Synthesis of a phosphinic acid analogue of cyclic AMP

Article abstract of DOI:10.1016/S0040-4039(00)01437-4

Synthesis of a cyclic phosphinate ester analogue of cyclic AMP was achieved by using a double Arbuzov-type cyclisation of bis-trimethylsilyl phosphonite onto a sugar dihalide derivative as the key step. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)01437-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 8211–8215
Synthesis of a phosphinic acid analogue of cyclic AMP
Andrew C. Regan,* Nunzio Sciammetta and Peter I. Tattersall
Department of Chemistry, University of Manchester, Oxford Road, Manchester M13 9PL, UK
Received 25 August 2000; accepted 29 August 2000
Abstract
Synthesis of a cyclic phosphinate ester analogue of cyclic AMP was achieved by using a double
Arbuzov-type cyclisation of bis-trimethylsilyl phosphonite onto a sugar dihalide derivative as the key step.
© 2000 Elsevier Science Ltd. All rights reserved.
Cyclic AMP 1 (Fig. 1) is a well-known ‘second messenger’ in cell signalling.1,2 It is formed in
eukaryotic cells by membrane-bound adenyl cyclases, activated by a variety of hormones, and
can have a wide range of effects in the cell, usually mediated by cyclic AMP-dependent protein
kinases, or by cyclic nucleotide-gated ion channels. Cyclic AMP is broken down to AMP by
cyclic nucleotide phosphodiesterases (PDEs), and this is an important mechanism for regulating
its intracellular concentration. Replacement of both the oxygen atoms in the phosphate diester
ring of cyclic AMP by methylene groups would give the corresponding phosphinic acid analogue
2, which should be resistant to hydrolysis by PDEs. To our knowledge, this phosphinic acid
analogue of cyclic AMP is novel, although the two phosphonic acid analogues 3 and 4, where
only one of the oxygen atoms is replaced by CH2, have previously been reported.3
Figure 1.
* Corresponding author. Fax: +44 161 275 4617; e-mail: andrew.regan@man.ac.uk
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01437-4

8212
A variety of different PDEs are known, from different sources and with varying substrate
specificities. Many of these PDEs cleave cyclic AMP specifically to give 5%-AMP, and consistent
with this, a PDE from amoebae of Dictyostelium discoideum has been reported to hydrolyse
phosphonate 3 as fast as cyclic AMP 1, but phosphonate 4 was not cleaved.4 A PDE from rabbit
brain was found to cleave 3 slowly, but again did not cleave 4.5 However, a PDE from a plant
source (Jerusalem artichoke) hydrolyses cyclic AMP to give a mixture of both 3%-AMP and
5%-AMP,6 and so both phosphonates 3 and 4 might be susceptible to hydrolysis by PDEs of this
type. The advantage of the phosphinate 2 is that both PꢀO bonds are replaced by PꢀCH2, and
so 2 should be resistant to cleavage in both the 3%- and the 5%-directions, as well as any
non-specific hydrolysis.
In this Letter, we report a synthesis of a methyl phosphinate ester analogue 5, and its
subsequent hydrolysis to the phosphinic acid analogue of cyclic AMP 2. The retrosynthetic
analysis is shown in Scheme 1, and involves a double Arbuzov-type reaction to generate a cyclic
phosphinic acid from a dihalo-sugar derivative e.g. 7 or 8, or ditriflate 9.
Scheme 1.
The starting material for the synthesis is diacetone glucose 10, which was converted into
the C3-homologated compound 11 using literature procedures,7–9 as shown in Scheme 2. The
C3a-hydroxyl group in 11 was protected as the benzyl ether using phase-transfer conditions,
and selective hydrolysis of the side-chain acetonide in 12 using warm aqueous acetic acid
afforded the diol 13.
Selective deoxygenation of the 5-hydroxyl group in 13 was investigated using several proce-
dures, and the most efficient involved formation of the dimesylate 14, elimination of both
mesylate groups using sodium iodide10,11 to give the alkene 15, followed by hydroboration–
oxidation to reintroduce the 6-hydroxyl group in 16. Removal of the C3a-benzyl group then
gave the diol 17, which was converted into the dibromide 7 using triphenyl phosphine and
NBS, with the addition of pyridine.9 The corresponding diiodide 8 and ditriflate 9 were also
prepared from the diol 17, and the stereochemistry of diiodide 9 was confirmed by an X-ray
structure determination.

8213
Scheme 2.
The key step in the synthesis is the formation of the cyclic phosphinic acid 18, by a double
Arbuzov-type reaction12,13 between bistrimethylsilyl phosphonite (BTSP) and the dibromide 7
(Scheme 3). Since pure BTSP is highly pyrophoric, it was formed in situ from ammonium
phosphinate and hexamethyldisilazane, as in our previous work.14 This reaction was initially
attempted using simple model dihalides 1,5-dibromopentane and 1,5-diiodopentane, and also
pentane-1,5-ditriflate. Various reaction conditions were explored, with the best results being
obtained in refluxing mesitylene, using the conditions of Frost.12 When the same reaction was
performed on the halides 7 and 8, and also on the triflate 9, only the dibromide 7 gave
satisfactory results. The resulting phosphinic acid 18 was not isolated, but was immediately
converted into the methyl ester 1915 using diazomethane, which allowed easier purification by
flash chromatography. The methyl ester 19 was formed as a mixture of two diastereoisomers,
resulting from the new stereogenic centre at the phosphorus atom.
The acetonide protecting group in 19 was removed using conditions which also converted the
resulting diol into the diacetate 6 in situ. Reaction of 6 with adenine in the presence of several
equivalents of tin(IV) chloride, using the methodology of Saneyoshi and Satoh,16 afforded the
protected nucleotide analogue 5.17 The 1-acetate group in 6 served as a leaving group for the
introduction of the adenine base, with neighbouring group participation from the 2-acetyl being
used to control the stereochemistry, ensuring formation of the b-substitution product. Hydroly-
sis of both the methyl ester and acetate protecting groups in 5 in one step, using lithium
hydroxide, was followed in the NMR tube by recording a time series of NMR spectra, with the
methyl ester being removed much more slowly than the 2%-acetate group.

8214
Scheme 3.
In conclusion we have demonstrated a viable synthetic route to a phosphinic acid analogue of
cyclic AMP, involving a double Arbuzov-type reaction on the dibromide intermediate 7 as the
key step.
Acknowledgements
We thank the EPSRC for financial support.
References
1. Francis, S. H.; Corbin, J. D. In Signal Transduction; Heldin, C.-H.; Purton, M., Eds.; Chapman & Hall: London,
1996.
2. Hancock, J. T. Cell Signalling; Longman: Harlow, 1997.
3. Albrecht, H. P.; Jones, G. H.; Moffatt, J. G. J. Am. Chem. Soc. 1970, 92, 5510–5511.
4. Malchow, D.; Fuchila, J.; Jastorff, B. FEBS Lett. 1973, 34, 5–9.
5. Drummond, G. I.; Powell, C. A. Mol. Pharmacol. 1970, 6, 24–30.
6. Giannattasio, M.; Sica, G.; Macchia, V. Phytochemistry 1974, 13, 2729–2733.
7. Legler, G.; Pohl, S. Carbo. Res. 1986, 155, 119–129.
8. Funabashi, M.; Sato, H.; Yoshimura, J. Bull. Chem. Soc. Jpn. 1976, 49, 788–790.
9. Mazur, A.; Tropp, B. E.; Engel, R. Tetrahedron 1984, 40, 3949–3956.
10. Jones, J. K. N.; Thompson, J. L. Can. J. Chem. 1957, 35, 955–959.
.

8215
11. Block, E. Org. React. 1984, 30, 457–566.
12. Montchamp, J.-L.; Tian, F.; Frost, J. W. J. Org. Chem. 1995, 60, 6076–6081.
13. Kurdyumova, N. R.; Ragulin, V. V.; Tsvetkov, E. N. Zh. Obshch. Khim. 1994, 64, 419-422 (Chem. Abs. 1994,
121, 205525).
14. Boyd, E. A.; Regan, A. C.; James, K. Tetrahedron Lett. 1992, 33, 813–816.
15. 1H NMR data for 19: lH (300 MHz, CDCl3) 1.27 (3H, s), 1.46 (3H, s), 1.46–2.35 (7H, m), 3.58 (1H, m), 3.68 (3
3
H, d, JPOCH 11.0), 4.43 (1H, dd, J 4.0, 4.0), 5.77 (1H, d, J 4.0). Evidence for a second diastereroisomer from
presence of extra set of peaks at l 4.52 and 5.77.
16. Saneyoshi, M.; Satoh, E. Chem. Pharm. Bull. 1979, 27, 2518–2521. For an example of the use of this
methodology, see: Maguire, M. P.; Feldman, P. L.; Rapoport, H. J. Org. Chem. 1990, 55, 948-955.
3
17. NMR data for 5: lH (300 MHz, CD3OD) 1.75–2.30 (7H, m), 2.02 (3H, s), 3.79 (3H, d, JPOCH 11.0), 4.04 (1 H,
m), 4.49 (1H, m), 6.03 (1 H , d, J 3.0), 7.98 (1H, s), 8.10 (1H, s). Evidence for a second diastereroisomer from
extra peaks superimposed on those at l 4.04, 4.49, and 6.03. lP (122 MHz, CD3OD) 56.94 and 58.50 (two
diastereoisomers).
.
.