Comparison of cyclic and acyclic phosphites by selective phosphorylation. Synthesis of phosphatidylinositol 4-phosphate

Article abstract of DOI:10.1246/cl.2002.292

To compare the synthetic utility of cyclic xylylene N,N-diethylphosphoramidite and its acyclic analog, dibenzyl ester, both phosphites derived from them with dipalmitoylglycerol were subjected to the selective phosphorylation at the OH-3 position in 1,2:4

Full text of DOI:10.1246/cl.2002.292

292
Chemistry Letters 2002
Comparison of Cyclic and Acyclic Phosphites by Selective Phosphorylation. Synthesis of
Phosphatidylinositol 4-Phosphate
Yutaka Watanabe,ꢀ Hirofumi Munetsugu, and Minoru Hayashi
Department of Applied Chemistry, Faculty of Engineering, Ehime University, Matsuyama 790-8577
(Received November 1, 2001; CL-011092)
To compare the synthetic utility of cyclic xylylene N,N-
Table 1. Comparison of phosphorylation
diethylphosphoramidite and its acyclic analog, dibenzyl ester,
both phosphites derivedfrom them with dipalmitoylglycerol were
subjected to the selective phosphorylation at the OH-3 position in
1,2:4,5-di-O-cyclohexylidene-myo-inositol in the presence of
pyridinium tribromide, and the cyclic version showed to be much
superior to the acyclic. The 3-O-phosphorylated product was
conveniently converted into phosphatidylinositol 4-phosphate.
We reported xylylene N,N-diethylphosphoramidite (XEPA,
1a) as a promising phosphitylating agent for the synthesis of
phosphoric monoesters1 and diesters.2 The N,N-dimethyl3 and
N,N-diisopropyl4 versions also appeared after our first report.
These reagents have been used for phosphorylation of inositols5
and sugars,6 and for glycosylation.7 Their acyclic analog,
dibenzyl phosphoramidite8 1b has been more widely utilized
than the cyclic ones. However, preparation9 and purification of
XEPA are much easier than those of 1b, and both reagents have an
identical reactivity.10 Furthermore, a cyclic phosphate derived
from 1a is particularly prone to crystallize. These advantages of
XEPA prompted us to demonstrate its efficiency in terms of
selective phosphorylation of diol 4 with the phosphite 3a derived
from 1a using the phosphite-pyridinium tribromide method,11
that has been well documented togive selectively 1-O-phosphates
starting from 1,2-free inositol derivatives.12 We describe here the
comparison of the selective phosphorylation and transformation
of the resultant 1-O-phosphate 5 to phosphatidylinositol 4-
phosphate {PI(4)Pg, which is known to be a metabolic precursor
of
biologically
important
phosphatidylinositol
4,5-
bisphosphate.13
phosphoryl group seems to be sterically more hindered rather than
the xylylene. A more plausible explanation for the selectivity may
be mainly put forward by the consideration of the reactive
intermediate.15 The phosphorylation first forms the bromophos-
phonium salt 7, that reacts with an alcohol and then decomposes
to the final phosphorylation product 9 (Scheme 1). The other
pathway involves the first decomposition of 7 into the
phosphorobromidate 10 and then the phosphorylation of the
alcohol. When both phosphites 3 were treated respectively with 3
molar equivalents of butyl alcohol (R2OH in Scheme 1) in the
presence of PTB and pyridine in CH2Cl2, 3a gave a fair amount of
the ester exchange product 12 (31%) accompanied with the
normal product 9 (47%). Since 9 did not react with butyl alcohol
under the identical conditions, the major pathway for the cyclic
phosphite involves the reaction of the phosphonium salt 7 with the
alcohol without decomposition to the phosphorobromidate, to
form 8, which, when further reacts with R2OH, is converted to the
undesired product16 12. In the case of 3b, the ester exchange
reaction scarcely took place. Therefore, an acyclic phosphite such
as dibenzyl ester would likely be transformed via the phosphoro-
bromidate pathway. These results shows that the acyclic 7 is more
The attempts to selectively introduce a phosphate function at
the OH-3 position in DL-1,2 : 4,5-di-O-cyclohexylidene-myo-
inositol14 were made using the reaction with DL-dipalmitoylgly-
cerol phosphites 3 in the presence of pyridinium tribromide
(PTB), 2,6-lutidine, and anhydrous calcium sulfate. Both
phosphites were derived respectively by the reaction of 1a and
1b with racemic 1,2-di-O-palmitoylglycerol 2 in the presence of
1H-tetrazole. Cyclic phosphite 3a gave a better yield of the
desired 3-O-phosphorylated product 5a with higher selectivity
(Table 1). In contrast to the result, acyclic dibenzyl phosphite 3b
yielded fair amounts of 3,6-diphosphate, even when a limited
amount of the phosphite was used.
The difference in selectivity of the phosphorylation between
the cyclic and acyclic derivatives can not be explained in terms of
the steric bulkiness of a protecting group, because the dibenzyl
Copyright Ó 2002 The Chemical Society of Japan

Chemistry Letters 2002
293
This work was supported financially by Venture Business
Laboratory of Ehime University.
This paper is dedicated to Professor Teruaki Mukaiyama on
the occasion of his 75th birthday.
References and Notes
1
2
3
4
5
Y. Watanabe, Y. Komoda, K. Ebisuya, and S. Ozaki, Tetrahedron
Lett., 31, 255 (1990).
Y. Watanabe, Y. Komoda, and S. Ozaki, Tetrahedron Lett., 33, 1313
(1992).
G. M. Salomonczyk and K. M. Pietrusiewicz, Tetrahedron Lett., 32,
6167 (1991).
´
N. Moitessier, F. Chretien, Y. Chapleur, andC. Humeau, Tetrahedron
Lett., 36, 8023 (1995).
S. Ozaki, Y. Kondo, N. Shiotani, T. Ogasawara, and Y. Watanabe, J.
Chem. Soc., Perkin Trans. 1, 1992, 729; L. Schmitt, B. Spiess, and G.
Schlewer, Tetrahedron Lett., 33, 2013 (1992).
Scheme 1. Mechanistic pathways.
6
7
8
K. Fukase, Y. Aoki, I. Kinoshita, Y. Suda, M. Kurosawa, U.
Zaehringer, E. T. Rietschel, and S. Kusumoto, Tetrahedron Lett., 36,
8645 (1995); K. Fukase, Y. Fukase, M. Oikawa, W.-C. Liu, Y. Suda,
and S. Kusumoto, Tetrahedron, 54, 4033 (1998).
Y. Watanabe, C. Nakamoto, and S. Ozaki, Synlett, 1993, 115; Y.
Watanabe, C. Nakamoto, T. Yamamoto, and S. Ozaki, Tetrahedron,
50, 6523 (1994); L. Qiao, B. W. Murray, M. Shimazaki, J. Schultz,
and C.-H. Wong, J. Am. Chem. Soc., 118, 7653 (1996).
For simple alcohols, amino acids, and peptides: J. W. Perich and R. B.
Johns, Tetrahedron Lett., 28, 101 (1987); W. Bannwarth and A.
Trzeciak, Helv. Chim. Acta, 70, 175 (1987); H. B. A. de Bont, G. H.
Veeneman, J. H. van Boom, and R. M. J. Liskamp, Recl. Trav. Chim.
Pays-Bas, 106, 641 (1987). For inositols: K.-L. Yu and B. Fraser-
sensitive toward the bromide attack than the cyclic intermediate
in accordance with the report.17 These also suggest that, since the
selectivity depends principally on the reactivity, a phosphoro-
bromidate is more reactive than a phosphonium salt.
The XEPA-derived 3-phosphate 5a thus obtained was used
for the synthesis of PI(4)P. Phosphorylation of 5a using XEPA
gave 13 quantitatively. The deprotection procedure was first
carried out by removal of the cyclohexylidene groups. To prevent
the migration of the phosphate functions, we applied the method
using pyridinium poly(hydrogen fluoride), which was recently
demonstrated to efficiently deprotect acetals,18 giving 14 in 76%
yield. The structure of 14 was chemically confirmed by its
transformation to the tetrakis(chloroacetyl) derivative and the
NMR analysis. The final deprotection by catalytic hydrogenolysis
on Pd black in the presence of triethylammonium acetate
followed by the cation exchange to the Naþ form gave a racemic
and diastereomeric mixture of PI(4)P sodium salt19 in 97%. The
reaction in the absence of the acetate considerably conducted the
decomposition of the diester moiety in the product, presumably
because hydrogen bromide generated during the reaction caused
hydrolysis of the function.
Reid, Tetrahedron Lett., 29, 979 (1988); G. Baudin, B. I. Glanzer, K.
¨
S. Swaminathan, and A. Vasella, Helv. Chim. Acta, 71, 1367 (1988);
C. E. Dreef, C. J. J. Elie, P. Hoogerhout, G. A. van der Marel, and J. H.
van Boom, Tetrahedron Lett., 29, 6513 (1988).
B. A. Arbuzov, R. A. Kadyrov, V. V. Klochkov, R. P. Arshinova, and
A. V. Aganov, Izvest. Akad. Nauk SSSR, Ser. Khimicheskaya, 1982,
588.
9
10 Both phosphites 1a and 1b are now commercially available.
11 Y. Watanabe, E. Inada, M. Jinno, and S. Ozaki, Tetrahedron Lett., 34,
497 (1993).
12 K. K. Reddy, J. Rizo, and J. R. Falck, Tetrahedron Lett., 38, 4729
(1997); Y. Watanabe, ‘‘Studies in Natural Products Chemistry,
Vol. 18, Stereoselective Synthesis (Part K),’’ ed. by A. Rahman,
Elsevier, Amsterdam (1996), pp 391–456.
13 R. F. Irvine, Biochem. Soc. Trans., 23, 27 (1995).
14 Selective phosphorylation at the OH-3 in 1,2:4,5-diisopropylidene-
myo-inositol was carried out by the reaction with dipalmitoylglycerol
phenyl phosphate in the presence of N-mesitylenesulfonyltriazole
(60% yield) [M. Jones, K. K. Rana, J. G. Ward, and R. C. Young,
Tetrahedron Lett., 30, 5353 (1989).], and with diphenyl phosphoro-
chloridate (30%) [S.-K. Chung and Y. Ryu, Carbohydr. Res., 258,
145 (1994).].
15 J. K. Stowell and T. S. Widlandski, Tetrahedron Lett., 36, 1825
(1995). V. B. Oza andR. C. Corcoran, J. Org. Chem., 60, 3680 (1995).
16 L. S. Colle and E. S. Lewis, J. Org. Chem., 43, 571 (1978).
17 J. Gloede, Z. Chem., 28, 352 (1988).
18 Y. Watanabe, Y. Kiyosawa, A. Tatsukawa, and M. Hayashi,
Tetrahedron Lett., 42, 4641 (2001).
19 1H-NMR (400 MHz,CDCl3) ꢀ 0.79–0.95 (6H, brt, CH3), 1.04–1.49
(48H, complex, CH2), 1.51–1.68 (4H, m, ꢁ-CH2), 2.22–2.48 (4H, m,
ꢂ-CH2), 3.34–3.78 (4H, complex, InsH3, H5, GlyH3), 3.80–4.45 (6H,
complex, InsH1, H2, H4, H6, GlyH1), 5.04–5.12 (0.3H, m, GlyH2),
5.20–5.30 (0.7H, m, GlyH2); 31P-NMR (162 MHz, CDCl3) 3.73 (P1),
6.36 (P4); MS (FABꢁ, triethanolamine, M: C41H78O16P2Na2) m=z
911(M-Na), 889 (M-2Na þ H).
Scheme 2. Synthesis of phosphatidylinositol 4-phosphate.
In this paper, difference in the synthetic merit between seven-
membered cyclic phosphite, XEPA and acyclic dibenzyl has been
first shown.