Synthetic Approaches for the Preparation of Phosphoramidate Prodrugs of 2′‐Deoxypseudoisocytidine

Abstract A synthetic procedure for the preparation of phosphoramidate prodrugs of C‐nucleosides is reported. Different phosphorochloridates were reacted with 3′‐O‐protected N‐acetyl‐2′‐deoxypseudoisocytidine or 3′‐O‐protected 2′‐deoxypseudoisocytidine, followed by acidic hydrolysis of the protecting group. In the presence of the N‐acetyl moiety, the enolisable keto group of the nucleobase was able to react (like the 5′‐OH) with the phosphorochloridates to give bisphosphorylated derivatives. Epimerisation (β to α) occurred if the amino group of the nucleobase was unprotected. These side reactions demonstrate the peculiar behaviour of C‐nucleosides compared to their nucleoside analogues. It was demonstrated that the first enzymatic activation step for this new class of prodrugs can be mediated by carboxypeptidase and that it follows the same pathway and rate reported for ProTides of more conventional nucleoside analogues. These new phosphoramidate derivatives deserve further investigation for their therapeutic potential as anti‐cancer agents.


Introduction
The C-nucleosidesr epresent ag roup of nucleoside analogues in which the sugar moiety is linked to the nucleobase by ac arbon-carbon bond. [1] Several C-nucleosidesa re naturally occurring compounds. Among them, pseudouridinew as the first to be isolated from yeast tRNA in 1957. [2] Subsequently, other C-nucleosides,i ncludingo xazinomycin, [3] pyrazomycin, [3] showdomycin, [4] and formycin A, [5] were isolated from culture filtrates of different bacterial strains.T hese compounds are antibioticsa nd exhibit anti-cancer and/or antiviral activity.T heir advantageous properties arise from the presence of aC ÀCg lycosidic bond, which gives ag reater resistance than N-nucleosides towards chemical hydrolysis and enzymatic hydrolysis by phosphorylase andd eaminasee nzymes. On the basis of these interesting chemical and biological properties, aw ide variety of synthetic analogues have been prepared thanks to the large array of novel synthetic methodologies developed in the last two decades. Several of these compounds have found numerous applications in medicinal chemistry and chemical biology. [1] Amongt hem, pseudoisocytidine (PIC, 1), an ucleoside isostere of cytidinew as developed as ac andidate fora nti-leukaemic therapy [6] (Figure 1). PIC was shown to be incorporated into both RNA and DNA and this incorporation was considered to be responsible for its therapeutic activity,w hich hasb een observed against several mousel eukaemias in vitro and in vivo. [7,8] In addition, PIC was found to disruptD NA methylation by inhibition of the enzymeD NA methyltransferase, most probablydue to the presence of anitrogen atom in the 5-position of the base. [9] However,t he development of PIC was halted due to hepatotoxicity observed during phase Ic linical evaluation. [10] The efficiency with which PIC is incorporated into RNA, andt he rapid RNA turnover,a ssociated with protein synthesis in the liver,w ere considered the main causes of its hepatotoxicity.T hisf inding prompted the investigation of 2'deoxypseudoisocytidine 2), [11] which, in preliminary tissue culture experiments,w as found to exhibit inhibitory activity against P815 cell lines. [11a] PIC, 2'd-PIC andtheir analogues were also used as novel base-pairinga gents in oligonucleotides to investigate DNA and RNA structures andf unctions. [12] Althoughs everal C-nucleoside analogues have been described as anti-cancer and/or antiviral agents, none have ever been developed as anti-cancer or antiviral drugs.T he recent advent of two novel C-nucleosides, BCX4430 (3) [13] and GS-6620 (4), [14] as potentialt herapeutic agentsf or the treatment of the Ebola virus and hepatitis Cv irus (HCV) infections, respectively,h as stimulated renewed interest in this class of compounds ( Figure 1).
As part of our currentr esearch we wereinterested to further investigate the potential utility of 2'd-PIC (2)a sa na nti-leukaemic agent by preparing as eries of phosphoramidate prodrugs for biological evaluation as anti-cancera gents. "ProTides" in As yntheticp rocedure for the preparation of phosphoramidate prodrugs of C-nucleosidesi sr eported.D ifferent phosphorochloridates were reactedw ith 3'-O-protected N-acetyl-2'-deoxypseudoisocytidine or 3'-O-protected 2'-deoxypseudoisocytidine, followed by acidic hydrolysis of the protecting group. In the presenceo fthe N-acetyl moiety,t he enolisable keto group of the nucleobase was ablet or eact (like the 5'-OH) with the phosphorochloridates to give bisphosphorylated derivatives. Epimerisation (b to a)o ccurred if the amino group of the nu-cleobasew as unprotected. These side reactions demonstrate the peculiar behaviour of C-nucleosidesc ompared to their nucleoside analogues.I tw as demonstrated that the first enzymatic activation step for this new class of prodrugs can be mediated by carboxypeptidase andt hat it follows the same pathway and rate reportedf or ProTides of more conventional nucleoside analogues. These new phosphoramidate derivatives deservef urtheri nvestigation for their therapeutic potential as anti-cancer agents.
The coupling reaction of 5-iodo base 11 with the protected ribofuranosyl glycal 12 using Pd(OAc) 2 as ac atalyst, AsPh 3 as as oft ligand and N,N-diisopropylethylamine as ab ase, formed selectively the b-C-nucleoside 13.A fter removal of the silyl groups with fluoride ions, the resulting2 '-deoxy-3'-keto C-nucleoside 14 was treated with sodium triacetoxyborohydride to reduce diastereoselectively the 3'-keto group from the b-face of the furanosyl ring, forming N-acetyl-2'-deoxypseudoisocytidine 15. [22] The cleavage of the acetyl group to afford nucleoside 2 was then accomplishedb yb asic hydrolysisu sing NH 3 in MeOH.T he assignment of the configuration at the 1'-position of 2 was based on the comparison of its 1 HNMR spectrum with that reported in the literature. [22] 2.2. Synthesis of N-Acetyl-2'-deoxypseudoisocytidine Phosphoramidates The two synthetic strategies commonly used for the preparation of phosphoramidate prodrugs (phosphorochloridate in the presence of either tert-butylmagnesiumc hloride or Nmethylimidazole as ab ase) [26] failed when appliedt o2,p robably due to the low solubility of the starting materiali nt he reaction medium, returning only unreacted startingm aterials. Attempts to improve the solubility of 2 using different solvents were unsuccessful.A pplication of the ProTide approacht op recursors 14 and 15 also failed, indicating that development of as uitable synthetic strategy to afford phosphoramidates of 2 was more challenging than originally expected. These results prompted us to use ad ifferent synthetic methodology with compound 17 as the key intermediate (Scheme 2).
We envisaged that introduction of a tert-butyldimethylsilyl ether at the 3'-OH group in 15 would help to improve its solubility and to achieve exclusive phosphorylation at the 5'-position.
In order to prepare compound 17,t he two hydroxy groups of deoxyribose present in N-acetyl-2'-deoxypseudoisocytidine (15)w ere first protected with a tert-butyldimethylsilyl group using tert-butyldimethylsilyl chloride in DMF for 24 ha tr oom temperature in the presence of 4-dimethylaminopyridine (DMAP)t op rovide, after flash chromatography,c ompound 16 in reasonabley ield. Then, selectives ilyl group deprotection was achievedw itha queous trifluoroacetic acid to give, after isolation by silica gel chromatography, 17 with af ree primary hydroxy group in moderate yield. Next, phosphorochloridates 18 a-f,p repared as am ixture of R P and S P diastereoisomers accordingt oaliterature procedure, [26] were reactedw ith 17 in the presence of tert-butylmagnesium chloride (1.0 m in THF), yielding3 '-O-tert-butyldimethylsilyl phosphoramidates 20 a-f (Scheme2)asdiastereoisomeric mixtures after column chromatography,e xcept for 20 d,w hich wasi solated after purification as as ingle diastereoisomer.D espite the almost complete consumption of the starting material, the desired products 20 a-f were recovered in low yields, which was ascribed in each case to the formation of ab isphosphorylated by-product, as exemplified in Figure 2. The bisphosphorylated compound 19 f was isolateda nd its structure was characterisedb ym asss pectrometry and 31 Pa nd 1 HNMR analysis, [31] which clearly suggested that the phosphorylationinvolved the oxygen atom of the pyrimidine ring rather than either one of the nitrogen atoms. N-Acetylisocytidine possesses an enolisablek eto group which, like the 5'-OH group, is able to react with ap hosphorochloridate to give an O-phosphorylated derivative. In support of this result,w ef ound in the literature that the reaction of 2-acetylamino-4-hydroxypyrimidines with phosphorochloridates gives O-phosphoryl rather than N-phosphoryl derivatives. [27] The substantials teric requirement of the phosphoryl chloride and the steric hindrance exerted to some extent by the acetylg roup were considered to be the key features for preventing phosphorylation at either one of the ring nitrogen atoms. [27] Acidic deprotection of 20 a-f afforded after preparative HPLC purification compounds 21 a-f in moderate yields (Scheme2and Table 1). Attempts to removet he acetyl protec-tion from 21 a with Schwartz's reagent as described by Ferrari et al., [28] failed due to the ring openingo ft he base. The difficultiese ncountered in removing the N-acetyl group from 21 af using mild conditions, and the fact that the labile PÀOb ond of the ProTide would not tolerate other harsh de-acetylating agentss uch as methanolic ammonia,p romptedu st oa bandon our attempts toward modification of 21 a-f.W et herefore continued our effort to conceiveamore efficient route that would allow the preparation of the N-deacetylated analogues.

Synthesis of 2'-Deoxypseudoisocytidine Phosphoramidates
As shown in Scheme3,c ompound 22,o btained by treatment of 16 with methanolic ammonia, underwent selective5 '-desilylation using aqueous trifluoroacetic acid in THF to afford the monosilyl compound 23 in excellent yield. Next, phosphorochloridates 18 a and 18 g were reactedw ith 23 in the presence of tert-butylmagnesium chloride (1.0 m in THF) to yield, after columnc hromatography,t he 3'-O-tert-butyldimethylsilyl-protected phosphoramidates 24 a and 24 g in moderate yield as diastereoisomeric mixtures (Table 2). No traces of bisphosphorylated products either due to O-o rN-phosphorylation were observed.
Acidic deprotectiono ft he tert-butyldimethylsilylm oieties in 24 a and 24 g with trifluoroacetic acid in dichloromethane (1:2 v/v;r oom temperature, overnight), afforded the final compounds 25 a and 25 g as mixtures of a and b isomersi na3:1 ratio after columnc hromatography.T he b-isomers of 25 a and   25 g were isolated in low yield after preparative HPLC purification (25 a as as ingle diastereoisomer and 25 g as am ixture; Scheme 3a nd Table 2). Most probably, the presence of ad issociable proton on N-1 facilitates the a,b-epimerisation in acidic conditions through aring opening-closure of the carbohydrate ring (Scheme 4) as previously reported for other C-nucleosides. [11b, c, 29] If am ild procedure for the cleavageo ftert-butyldimethylsilyl ethers to alcohols (based on an exchange reactionw ith trimethylsilyl triflate at À78 8C) [30] was used, no epimerisation was observed.

Enzymatic Studieso nt he Activation of C-Nucleoside ProTides
To exert their biological activity,P roTides must be metabolised in vivo into the monophosphate form,w hich in turn generates the active triphosphate form by two consecutive phosphorylation reactions. [31] In the process of intracellular activation of ProTides, the first step is catalysed by ac arboxyesterase-type enzyme,s uch as cathepsin A, which was shown to be responsible for the cleavage of the amino acid ester moiety. [32] In order to demonstrate that the ProTides of C-nucleosidesa re activated in as imilar manner,t he interaction of compound 21 e with ac arboxyesterase-typee nzyme was investigated. Carboxypeptidase Yw as used as as urrogate of cathepsin Ab ecause it belongs to the same family of C-type carboxypeptidases and it was reported to share similarities in the active site. [33] Compound 21 e in [D 6 ]acetonew as therefore incubated in an NMR tube with carboxypeptidase Yi nT rizma buffer (pH 7.6), and the progress of the reactionw as monitored by 31 PNMR analysiso ver1 4h.T he stacked spectra ( Figure 3) show the formation of an ew peak after 10 min of incubation, which corresponds to intermediate I (d P = 5.06 ppm, t = 10 min). Complete conversion of the ProTide 21 e (which in [D 6 ]acetonea ppears as as ingle peak at d P = 4.26 ppm) into the correspondinga minoacyl phosphoramidate ester (II: d P = 7.19 ppm) was observed in 40 min. In vivo, the aminoacyl phosphoramidate ester metabolite is then believed to undergo PÀNb ond cleavage, mediated by ap hosphoramidase-type enzymet oe ventually release the parent drug in its monophosphate form.

Conclusions
An alternative route to C-nucleoside ProTides has been developed and used to preparep hosphoramidate derivatives of 2'd-PIC (2)a nd N-acetyl 2'd-PIC (15). Unexpecteds ide reactions such as phosphorylationo ft he enolisable keto group of the nucleobase and epimerisation throughr ing openingh ighlighted the differentr eactivity of C-nucleosides compared to nucleoside analogues.The first carboxypeptidase-mediated bioactivations tep for this new class of prodrugs followed the same pathway and rate as reportedf or ProTides of conventional nucleoside analogues. Biological evaluation of these novel nucleoside analogues shoulde nhanceo ur understanding of the potentialo fC-nucleosides as anti-tumour agents and in particular of 2'd-PIC as an anti-leukaemic drug.T ogether with derivatives 25 a and 25 g,w ep lan to evaluate the N-acetylated derivatives 21 a-f for their anti-tumour activity.W ec onsidered that the acetyl moiety would further enhance the lipophilicity of these compounds and remove the potentialf or their protonation in vitro, whereas in vivo the acetyl moietyw ould most probably be able to undergo cleavage (thus acting as ad ual prodrug). The results of these investigationsw ill be disclosed in due course.

Experimental Section
Chemistry All anhydrous solvents were purchased from Sigma-Aldrich and amino acid esters from Novabiochem. All commercially available reagents were used without further purification.
For practical purposes, in some cases standard procedures are given. Procedures that differ from the standard are fully described.

N-(5-{(2'R,4'S,5'R)-4'-[(tert-Butyldimethylsilyl)oxy]-5'-(hydroxymethyl)tetrahydrofuran-2'-yl}-6-oxo-1,6-dihydropyrimidin-2yl)acetamide (17)
Am ixture of TFAa nd H 2 O( 1:1v /v,2 .4 mL) was added dropwise to as olution of 16 (0.3 g, 0.6 mmol) in THF (4.8 mL) at 0 8C. The reaction mixture was stirred at room temperature for 2h under an argon atmosphere, then quenched with aqueous NaHCO 3 ;C H 2 Cl 2 was added and the aqueous phase was extracted twice with CH 2 Cl 2 .T he combined organic layers were dried over MgSO 4 ,f iltered and concentrated under vacuum to yield 17 as ag lassy solid (0.172 g, 75 %), which was used in the next step without further purification. 1  Standard Procedure 1: Synthesis of Phosphorochloridates 18 a-g Anhydrous triethylamine (2.0 mol equiv.) was added dropwise at À78 8Ct oas tirred solution of the appropriate amino ester hydrochloride/tosylate salt (1.0 mol equiv.) and the appropriate dichlorophosphate (1.0 mol equiv.) in anhydrous dichloromethane (61.6 mol) under an argon atmosphere. After 1hthe reaction mixture was allowed to warm to room temperature and was stirred for an additional 1-2 h. Formation of the desired phosphorochloridate was monitored by 31 PNMR spectroscopy.A fter the reaction was completed, the solvent was removed under reduced pressure and the resulting residue was re-dissolved in anhydrous diethyl ether and the triethylammonium salt was removed by filtration. The filtrate was evaporated to dryness and the crude material was purified by flash column chromatography with ethyl acetate/ hexane (1:1 v/v) as the eluent to give the desired phosphorochloridate as an oil.