyrrolidine and 1b did not yield tractable product. This led us to question the underlying reasons, and one was the possible competitive formation of tetra(dialkylamino)phosphonium salts (Scheme 4). Such salts are described in the literature.^[19]

Scheme 4.

Plausible reactions of HAPT with I₂ and subsequent conversion to a tetra(dialkylamino)phosphonium salt.

We interpreted this unexpected difficulty as follows. When excesses of fairly nucleophilic amines are utilized, there can be two competing processes: (a) reaction of the nucleoside O6 with the iodo-HAPT intermediate formed from reaction of the HAPT with I₂, and (b) reaction of the amine with the iodo-HAPT intermediate produced. It is conceivable that in the present cases the tetra(dialkylamino)phosphonium species represents a dead end intermediate in the present cases that does not react at the O6 of the nucleoside.

In order to test this working hypothesis, we conducted two experiments. In one case, piperidine (6 molar equiv) was carefully added to a well-stirred solution of PCl₃ in anhydrous toluene at 0 °C. After bringing the mixture to room temperature and stirring for about 30 minutes, I₂ (1.2 mol equivalents) was added. The suspension was stirred at room temperature for about 23 h and then filtered. The filtrate, when concentrated to dryness and analyzed by ³¹P{¹H} NMR (Figure 2A), showed a predominant resonance at δ 26.55 ppm (relative to 85% H₃PO₄ as external reference). A second experiment was conducted in a similar manner with the exception that 30 molar equiv of piperidine was used. ³¹P{¹H} NMR analysis (Figure 2B) showed an entirely different resonance at δ 20.69 ppm.

Figure 2.

³¹P{¹H} NMR spectra (in CDCl₃) of the reaction mixture obtained with A: PCl₃ + 6 mol equiv piperidine + 1.2 mol equiv I₂ and B: PCl₃ + 30 mol equiv piperidine + 1.2 mol equiv I₂.

The results clearly indicated formation of a new phosphorus containing species when a large excess of piperidine was present. To further verify these results, another experiment was conducted (see Scheme above Figure 3). A toluene solution of commercially available bromotripyrrolidinophosphonium hexafluorophosphate (PyBroP) was exposed to 15 molar equiv of pyrrolidine for about 20 h. The reaction mixture was concentrated to dryness and analyzed by ³¹P{¹H} NMR. The initial ³¹P resonance of PyBroP at δ 27.93 ppm (Figure 3A) was replaced by new resonances, with the major one at δ 25.14 ppm, Figure 3B).

Figure 3.

³¹P{¹H} NMR spectra (in CDCl₃) of A: PyBroP and B: PyBroP + 15 molar equiv pyrrolidine. (The PF₆^- resonance is not shown, but it appears at δ −143.97 ppm in each case).

Having gained a better understanding of the underlying chemical processes involved in these transformations, we decided to evaluate reactions that only involved stoichiometric ratios of PCl₃ and the 2° amine (1:6 PCl₃/amine). This would in principle result in the formation of the HAPT and amine hydrochloride. Then, in order to prevent formation of the tetra(dialkylamino)phosphonium species, the reaction with the nucleoside was conducted in the presence of a 3° amine, (iPr)₂NEt. The 3° amine not only acts as proton sponge but can also liberate free 2° amine from the initially formed hydrochloride salt for the final displacement step. A brief analysis of solvents [toluene (PhMe), 1,2-dimethoxyethane (DME) and CH₂Cl₂] was also conducted at this stage in order to optimize the reactions. The results of these experiments are shown in Table 2.

Table 2.

Modified conditions tested for the synthesis of N,N-modified adenine nucleoside derivatives.

Entry	Substrate	Amine	Substrate/PCl3/Amine/I2/(iPr)2NEt molar equiv, Solvent, time at T °C	Product, yield[a]
1	1a		1:3:18:3:9, PhMe, 21 h at r.t.	2b: 71%
2	1b		1:3:18:3:9, PhMe, 24 h at r.t.	3b: 65%
3	1b		1:3:30:2.2:0, CH2Cl2, 21 h at r.t.	Inc[b]
4	1b		1:3:18:3:9, PhMe, 21 h at r.t.	3c: 62%
5	1b		1:3:18:3:9, DME, 18 h at r.t.	3d: 43%
6	1b		1:3:18:3:9, DME, 1 h at 85 °C then 4.5 h at r.t.	3d: 35%[c]

^[a]Where reported, yield is of isolated, purified product.

^[b]About 80% 1b remained.

^[c]About 20% 1b was still present at workup.

From Table 2 it becomes clear that in many cases, complete reaction can be attained in the presence of the 3° amine. As seen in entries 1, 2, 4 and 5, complete reaction with the nucleosides can be accomplished in the presence of (iPr)₂NEt but with only 18 molar equiv of the 2° amine. DME proved to be a good solvent for reaction of 1b with pyrrolidine (entry 5), a reaction that was generally problematic/low yielding.

As shown in Scheme 5, the imidazol-1-yl nucleoside 2e can also be synthesized from 1a using 3 molar equiv PCl₃/30 molar equiv imidazole/2.2 molar equiv I₂. The reaction of 1a, which was conducted in 1,2-dichloroethane (DCE) for better solubility of imidazole, proceeded to completion at room temperature in 72 h (63% yield). In contrast to this successful reaction of disilyl derivative 1a, reaction of triacetate 1b was unsuccessful. We then investigated the use of Et₂NH as the nucleophile (Scheme 5). In this case, surprisingly, successful reaction was observed with the triacetate 1b (82% yield of 3e) but reaction with the disilyl derivative 1a was unsuccessful. In contrast, O⁶-(benzotriazol-1-yl)-3′,5′-bis-O-(t-butyldimethylsilyl)-2′-deoxyinosine underwent smooth reaction with Et₂NH leading to the diethylamino product.^[20]

Scheme 5.

Synthesis of the C-6 imizadol-1-yl and diethylamino derivatives. Conditions (a): 3 molar equiv PCl₃/30 molar equiv imidazole/2.2 molar equiv I₂, DCE, r.t.; (b) 3 molar equiv PCl₃/30 molar equiv Et₂NH/2.2 molar equiv I₂, PhMe, r.t.; (c) 3 molar equiv PCl₃/30 molar equiv imidazole/2.2 molar equiv I₂, DCE, 90 °C; (d) 3 molar equiv PCl₃/30 molar equiv Et₂NH/2.2 molar equiv I₂, DCE, 90 °C.

In order to determine what role if any the 2′-substituent plays, we conducted two additional experiments on the trisilyl-protected inosine 1c (Scheme 5). We chose to conduct experiments using imidazole (with which 1b did not yield any product) and Et₂NH (with which 1a did not yield any product). In these reactions, the imidazolyl product 4a was obtained in 47% yield within 36, h and the diethylamino product 4b was obtained in 81% yield within 8 hours. Although we are unable explain the protecting group based differences in reactivity, there appears to be a link between the protecting group on the saccharide and reactivity at the purine. A similar effect has previously been observed in direct displacement reactions at the C-6 position of purine nucleosides.^[21]

Finally, we wanted to assess if a relatively hindered HAPT such as (Et₂N)₃P and I₂ could be used to activate the C-6 amide carbonyl with a second nucleophilic amine performing the displacement (Scheme 6). Thus, 3.0 molar equiv PCl₃ were exposed to 9 molar equiv Et₂NH and 2.2 molar equiv I₂ in DCE. Addition of either 1a or 1b, 30 molar equiv imidazole and heating at 90 °C then led to formation of the imidazol-1-yl derivative from both 1a and 1b (67% and 86% product yields, respectively). These results are particularly relevant, as 1b did not undergo successful reaction when the imidazole/PCl₃/I₂ combination was used for amide carbonyl activation.

Scheme 6.

Synthesis of 6-imidazolyl derivatives from disilyl-protected 2′-deoxyinosine and triacetyl-protected inosine using PCl₃/Et₃NH/I₂ for amide activation.

Conclusions

Although we and others have developed the use of Pd-catalyzed amination methods for nucleoside modification,^[22] the approach using S_NAr displacement continues to play an important role. In this paper we have studied the use of in situ generated tris(dialkylamino)phosphines (HAPT) for the conversion of protected inosine and 2′-deoxyinosine derivatives into adenosine analogues. In several instances, excess 2° amine can be used for formation of the HAPT, as well as a base and nucleophile. However, it appears that formation of tetra(dialkylamino)phosphonium salts can become a competing process with the strongly nucleophilic^[23] and sterically less-demanding amines. Both silyl and acetyl protecting groups are suitable for these reactions although some differences in reactivity were observed. In a distantly related reaction, adenine has been produced in 16% yield upon heating hypoxanthine with (MeO)₃P=O at 100 °C followed by treatment with NH₄Cl/K₂CO₃ at 100 °C.^[24] The leaving group proposed in that reaction was dimethyl phosphate whereas in the present case a phosphoramide species is the likely leaving group.

In comparing the present method to our recently reported chemistry using O⁶-(benzotriazol-1-yl)inosine derivatives,^[10-12] this method offers a one-step conversion of hypoxanthine nucleosides to adenosine derivatives, and is particularly useful with cyclic 2° amines. However, the previously reported procedure^[10-12] is overall more versatile in terms of the diversity of nucleophiles that can be used for the C-6 functionalization.

Experimental Section

Thin layer chromatography was performed on 250 μm silica plates and column chromatographic purifications were performed on 200-300 mesh silica gel. PhMe was distilled over Na, DCE, (iPr)₂NEt, and CH₂Cl₂ were distilled over CaH₂, anhydrous DME was obtained from a commercial supplier. All other reagents were obtained from commercial sources and used without further purification. The conventional numbering system for purine nucleosides is used. ¹H NMR spectra were recorded at 500 MHz and were referenced to the residual protonated solvent. ³¹P{¹H} NMR spectra were recorded at 202 MHz and were referenced to 85% H₃PO₄ as external standard. Chemical shifts (δ) are reported in parts per million and coupling constants (J) are in hertz. Some representative synthetic procedures are given below. HRMS data is reported for new compounds.

9-[2-Deoxy-3,5-bis-O-(t-butyldimethylsilyl)- β-d-ribofuranosyl]-6-(morpholin-4-yl)purine (2b).^[10]

Using (iPr)₂NEt and stoichiometric morpholine

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (43.6 μL, 0.499 mmol) and dry toluene (10.7 mL) under nitrogen gas, and the mixture was cooled to ca 5 °C in an ice-water bath. After 20 min at this temperature morpholine (0.26 mL, 2.99 mmol) was added slowly dropwise while maintaining the temperature below 10 °C. The mixture was brought to room temperature and allowed to stir for 30 min. I₂ (126.7 mg, 0.499 mmol) was added to the reaction mixture and the stirring was continued for 10 min at room temperature. Disilyl 2′-deoxyinosine 1a (80.0 mg, 0.166 mmol) and DIPEA (0.26 mL, 1.49 mmol) were added and the mixture was stirred at room temperature for 21 h. The reaction mixture was diluted with EtOAc (50 mL) and washed with water (2 × 15 mL), followed by brine (15 mL). The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on silica gel using 20% acetone in hexanes afforded 64.5 mg (71% yield) of the morpholinyl nucleoside 2b as a thick, pale yellow oil. ¹H NMR (CDCl₃): δ 8.34 (s, 1 H, Ar-H), 8.02 (s, 1 H, Ar-H), 6.46 (t, J = 6.8, 1 H, H-1′), 4.59 (app dt, J_app = 6.2, 3.4, 1 H, H-3′), 4.29 (br s, 4 H, 2 x -OCH₂), 4.00 (app q, J_app = 3.9, 1 H, H-4′), 3.85–3.81 (m, 5 H, 2 x -NCH₂ and H-5′), 3.76 (dd, J = 11.2, 4.4, 1 H, H-5′), 2.59 (app dt, J_app = 13.6, 6.4, 1 H, H-2′), 2.41 (ddd, J = 13.6, 5.8, 4.0, 1 H, H-2′), 0.90 (s, 18 H, t-Bu), 0.09, 0.07 (2s, 12 H, SiCH₃).

Using excess morpholine

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (40.8 μL, 0.468 mmol) and dry toluene (4.0 mL) under nitrogen gas, and the mixture was cooled to 0 °C in an ice bath. After 5 min at this temperature morpholine (0.41 mL, 4.68 mmol) was added slowly dropwise and a white precipitate formed. The mixture was brought to room temperature and allowed to stir for 30 min. I₂ (87.1 mg, 0.343 mmol) was added followed by the addition of disilyl 2′-deoxyinosine 1a (75.0 mg, 0.156 mmol). The mixture was heated at 90 °C for 6 h. The orange-colored reaction mixture was diluted with EtOAc, and washed with water. The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on a silica gel column packed in CH₂Cl₂ using 20% acetone in hexanes afforded 59.2 mg (69% yield) of the morpholinyl nucleoside 2b.

9-[2-Deoxy-3,5-bis-O-(t-butyldimethylsilyl)- β-d-ribofuranosyl]-6-(imidazol-1-yl)purine (2e).^[10]

Using Et₂NH and imidazole

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (40.8 μL, 0.468 mmol) and dry DCE (4.0 mL) under nitrogen gas, and the mixture was cooled to 0 °C in an ice bath. After 10 min at this temperature Et₂NH (0.144 mL, 1.404 mmol) was added slowly dropwise while maintaining the temperature at 0 °C, and a white precipitate formed. The mixture was brought to room temperature and allowed to stir for 30 min during which time the precipitate appeared to dissolve. I₂ (87.1 mg, 0.343 mmol) was added to the reaction mixture and the stirring was continued for 10 min at room temperature. Disilyl 2′-deoxyinosine 1a (75.0 mg, 0.156 mmol) and imidazole (318.6 mg, 4.68 mmol) were added and the mixture was stirred at 90 °C for 2 h. The reaction mixture was diluted with CH₂Cl₂ (30 mL) and washed with water (2 × 15 mL), followed by brine (15 mL). The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on silica gel using 60% EtOAc in hexanes afforded 56.0 mg (67% yield) of the imizadolyl nucleoside 2e as a colorless oil. ¹H NMR (CDCl₃): δ 9.20 (s, 1 H, Ar-H), 8.80 (s, 1 H, Ar-H), 8.46 (s, 1 H, Ar-H), 8.40 (s, 1 H, Ar-H), 7.27 (s, 1 H, Ar-H), 6.55 (t, J = 6.0, 1 H, H-1′), 4.59 (app dt, J_app = 5.6, 4.2, 1 H, H-3′), 4.06 (app q, J = 4.2, 1 H, H-4′), 3.91(dd, J = 11.4, 3.9, 1 H, H-5′), 3.80 (dd, J = 11.4, 2.9, 1 H, H-5′), 2.65 (app dt, J_app = 12.7, 6.0, 1 H, H-2′), 2.51 (ddd, J = 12.7, 5.9, 2.3, 1 H, H-2′), 0.92, 0.91 (2s, 18 H, t-Bu), 0.11, 0.10 (2s, 12 H, SiCH₃).

9-(2,3,5-Tri-O-acetyl-β-d-ribofuranosyl)-6-(morpholin-4-yl)purine (3b).^[8b,20]

Using (iPr)₂NEt and stoichiometric morpholine

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (53.1 μL, 0.61 mmol) and dry toluene (10 mL) under nitrogen gas, and the mixture was cooled to ca 5 °C in an ice-water bath. After 15 min at this temperature morpholine (0.32 mL, 3.66 mmol) was added slowly dropwise while maintaining the temperature below 10 °C. The mixture was brought to room temperature and allowed to stir for 30 min. I₂ (154.4 mg, 0.609 mmol) was added to the reaction mixture and the stirring continued for 10 min at room temperature. Inosine triacetate 1b (80.0 mg, 0.203 mmol) and (iPr)₂NEt (0.32 mL, 1.84 mmol) were added and the mixture was stirred at room temperature for 24 h. The reaction mixture was diluted with EtOAc (50 mL) and washed with water (2 × 15 mL) followed by brine (15 mL). The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on silica gel using 30% acetone in hexanes afforded 60.9 mg (65% yield) of the morpholinyl nucleoside 3b as a thick, pale yellow oil. ¹H NMR (CDCl₃): δ 8.34 (s, 1 H, Ar-H), 7.89 (s, 1 H, Ar-H), 6.21 (d, J = 5.5, 1 H, H-1′), 5.89 (t, J = 5.4, 1 H, H-2′), 5.64 (dd, J = 5.3, 4.6, 1 H, H-3′), 4.43 (dd, J = 4.6, 2.0, 1 H, H-4′), 4.42 (dd, J = 12.8, 3.2, 1 H, H-5′), 4.36 (dd, J = 12.8, 5.2, 1 H, H-5′), 4.29 (br s, 4 H, 2 x -OCH₂), 3.82 (t, J = 4.9, 4 H, 2 x -NCH₂), 2.14, 2.13, 2.07 (3s, 9 H, OCOCH₃).

Using excess morpholine

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (42.7 μL, 0.489 mmol) and dry toluene (6 mL) under nitrogen gas, and the mixture was cooled to 0 °C in an ice bath. After 5 min at this temperature morpholine (0.33 mL, 3.77 mmol) was added slowly dropwise and a white precipitate formed. The mixture was brought to room temperature and allowed to stir for 30 min. I₂ (0.106 mg, 0.418 mmol) was added followed by the addition of inosine triacetate 1b (75.0 mg, 0.19 mmol). The mixture was heated at 90 °C for 2 h 10 min. The reaction mixture was diluted with EtOAc and washed with water. The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on a silica gel column packed in CH₂Cl₂, using 5% MeOH in CH₂Cl₂ gave a slightly impure product that was rechromatographed using 20% acetone in CH₂Cl₂ to yield 62.7 mg (71% yield) of the morpholinyl nucleoside 3b.

2′,3′,5′-Tri-O-acetyl-N⁶,N⁶-diethyladenosine (3e).^[21,25]

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (40.8 μL, 0.468 mmol) and dry toluene (4.0 mL) under nitrogen gas, and the mixture was cooled to 10 °C in an ice-water bath. After 10 min at this temperature Et₂NH (0.59 mL, 5.70 mmol) was added dropwise while maintaining the temperature 10 °C. A very viscous mixture was formed and the stirring was continued for 30 min at room temperature. I₂ (106.1 mg, 0.418 mmol) was added to the reaction mixture and the stirring was continued for 10 min. Inosine triacetate 1b (75.0 mg, 0.190 mmol) was added, and the mixture was stirred at room temperature for 4 h. The reaction mixture was diluted with CH₂Cl₂ (30 mL) and washed with water (2 × 15 mL), followed by brine (15 mL). The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on silica gel using 70% EtOAc in hexanes afforded 69.7 mg (81% yield) of 3e as a light-yellow, viscous liquid. ¹H NMR (CDCl₃): δ 8.32 (s, 1 H, Ar-H), 7.87 (s, 1 H, Ar-H), 6.21 (d, J = 5.7, 1 H, H-1′), 5.90 (t, J = 5.6, 1 H, H-2′), 5.65 (dd, J = 5.5, 4.2, 1 H, H-3′), 4.43 (dd, J = 4.2, 2.0, 1 H, H-4′), 4.41 (dd, J = 12.7, 3.3, 1 H, H-5′), 4.36 (dd, J = 12.7, 5.4, 1 H, H-5′), 4.17-3.72 (br s, 4 H, 2 x CH₂), 2.13 (s, 6 H, 2 x OCOCH₃), 2.06 (s, 3 H, OCOCH₃), 1.28 (t, J = 6.9, 6 H, 2 x CH₃).

9-(2,3,5-Tri-O-acetyl-β-d-ribofuranosyl)-6-(imidazol-1-yl)purine (3f).^[8b,21

Using Et₂NH and imidazole

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (42.7 μL, 0.571 mmol) and dry DCE (6.0 mL) under nitrogen gas, and the mixture was cooled to 0 °C in an ice bath. After 10 min at this temperature Et₂NH (0.18 mL, 1.74 mmol) was added slowly, dropwise while maintaining the temperature 0 °C, and a white precipitate formed. The mixture was brought to room temperature and allowed to stir for 30 min, during which time the precipitate appeared to dissolve. I₂ (106.2 mg, 0.418 mmol) was added to the reaction mixture, and the stirring was continued for 10 min at room temperature. Inosine triacetate 1b (75.0 mg, 0.190 mmol) and imidazole (388.4 mg, 5.71 mmol) were added, and the mixture was stirred at 90 °C for 5 h. The reaction mixture was diluted with CH₂Cl₂ (30 mL) and washed with water (2 × 15 mL), followed by brine (15 mL). The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on silica gel using EtOAc afforded 76.0 mg (86% yield) of 3f as a light-brown syrup. ¹H NMR (CDCl₃): δ 9.17 (s, 1 H, Ar-H), 8.79 (s, 1 H, Ar-H), 8.39 (s, 1 H, Ar-H), 8.27 (s, 1 H, Ar-H), 7.25 (s, 1 H, Ar-H), 6.27 (d, J = 5.3, 1 H, H-1′), 5.97 (t, J = 5.4, 1 H, H-2′), 5.67 (t, J = 5.3, 1 H, H-3′), 4.49 (app q, J_app = 4.1, 1 H, H-4′), 4.47 (dd, J = 12.7, 3.1, 1 H, H-5′), 4.40 (dd, J = 12.7, 4.3, 1 H, H-5′), 2.17, 2.14, 2.09 (3s, 9 H, OCOCH₃).

6-(Imidazol-1-yl)-9-[2,3,5-tri-O-(tert-butyldimethylsilyl)- β-d-ribofuranosyl]purine (4a)

Using excess imidazole

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (33.0 μL, 0.368 mmol) and dry DCE (4.0 mL) under nitrogen gas, and the mixture was cooled to 0 °C in an ice bath. After 10 min at this temperature imidazole (249 mg, 3.66 mmol) was added and a white precipitate formed. The mixture was brought to room temperature and allowed to stir for 30 min. I₂ (68.1 mg, 0.268 mmol) was added to the reaction mixture, and the stirring was continued for 10 min at room temperature. Trisilyl inosine 1c (75.0 mg, 0.122 mmol) was added, and the mixture was stirred at 90 °C for 36 h. The reaction mixture was diluted with CH₂Cl₂ (30 mL), and washed with water (2 × 15 mL), followed by brine (15 mL). The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on silica gel using 25%EtOAc in hexanes afforded 38.0 mg (47% yield) of 4a as a white solid. Characterization of this product has been previously reported.^[26]

2′,3′,5′-Tri-O-(t-butyldimethylsilyl-N⁶,N⁶-Diethyl-2′-deoxyadenosine (4b)

In a clean, dry, round-bottomed flask, equipped with a stirring bar were placed PCl₃ (33.0 μL, 0.368 mmol) and dry toluene (4.0 mL) under nitrogen gas, and the mixture was cooled to 10 °C in an ice-water bath. After 10 min at this temperature Et₂NH (0.38 mL, 3.67 mmol) was added dropwise, while maintaining the temperature 10 °C. A very viscous mixture was formed, and the stirring was continued for 30 min at room temperature. I₂ (68.1 mg, 0.268 mmol) was added to the reaction mixture, and the stirring was continued for 10 min. Trisilyl inosine 1c (75.0 mg, 0.123 mmol)) was added and the mixture was stirred at 90 °C for 8 h. The reaction mixture was diluted with CH₂Cl₂ (30 mL), and washed with water (2 × 15 mL), followed by brine (15 mL). The organic layer was separated, dried over Na₂SO₄, and evaporated to dryness. Chromatographic purification on silica gel using 20% EtOAc in hexanes afforded 67.0 mg (81% yield) of 4b as a light-yellow, thick oil. ¹H NMR (CDCl₃): δ 8.29 (s, 1 H, Ar-H), 7.94 (s, 1 H, Ar-H), 6.00 (d, J = 5.5, 1 H, H-1′), 4.79 (app t, J_app = 5.0, 1 H, H-2′), 4.33 (app t, J_app = 3.7, 1 H, H-3′), 4.10 (app q, J_app = 4.5, 3.5, 1 H, H-4′), 4.02 (dd, J = 11.5, 5.0, 1 H, H-5′), 4.02-3.92 (br s, 4 H, 2 x CH₂), 3.76 (dd, J = 11.5, 3.5, 1 H, H-5′), 1.27 (t, J = 7.0, 6 H, 2 x CH₃), 0.93, 0.78 (2s, 27 H, t-Bu), 0.10, −0.05, −0.02 (3s, 18 H, SiCH₃). HRMS calculated for C₃₂H₆₃N₅O₄Si₃ (M⁺) 665.4188, found 665.4201.

Scheme 1.

In situ formation of C-6 phosphonium derivatives and their conversion to adenosine analogues.