An unsymmetrical approach to the synthesis of bismethylene triphosphate analogues

Article abstract of DOI:10.1021/ol0615432

(Chemical Equation Presented) A protected, unsymmetrical bismethylene triphosphate analogue was prepared by sequential Michaelis-Arbuzov reactions on ethyl bis(halomethyl)phosphinates. This species was monodeprotected at one of the terminal phosphonate gr

Full text of DOI:10.1021/ol0615432

ORGANIC
LETTERS
2006
Vol. 8, No. 19
4243-4246
An Unsymmetrical Approach to the
Synthesis of Bismethylene Triphosphate
Analogues
Scott D. Taylor,*,† Farzad Mirzaei,† and Stephen L. Bearne‡
Department of Chemistry, UniVersity of Waterloo, 200 UniVersity AVenue West,
Waterloo, Ontario, Canada, N2L 3G1, and Department of Biochemistry and Molecular
Biology, Dalhousie UniVersity, Halifax, NoVa Scotia, Canada, B3H 1X5
s5taylor@sciborg.uwaterloo.ca
Received June 22, 2006
ABSTRACT
A
protected, unsymmetrical bismethylene triphosphate analogue was prepared by sequential Michaelis−Arbuzov reactions on ethyl
bis(halomethyl)phosphinates. This species was monodeprotected at one of the terminal phosphonate groups in high yield. The resulting
monodeprotected compound was used to achieve the first syntheses of the bismethylene triphosphate analogues of UTP and CTP.
Polyphosphorylated compounds, such as nucleotides, co-
enzymes, vitamins, and isoprenoids, are widespread in nature
and have key roles in numerous biological functions. Non-
hydrolyzable polyphosphate analogues have been used
extensively as probes and inhibitors of enzymes and other
proteins that bind or hydrolyze polyphosphates.1 The meth-
ylene group figures prominently among the various groups
that have been used to replace the labile pyrophosphate
oxygens. Although methylene analogues of diphosphates are
readily constructed, the synthesis of bismethylene triphos-
phate (BMT) analogues still represents a considerable
challenge.2 BMT derivatives are usually prepared by reacting
a nucleophile with trimetaphosphate 13 or, more commonly,
by reacting alkylammonium salts of 2 with tosyl ester
acceptors1e-h,4a,b (Scheme 1). However, these reactions
Scheme 1
† University of Waterloo.
‡ Dalhousie University.
proceed in very low yields and usually require a lengthy and
difficult purification. An alternative approach is to couple
the monoacid of a protected BMT of type 3 to an acceptor
(Scheme 2). The advantage of this route is that the purifica-
tion issues encountered using the approaches outlined in
(1) For some examples see: (a) Biller, S.; Forster, C. Tetrahedron 1990,
46, 6645-6658. (b) Blackburn, G. M.; Langston, S. P. Tetrahedron Lett.
1991, 32, 6425-6428. (c) Morr, M.; Wray, V. Angew. Chem., Int. Ed. Eng.
1994, 33, 1394-1396. (d) Yanachkov, I.; Pan, J. Y.; Wessling-Resnick,
M.; Wright, G. E. Mol. Pharmacol. 1997, 51, 47-51. (e) Ma, Q.-F.; Kenyon,
G. L.; Markham, G. D. Biochemistry 1990, 29, 1412. (f) Dalton, A.; Hornby,
D. P.; Langston, S. P.; Blackburn, G. M. Biochem. J. 1992, 287, 871-879.
(g) Labataille, P.; Pelicano, H.; Maury, G.; Imbach, J.-L.; Gosselin, G.
Biooorg. Med. Chem. Lett. 1995, 5, 2315-2320. (h) Spelta, V.; Mekhalfia,
A.; Rejman, D.; Thompson, M.; Blackburn, G. M.; North, R. A. Br. J.
Pharmacol. 2003, 140, 1027-1034.
(3) Towbridge, D. B.; Yamamoto, D. M.; Kenyon, G. L. J. Am. Chem.
Soc. 1972, 94, 3816.
(4) (a) Stock, J. A. J. Org. Chem. 1979, 44, 3997-4000. (b) Blackburn,
G. M.; Guo, M. J.; Langston, S. P.; Taylor, G. E. Tetrahedron Lett. 1990,
31, 5637-5640.
(2) Klein, E.; Mons, S.; Valleix, A.; Mioskowski, M.; Lebeau, L. J. Org.
Chem. 2002, 67, 146-153 and references therein.
10.1021/ol0615432 CCC: $33.50
© 2006 American Chemical Society
Published on Web 08/18/2006

not possible, then an alternative would be to prepare an
unsymmetrical BMT (R1 * R2 in Scheme 4) since this would
Scheme 2
Scheme 4
Scheme 1 are for the most part overcome since purification
of the coupled product is readily accomplished with flash
chromatography. The final BMT derivative is obtained after
deprotection of the coupled product under mild conditions.
This route, pioneered by Mioskowski and co-workers, has
been used to prepare BMT analogues of AZT,5 thiamine
triphosphate,6 and Ap3A and Gp3A.2 The AZT analogue was
recently used to obtain antibodies against AZT triphosphate.7
The tetrabenzyl derivative 4 was used for these studies and
the benzyl groups were removed at the end of the syntheses
with TMSBr5,6 or by hydrogenolysis.1 Although this route
appears attractive, the synthesis of compound 4 and its
precursor, symmetrical pentabenzyl ester 5 (Scheme 3), has
allow for the selective cleavage of one of the terminal esters.
Methyl and ethyl protecting groups should be applicable
since TMSBr can also readily remove such esters. Here we
report a straightforward synthesis of 5 and our attempts to
achieve a selective deprotection of one of its terminal
phosphonate moieties. We also report the synthesis of the
unsymmetrical BMT 7 via sequential Michaelis-Arbuzov
reactions on bishalomethylene phosphinates. A highly selec-
tive deprotection of the terminal methylphosphonate moiety
in 7 yielded monodeprotected 8 that was used for the first
synthesis of the BMT analogues of UTP and CTP.
Due to the conflicting reports on the monodeprotection
of 5, we initiated our studies by preparing 5 and determining
if a selective monodeprotection of one its terminal benzyl
groups could be achieved. Saady et al. reported that benzyl
derivative 5 could be prepared in a 71% yield by a
Michaelis-Arbuzov reaction between benzyl bischloro-
methylphosphinate (9) with 4 equiv of tribenzyl phosphite
(TBP) at 6-10 Torr at 140 °C (Scheme 5).9 However, Mons
Scheme 3
Scheme 5
been fraught with difficulties which limits the appeal of this
approach. For example, Saady et al. reported that 5 could
be selectively monodeprotected at one of the terminal
phosphonate moieties in 95-97% yield with 1 equiv of
DABCO or quinuclidine in refluxing toluene (Scheme 3).8
However, it was later reported that this reaction gives a
mixture of 4 and 6 that could not be separated.5,6 Conse-
quently, a mixture of 4 and 6 was used to prepare BMT’s of
thiamine and AZT.5,6 After complete deprotection of the
coupled products the resulting regioisomers were separated
by HPLC. This approach would be considerably improved
if a highly selective monodeprotection of one of the terminal
phosphonates of compound 5 could be achieved. If this is
et al. stated that the synthesis of 5 using this approach is
difficult to reproduce.10 This is due to a competing reaction
of the in situ generated benzyl chloride with TBP. The TBP
is rapidly consumed by reaction with benzyl chloride unless
the benzyl chloride is efficiently removed from the reaction
mixture as it is formed.9 Consequently, these workers
reported an alternative approach to 5 by first preparing
selenophosphonate 10 and then reacting the anion of 10 with
N,N-dimethylphosphonamidous dichloride, followed by reac-
tion with benzyl alcohol and then oxidation with MCPBA
(Scheme 5). A yield of 37% was reported over these three
steps.2,10
(5) Brossette, T.; Le Faou, A.; Goujon, L.; Valleix, A.; Creminon, C.;
Grassi, J.; Mioskowski, C.; Lebeau, L. J. Org. Chem. 1999, 64, 5083-
5090.
(6) Klein, E.; Nghiem, H.-O.; Valliex, A.; Mioskowski, C.; Lebeau, L.
Chem. Eur. J. 2002, 8, 4649-4655.
(7) Becher, F.; Schlemmer, D.; Pruvost, A.; Nevers, M.-C.; Goujard, S.
J.; Guerreiro, C.; Brossette, T.; Lebeau, L.; Creminon, C.; Grassi, J.; Benech,
H. Anal. Chem. 2002, 74, 4220-4227.
(9) Saady, M.; Lebeau, L.; Mioskowski, C. HelV. Chim. Acta 1995, 78,
670-678.
(8) Saady, M.; Lebeau, L.; Mioskowski, C. J. Org. Chem. 1995, 60,
2946-2947.
(10) Mons, S.; Klein, E.; Mioskowski, C.; Lebeau, L. Tetrahedron Lett.
2001, 2, 5439-5442.
4244
Org. Lett., Vol. 8, No. 19, 2006

We initially examined the selenophosphonate route2,10 to
prepare 5. However, in our hands, this procedure gave 5 in
yields <10%. Consequently, we decided to examine Saady
et al.’s original approach using tribenzyl phosphite and
phosphinate 9.9 Using the conditions reported by Saady et
al.9 mentioned earlier only trace amounts of 5 were obtained.
Nevertheless, it was found that by reacting 9 in the presence
of 6.5 equiv of tribenzyl phosphite at 155 °C and 0.05 Torr
for 16 h, a mixture of 5, 11, and dibenzyl benzylphosphonate
could be obtained, which was separated by careful chroma-
tography (Scheme 6).11 Phosphinate 11 was then reacted
of products. After some experimentation, we found that a
mixture of 4 and 6 in a 4:1 ratio could be obtained in a 67%
yield by reacting 5 with 1.0 equiv of DABCO in benzene at
60 °C for 6 h. We also attempted the selective removal of
one of the terminal benzyl groups using a variety of
nucleophiles (NaI, LiBr, thiols, amines) in a variety of
solvents; however, all of these procedures gave inseparable
mixtures of 4 and 6, which is consistent with the results of
Brossette et al. and Klein et al.5,6 Consequently, we decided
to prepare unsymmetrical BMT 7.
Our approach to 7 was by sequential Michaelis-Arbuzov
reactions on ethyl bis(halomethyl)phosphinates (Scheme 8).
Scheme 6
Scheme 8
again in a similar manner with 4 equiv of tribenzyl phosphite
to give 5 in an overall 52% yield (Scheme 6). Although this
procedure yielded 5, it was tedious and the separation of 5
from dibenzyl benzylphosphonate and 11 required several
columns. Consequently, an alternative route to 5 was
developed (Scheme 7). Ethyl derivative 13 was readily
Scheme 7
Compound 15 was easily prepared in a 54% yield with the
procedure of Medved et al., which involved reacting 12 with
1.8 equiv of TEP at 175 °C for 7 h.14 However, attempts to
produce unsymmetrical BMT 7 by reaction of 15 with
refluxing excess trimethyl phosphite (TMP) were not suc-
cessful due to the low reactivity of 15 at this temperature.15
Therefore, bromo and iodo derivatives 16 and 18 were
prepared anticipating that these would be more reactive than
15. Iodo phosphinate 18 was prepared in 59% yield by
reacting ester 12 with KI in DMF at 105 °C for 3 h.16 Bromo
derivative 16 has never before been reported. However, it
was easily prepared by reacting 1417 with thionyl bromide
in refluxing CHCl3 for 6-7 h and then reacting the crude
phosphinyl bromide with ethanol in the presence of triethyl-
amine (61% yield, 2 steps). Reaction of 16 with 1.8 equiv
of TEP at 150 °C for 2 h gave bromo derivative 17 in 56%
yield while reaction of 18 with 5 equiv of TEP at 150 °C
for 75 min gave iodo derivative 19 in 52% yield. Reaction
prepared in 85% yield by reaction of phosphinate 12 with
triethyl phosphite (TEP).12a,b Reaction of 13 with 7.5 equiv
of TMSBr followed by treatment with MeOH and then
reaction of the resulting crude acid with 10 equiv of
tribenzylorthoformate at 150 °C for 3 h gave compound 5
in a 70% yield (from compound 13).13 Purification by column
chromatography was straightforward and compound 5 was
consistently obtained in good yield even on a multigram
scale. The selective removal of a terminal benzyl group in 5
was first attempted with the conditions described by Saady
et al. (1.0 equiv of DABCO or quinuclidine in toluene, reflux
2 h).8 However, in our hands, this gave a complex mixture
(11) The addition of more tribenzyl phosphite to the reaction resulted in
an increase in byproducts.
(14) Medved, T. Y.; Polikarpov, Y. M.; Pisareva, S. A.; Matrosov, E. I.,
Kabachnik, M. I. IzV. Akad. Nauk. USSR, Ser. Khim. 1968, 9, 2062-2070.
(15) We also attempted this reaction with high boiling point solvents
such as various xylenes; however, the reaction was very slow and after 24
h only small amounts of the desired products formed.
(16) Mukhametzyanova, E. K.; Panteleeva, A. R.; Shermergorn, I. M.
IzV. Akad. Nauk. SSSR, Ser. Khim. 1967, 7, 1597-1598.
(17) Maier, L. J. Organomet. Chem. 1979, 178, 157-169.
(12) (a) Bel’skii, V. E.; Zyablikova, T. A.; Panteleeva, A. R.; Sherm-
ergorn, I. M. Dokl. Akad. Nauk SSSR 1967, 177, 340-343. (b) Maier, L.
HelV. Chim. Acta 1969, 52, 827-845.
(13) The tetrabenzyl ester of methylene diphosphonate has been prepared
with this approach: Gil, L.; Opas, E. E.; Rodan, G. A.; Ruel, R.; Seedor,
J. G.; Tyler, P. C.; Young, R. N. Bioorg. Med. Chem. 1999, 7, 901-919.
Org. Lett., Vol. 8, No. 19, 2006
4245

of iodo derivative 19 with excess TMP at 125 °C gave
compound 7 in a 72% yield. However, this reaction was
tedious to perform since the methyl iodide produced in the
reaction rapidly converted all of the TMP to dimethyl
methylphosphonate well before the reaction was complete.
Addition of more TMP resulted in an increase in byproducts.
To obtain reasonable yields, the reaction had to be monitored
by 31P NMR and when the TMP was consumed the reaction
was cooled to room temperature and the dimethyl meth-
ylphosphonate removed by high vacuum rotary evaporation.
More TMP was then added and the reaction continued and
this process was repeated until all of 19 was consumed. The
reaction of bromo derivative 17 with excess TMP at 130 °C
gave a similar yield of compound 7. However, dimethyl
methylphosphonate was produced at a much lower rate and
the reaction was much easier to perform. Selective depro-
tection of 7 was achieved by reaction with 1.0 equiv of KCN
in DMF at 70 °C for 5 h followed by conversion of the
potassium salt to the free acid by passage through a Dowex
H+ ion exchange column. This gave acid 8 in 90% yield.
Alternatively, 8 could be obtained in a 90% yield as its
triethylammonium salt (8a) by reacting 7 with 1.2 equiv of
PhSH and 3 equiv of Et3N in THF at room temperature for
24 h.
Scheme 9
products suggested that some cleavage of the BMT moiety
from the nucleoside had occurred, most likely during the
TMSBr reaction. Nevertheless, after purification by RP-
HPLC and conversion to their sodium salts with a Dowex
Na+ ion-exchange resin, pure uridine and cytidine BMT’s
24 and 25 were obtained in 62% and 50% yield, respectively.
In summary, straightforward syntheses of the symmetrical
BMT 5 and unsymmetrical BMT 7 were developed. A
selective monodeprotection could not be accomplished with
5 but was readily achieved in high yield with 7. The resulting
monodeprotected compounds 8 and 8a were used to effect
the first syntheses of the BMT analogues of UTP (24) and
CTP (25). This approach should be readily applicable to the
synthesis of other BMT derivatives of biological interest.
To determine if compound 8 or 8a could be used in the
synthesis of BMT’s of biomolecules, 8 was coupled to 2′,3′-
O,N3-tribenzoyluridine (20) via a Mitsunobu reaction to give
compound 22 in a 79% yield (Scheme 9). Surprisingly, we
failed to obtain coupled product 23 when reacting 8 with
2′,3′-O,N4-tribenzoylcytidine (21) under the same conditions.
However, employing the triethylammonium salt 8a and using
modified Mitsunobu conditions18 the desired coupled product
23 was obtained in 61% yield. Complete deprotection of 22
and 23 was achieved by subjecting them to TMSBr for 24 h
followed by treatment with aq NH4OH-MeOH.19 The nega-
tive electrospray mass spectra of the crude deprotected
Acknowledgment. This work was supported by a Natural
Sciences and Engineering Research Council (NSERC) of
Canada Discovery grant to S.D.T. and an NSERC Collabora-
tive Health Research Project Grant to S.L.B. and S.D.T.
Supporting Information Available: Preparation proce-
dures and characterization data for 5, 7, 8, 8a, 16, 17, 19,
and 22-25. This material is available free of charge via the
Internet at http://pubs.acs.org.
(18) Norbeck, D. W.; Kramer, J. B.; Lartey, P. A. J. Org. Chem. 1987,
52, 2174-2179.
(19) Brossette et al. have reported that some anomerization occurred after
subjecting a benzyl-protected BMT analogue of an AZT derivative to
1
TMSBr (ref 5). However, 31P and H NMR analysis of our crude reaction
mixtures revealed that anomerization had not taken place. This could be
due to participation of the benzoyl protecting group at the 2′ position.
OL0615432
4246
Org. Lett., Vol. 8, No. 19, 2006