New synthetic methodologies have been developed to prepare novel nucleoside and nucleotide analogues for enlarging the collection of potential antitumoral and antiviral drugs and for improving the pharmacological and pharmacokinetic properties of the existing ones [1,2,3,4,5,6,7,8,9,10]. Particular attention has been paid to imidazole nucleosides and nucleotides considering that their precursors, 5-aminoimidazole-4-carboxamide riboside (acadesine or AICAR) and 5-aminoimidazole-4-carboxamide ribotide (ZMP, Figure 1), are involved in the activation of AMP-activated protein kinase (AMPK) [11,12]. In the cells AICAR is phosphorylated to ZMP that is a mimic of AMP. The direct binding of ZMP to an allosteric site of AMPK causes a cascade of important metabolic events, including inhibition of the basal and insulin-stimulated glucose uptake, lipogenesis and glucose oxidation [13]. In the treatment of ischemia, AICAR has also been shown to be cardioprotective [14], preventing the endoplasmic reticulum stress, the inhibition of the sodium/hydrogen exchanger and thus the rapid rise of intracellular sodium, which normally occurs during the reperfusion [15].

The AMPK pathway is also implicated in the regulation of cell proliferation and activation by AICAR could result in pro-apoptotic effects [16,17]. Moreover, it was established that mutations affecting the Ca²⁺ releasing channel RYR1 are associated with a broad spectrum of human disorders, including malignant hyperthermia, central core disease and core-rod myopathy [18]. By using a mouse model of malignant hyperthermia having mutation in the RYR1 gene, Lanner et al. have recently demonstrated that AICAR can inhibit Ca²⁺ leakage through RYR1 by a mechanism independent from AMPK activation, thus preventing heat-induced sudden death in the mutated mouse [18]. Considering that AICAR has a short half-life in cells, does not cross efficiently the blood-brain barrier, is poorly adsorbed by the gastrointestinal tract, increases uric acid production and the lactic acidosis [15], the development of novel AICAR derivatives/analogues endowed with an improved pharmacological profile is an appealing goal in the field of medicinal chemistry.

In recent years we have focused our attention on the preparation of new base- and sugar-modified nucleosides and nucleotides [4,10,19,20,21,22,23,24] with the aim to enlarge the pool of new potential antimetabolites and to obtain new probes to better understand how certain metabolic pathways work. It is well known that the introduction of fluorine in a potential drug can dramatically change its chemical behaviour, influencing its pharmacokinetic and pharmacodynamics properties; furthermore, increased metabolic stability and a more efficient permeation of biological membranes are a consequence, respectively, of the strength of the C-F bond and of the higher hydrophobicity of fluorine respect to hydrogen [25,26,27].

In the last years several novel synthetic methodologies have been proposed for the selective introduction of fluorine at the 2′-, 3′- or 5′-position of a nucleoside [28,29]; in some cases these substitutions were revealed to be so important that novel anti-tumor and anti-viral drugs have been obtained [28]. In this paper we report on the synthesis of the novel analogue 5′-fluoro-5′-deoxy-acadesine (5′-F-AICAR, 8, Scheme 1) in which the 5′-hydroxyl group is replaced by a fluorine atom. The latter is an isopolar mimic of the hydroxyl group, but prevents the kinase-mediated 5′-phosphorylation by cellular kinases. We hypothesized that this new derivative, which can be considered the most similar and non-phosphorylable AICAR analogue, could be a useful tool to explore the wide and not completely understood physiological roles of AICAR and ZMP. Futhermore, 5′-F-AICAR may be also considered as a mimic of 5′-fluoro-5′-deoxyadenosine (5′-FDA) that has been employed as a tool for studying the Actinomycete bacterium Strepromyces cattleya [30]. Interestingly, some 5′-FDA derivatives were revealed to be inhibitors of the enzyme S-adenosyl-l-homocysteine hydrolase (an enzyme involved in the replication of some viruses) thanks to their chemical inertness toward the phosphorylation [30,31]. Hence, it would not be surprising if the novel compound 5′-F-AICAR could also work as an antiviral agent.

At first, we tried the selective insertion of the fluorine atom at the 5′-position of AICAR starting from its 2′,3′-O-isopropilydene derivative, using common fluorinating agents, such as: (a) DAST, (b) deoxofluor, (c) the Olah reagent [28] and (d) the tosyl fluoride (TsF)/tetrabutylammonium fluoride system (TBAF) [30]. Unfortunately, all these procedures failed, resulting in the recovery of only the unreacted starting material (reagent d) or decomposition of the nucleoside (reagents a–c). These results thus prompted us to explore an alternative synthetic route. We hypothesized that the construction of the 5-aminoimidazole-4-carboxamide moiety could be performed starting from an easily fluorinable purinic substrate, from which AICAR could be obtained through the degradation of the purine ring. In this context, we have recently reported efficient synthetic strategies to obtain AICAR, ZMP and their 4-N-alkyl derivatives, through the purine ring degradation of suitable inosine or 5′-phosphate inosine precursors, respectively, after the introduction of a suitable strong electron-withdrawing group (2,4-dinitrophenyl, DNP) at the N1 position of the base moiety [20,21,22,24].

Accordingly, for the preparation of 5′-F-AICAR (8) we used as starting material commercially available 6-chloronebularine (1, Scheme 1), which was readily transformed into its 2′,3′-O-isopropylidene derivative 2 [32]. Then, the 5′-OH group of 2 was replaced by a 5′-F in a one-pot tosylation/fluorination sequence, treating 2 with the TsF/TBAF reagent system, according to the Schimizu’s procedure [30]. It should be noted that the presence of an electron-withdrawing atom, such as chlorine, at the C6 position is necessary to prevent the intramolecular nucleophilic attack of N3 on the first-formed C5′-OTs, that would led to the unwanted cyclic side product containing the N3-C5′ bond [29]. As previously observed by Ashton et al. [30], we noted (TLC monitoring) that the process led not only to the formation of the expected fluorinated product 3a, but also to the formation of the bis-fluorinated product 3b (Scheme 1, X = F), as a consequence of the substitution in 2 of the C6 chlorine atom by fluoride. This was in no way detrimental to the continuation of the synthetic plan since both 3a and 3b could be used in the next step. Thus, the 3a/3b mixture was converted into the corresponding inosine derivative 4 by reaction with a 0.1 M NaOH solution, affording 4 in a 60% yield. In particular, 3a and 3b were revealed to be reactive to the same extent toward the hydroxyl ion, and the displacement of chlorine and fluorine atoms at C6 could both be monitored by TLC.

Importantly, the success of this reaction was strongly dependent upon the concentration of the base employed. In fact, concentrations of NaOH higher than 0.1 M caused the decomposition of the nucleosidic material leading to the formation of several side-products, one of which was identified as 2′,3′-O-isopropylideneinosine.

Then, removal of the isopropylidene group in 4 with 70% trifluoroacetic acid (TFA) gave inosine derivative 5 [33], which was acetylated with Ac₂O in pyridine under standard conditions to give the 2′,3′-di-O-acetylated product 6 quantitatively. The change of the 2′,3′-protecting group was necessary since the deprotection of 2′,3′-O-isopropylidene group in the final step led to a partial decomposition of the 5′-fluoro-acadesine 8. Next, introduction in 6 of the strong DNP electron-withdrawing group at the N1 position was carried out by treatment with K₂CO₃ and 2,4-dinitrochlorobenzene (DNClB) to give 7 [34,35]. This process served to activate the C2 position of the purine, rendering it susceptible to nucleophilic attack by amines, thus leading to the cleavage of the N3-C2 bond. In particular, we have previously demonstrated that when a 1,ω-diaminoalkane is employed, the fate of the open intermediate depends on the length of the alkyl chain separating the two amino groups. If the diamine comprises two or three methylene groups, AICAR is formed as the sole product [22].

Eventually, when compound 7 was left into contact with a solution of ethylendiamine (EDA) in DMF, 5′-F-AICAR 8 was obtained by purine-ring opening and concomitant deacetylation of the hydroxyl groups at the 2′-3′ of the ribose moiety. During this reaction no loss of the fluorine at the 5′C position was observed. Compound 8 was obtained in 70% yield starting from 7 and an overall 25% yield from 1. Its structure was confirmed by 1D/2D-NMR evidence and by high-resolution mass spectrometry (HRMS).

In this paper we have reported a facile synthetic approach for the preparation of the novel compound 5′-fluoro-5′-deoxyacadesine (5′-F-AICAR) starting from commercially available 6-chloro-nebularine (1). This product may be considered a novel non-phosphorylable AICAR derivative, whose biological activity has to be explored, and that could be a useful tool to understand the metabolic pathways involving AICAR and ZMP. The described synthetic procedure may be extended to the preparation of novel fluorinated base modified AICAR.