Jacek CIEŚLAK^a, Jadwiga JANKOWSKA^a, Michał SOBKOWSKI^a, Annika KERS^b, Inger KERS^b, Jacek STAWIŃSKI^b and Adam KRASZEWSKI^a*

^aInstitute of Bioorganic Chemistry, Polish Academy of Sciences, 61-704 Poznań, Poland;
 e-mail: akad@ibch.poznan.pl
^bDepartment of Organic Chemistry, Arrhenius Laboratory, Stockholm University,
  S-106 91 Stockholm, Sweden.

Aryl nucleoside H-phosphonates, derivatives of controlled reactivity, have been developed as useful synthetic intermediates for the preparation of variety of nucleotide analogues containing P-O, P-N and P-S bonds.

The concept of aryl nucleoside H-phosphonates as reactive intermediates in synthesis of nucleotides, oligonucleotides and their analogues arose from our earlier studies on transesterification reactions of various H-phosphonate diesters¹. A synthetic utility of aryl H-phosphonate derivatives have been demonstrated, inter alia, in efficient preparation of various dinucleoside H-phosphonates², nucleoside H-phosphonamidates³, during functionalization of support-bound oligonucleotides⁴, and in the synthesis of various auxiliary reagents for nucleotide analogues synthesis⁵.
Aryl nucleoside H-phosphonates can be considered as kind of activated nucleoside H-phosphonates which, in contrast to other activated H-phosphonate species, e. g. phosphonate-pivaloyl⁶, phosphonate-carbonate⁷ or phosphonate-phosphate8 mixed anhydrides, have only one electrophilic centre localised on the phosphorus atom. In addition, reactivity of these compounds can be modulated by varying inductive and mesomeric effects of the aryl substituent, what distinguish this class of compounds from mixed-anhydride type of intermediates, which reactivity is fixed and determined by the nature of the activator used. It should also be kept in mind that electronic properties of the aryloxy groups may affect the phosphonate <-> phosphite equilibria in these compounds, and in consequence, the phosphorus center in aryl H-phosphonates can change its character from electrophilic (in the H-phosphonate form) to a nucleophilic one (in the phosphite form), with all chemical implications of this fact.
In this report we wish to present selected examples of reactions of aryl nucleoside H-phosphonate which demonstrate how different aryl substituents influence rates and pathways of the investigated reaction and show how reactivity of aryl H-phosphonates can be modulated by changing electronic properties of the aryl group.

RESULTS AND DISCUSSION

Reactions of aryl nucleoside H-phosphonates with O-nucleophiles.
 Aryl nucleoside H-phosphonates 3a-g are easy accessible from nucleoside H-phosphonate 1 and respective phenol derivatives 2a-g in a coupling reaction⁹ aided by pivaloyl chloride (Pv-Cl) or appropriate chlorophosphates (e. g. NEPCl) (Scheme 1).
 
 

Recently, we assessed reactivity of aryl H-phosphonate diesters as a function of the aryl moiety present, by reacting 3a-g with N4,3’-O-dibenzoyldeoxycytidine 4a. Progress of the reaction was monitored by oxidising the produced dinucleoside H-phosphonate 5a to a stable dinucleoside phosphate 6a followed by TLC analysis. The most reactive among investigated aryl H-phosphonate derivatives were found those bearing p-nitrophenyl (3f) and 2,4,6-trichlorophenyl (3g) groups, which produced 5a in less than 3 min. The relative order of reactivity in this reaction, 3a : 3b : 3c : 3d : 3e : 3f : 3g was found to be 1 : 4 : 10 : 40 : 350 : 1100 : 1100, which parallels that observed for transesterification of 3 with simple alcohols¹⁰, and reflects an extent of modulation of reactivity available for compounds of type 3.

Reactions of aryl nucleoside H-phosphonates with N-nucleophiles.
 Recently, we have shown³ that direct coupling of nucleoside H-phosphonates with N-nucleophiles (e. g. primary and secondary amines or 5'-amino-5'-deoxynucleosides) is not completely chemoselective and produced N-acylated or N-phosphorylated amines together with  the desired nucleoside phosphonamidates. Considering the fact that aryl nucleoside H-phosphonates bear only one electrophilic centre, we investigated these compounds as possible substrates for the nucleoside phosphonamidates formation.

Aminolysis of aryl nucleoside H-phosphonates generated with aid of pivaloyl chloride.
 Aryl nucleoside H-phosphonates 3d or 3g, produced in situ form nucleoside H-phosphonates 1 and appropriate phenol derivative (2d or 2g, 1.2 - 2.0 molar equiv.) in the presence of Pv-Cl (1.5 molar equiv.) in methylene dichloride/pyridine 9 : 1 (v/v), was treated with amines 7a-f of rather narrow range of basicity but different steric hindrance at the nitrogen atom (Scheme 2).

All the investigated reactions were rapid (less than 3 min, ³¹P NMR) but differed in products distribution. For primary amine 7a, the main product of the reaction was the desired nucleoside phosphonamidate 8a (95 %) and only small amounts of side-product 9a (<5 %; phosphonate-phosphate derivative, giving rise to two groups of resonances centred at ~20 ppm and ~ -6 ppm in the  ³¹P NMR spectrum), was formed. With growing steric hindrance in an amine moiety used for the reaction, the amount of 9a gradually increased and for diisopropylamine, it constituted the sole product of the reaction¹¹.

The formation of phosphonate-phosphate 9a as a side-product in these reactions can be explained (Scheme 3) by generation of P-acylation species 11 from substrate 3d and pivaloyl chloride (or phosphono-pivalic mixed anhydride)12, followed by its reaction (base catalysis) with another molecule of 3d and a spontaneous rearrangement of the produced gem-diphosphonate 12 to phosphonate-phosphate 9a. This transformation, which seems to be analogous to that reported for simple dialkyl H-phosphonates13, competes with aminolysis of aryl H-phosphonate 3d to produce H-phosphonamidates 8, and explain why formation of phosphonate-phosphate side product 9 is most pronounced for sterically hindered amines 7c, 7e and 7f.
 We expected, that by increasing electrophilicity of the phosphorus centre  in 3  we should be able to speed-up the aminolysis and suppress (or eliminates completely) the competing P-acylation that triggered formation of phosphonate-phosphate side product 9. Indeed, when 2,4,6-trichlorophenyl nucleoside H-phosphonate 3g (prepared in situ as described for 3d) was allowed to react with amines 7a-e, a quantitative formation of the desired H-phosphonamidates 8a-e was observed (< 3 min, ³¹P NMR). The only exception was diisopropylamine, which also in this instance produced nearly quantitatively phosphonate-phosphate 9b.

Aminolysis of aryl nucleoside H-phosphonates generated with aid of 2-chloro-5,5-dimethyl-2-oxo-2lambda⁵-1,3,2-dioxaphosphinane (NEPCl)¹⁴.
 The use of chlorophosphates as condensing reagents for the in situ generation of aryl nucleoside H-phosphonates of type 3 should alleviate problems connected with P-acylation of 3 (vide supra) and thus increase efficiency of the aminolysis. We found, that using sterically hindered chlorophosphate NEPCl as a condensing agent it was possible to generate 3d and 3g from nucleoside H-phosphonate 1 and phenols 2d or 2g (ca 30 min) as the only nucleotidic species (³¹P NMR).
 As expected, treatment of the prepared in this way aryl nucleoside H-phosphonates 3d and 3g with amines 7a-e resulted in rapid (ca. 3 min) and clean formation of the corresponding alkyl nucleoside phosphonamidates 8a-e, which were purified by silica gel column chromatography and characterised by ¹H and ³¹P NMR, HRMS and elemental analysis. Thus, due to its  simplicity and effectiveness, this approach provides a convenient entry to various alkyl nucleoside H-phosphonamidates.³
 Also in this instance, the exception was diisopropylamine 7f, which being a strong base an ineffective nucleophile, promoted disproportionation of 3g to produce equimolar amounts of nucleoside bisaryl phosphite 10 and nucleoside  H-phosphonate 1.¹⁵ It is worth noting, however, that diisopropyl H-phosphonamidate 8f can be obtained in high yield is a direct coupling of H-phosphonate 1 and diisopropylamine 7f promoted by NEPCl.³

Reactions of aryl nucleoside H-phosphonates with hydrogen sulfide.
During our recent studies on sulfhydrolysis of aryl nucleoside H-phosphonates, we observed that treatment of nucleoside phenyl H-phosphonate 3c (generated in situ  from 1 and phenol using 1.1 equiv. of diphenyl chlorophosphate) in pyridine with hydrogen sulfide afforded, instead of the expected nucleoside H-phosphonothioate 13, equimolar mixture of  nucleoside H-phosphonodithioate 16 and nucleoside H-phosphonate 1. H-Phosphonates 3d, 3e and 3f, bearing more electron-withdrawing aryl moieties, reacted faster but with the same product distribution as that during sulfhydrolysis of 3c. The reaction was very sensitive to the amount of a condensing agent used for generation of aryl nucleoside H-phosphonates 3, and with 3 molar equiv., instead of usually used 1.1, exclusively H-phosphonodithioate 16 was formed during the subsequent sulfhydrolysis.

To explain these findings we proposed a plausible mechanism for sulfhydrolysis of aryl H-phosphonate 3 in pyridine (Scheme 4). It involved an initial formation of nucleoside H-phosphonothioate 13 which reacted with aryl nucleoside H-phosphonate 3 to produce mixed H-pyrophosphonate 14, and this in turn, in sulfhydrolysis with H₂S afforded H-phosphonodithioate 16 and nucleoside H-phosphonate 1. The last step may involve, under the reaction conditions, intermediacy of aryl nucleoside H-phosphonothioate 15 (formed from mixed H-pyrophosphonate 14 and the appropriate phenol), followed by its reaction with hydrogen sulfide to produce 16. This possibility was substantiated by findings that aryl nucleoside H-phosphonothioates of type 15 readily underwent transformation to H-phosphonodithioates 16, when treated in pyridine with H₂S (³¹P NMR). The exclusive formation of H-phosphonodithioate 16 when excess of the condensing agent was present during sulfhydrolysis, can be explained by generation of various reactive species from the produced H-phosphonate 1, which then collapsed to H-phosphonothioate 13 and ultimately, to the final product 16.

Financial support from the State Committee for Scientific Research, Republic of Poland [3 T09A 118 14] and the Swedish Natural Science Research Council is gratefully acknowledged.

REFERENCES AND NOTES
1. Sobkowski M., Stawiński J., Sobkowska A. and Kraszewski A.: J. Chem. Soc. Perkin Trans. 1, 1994, 1803.
2. Cieślak J., Sobkowski M., Kraszewski A., Stawiński J.: Tetrahedron Lett. 1996, 37, 4561.
3. Sobkowska A., Sobkowski M., Cieślak J., Kraszewski A., Kers I., Stawiński J.: J. Org. Chem. 1997, 62, 4791.
4. Sobkowski M., Kraszewski A., Stawiński J.: Nucleosides Nucleotides 1998, 17, 253
5. a) Jankowska J., Sobkowski M., Stawiński J., Kraszewski A.: Tetrahedron Lett. 1994, 35, 3355.; b) Kers A., Kers I., Stawiński J., Sobkowski M., Kraszewski A.: Synthesis 1995, 427.
6. a) Garegg P. J., Lindh L., Regberg T., Stawiński J., Strömberg R., Henrichson L.: Tetrahedron Lett. 1986, 27, 4051.; b) Froehler B. C., Ng P. G., Matteucci M. D.: Nucleic Acids Res. 1986, 14, 5399.
7. Efimov V. A., Kalinkina A. L., Chakhmakhcheva O. G.: Nucleic Acids Res. 1993, 21, 5337.
8. Strömberg R., Stawiński J. Nucleic Acids Res. 1987, 18, 185.
9. Cieślak J., Jankowska J., Szymczak M., Kers A., Kers I., Stawiński J. and Kraszewski A. Nucleosides Nucleotides 1999, 0000.
10. Cieślak J., Jankowska J., Kers A., Kers I., Sobkowska A., Sobkowski M., Stawiński J. and Kraszewski A. Collect. Czech. Chem. Commun. Special Issue 1996, 61, S242.
11. Phosphonate-phosphate derivative 9a was isolated and its structure determined by 1H, 31P NMR spectroscopy and elemental analysis.
12. Kers A., Kers I., Kraszewski A., Stawiński J. Coll. Czech. Chem. Commun. (Special Issue) 1996, 61, S246.
13. Fitch S. J. and Moedritzer K. J. A. Chem. Soc. 1962, 84, 1876.
14. Strömberg R., Stawiński J. Nucleic Acids Res. Symp. Ser. 1987, 18, 185.
15.  Kers A., Kers I., Stawiński J., Sobkowski M., Kraszewski A. Tetrahedron 1996, 52, 9931.