Facile dephosphonylation of β-ketophosphonic acids: Mechanistic studies

Article abstract of DOI:10.1021/ol060519l

We have found that β-ketophosphonic acids can undergo facile dephosphonylation under fairly mild conditions. The rate of dephosphonylation is dependent on the electronic nature of the substituent on the carbon atom α to phosphorus, with electron-withdrawing groups accelerating the process. ³¹P NMR studies were used to probe the mechanism for the process.

Full text of DOI:10.1021/ol060519l

ORGANIC
LETTERS
2006
Vol. 8, No. 16
3429-3431
Facile Dephosphonylation of
-Ketophosphonic Acids: Mechanistic
Studies
â
Michael J. Hawkins, Eugene T. Powell, Gregory C. Leo, Diane A. Gauthier,
Michael N. Greco,* and Bruce Maryanoff
Research & Early DeVelopment, Johnson & Johnson Pharmaceutical Research &
DeVelopment, Spring House, PennsylVania 19477-0776
mgreco@prdus.jnj.com
Received March 2, 2006 (Revised Manuscript Received June 16, 2006)
ABSTRACT
We have found that
is dependent on the electronic nature of the substituent on the carbon atom
the process. 31P NMR studies were used to probe the mechanism for the process.
â-ketophosphonic acids can undergo facile dephosphonylation under fairly mild conditions. The rate of dephosphonylation
r
to phosphorus, with electron-withdrawing groups accelerating
The phosphonate functionality has played an important role
in organic chemistry, especially through the Horner-
Wadsworth-Emmons olefination reaction, which has been
used extensively to construct the carbon framework of
diverse synthetic targets.1-2 Phosphonates and phosphonic
acids have also been important in the bioorganic arena. In
fact, there are several therapeutic agents that contain the
phosphonate moiety as a critical pharmacophore.3 With
certain antiviral nucleosides, replacement of a phosphate by
a phosphonate has provided analogues that are resistant to
dephosphorylation, thereby paving the way for drugs against
human immunodeficiency virus (HIV), hepatitis B virus, and
cytomegalovirus.4
decarboxylation of â-ketocarboxylic acids, the spontaneous
dephosphonylation of â-ketophosphonic acids is quite rare.
A literature search revealed just a single report of dephos-
phonylation of a â-ketophosphonic acid.6 Specifically, Kluger
found that the monosodium salt of acetonylphosphonic acid
thermally decomposes on melting around 150 °C to produce
acetone and sodium metaphosphate.6 A related base-catalyzed
dephosphonylation of diethyl â-ketophosphonate derivatives
containing electrophilic keto groups has also been reported.7
To shed light on this area, we have examined the dephos-
phonylation of structural analogues of 1 with respect to
stereoelectronic influences. In addition, we conducted 31P
(5) (a) de Garavilla, L.; Greco, M. N.; Sukumar, N.; Chen, Z.; Pineda,
A. O.; Mathews, F. S.; Di Cera, E.; Giardino, E. C.; Wells, G. I.; Haertlein,
B. J.; Kaufman, J. A.; Corcoran, T. J.; Derian, C. K.; Eckardt, A. J.;
Damiano, B. P.; Andrade-Gordon, P.; Maryanoff, B. E. J. Biol. Chem. 2005,
280, 18001. (b) Greco, M. N.; Hawkins, M. J.; Powell, E. T.; Almond, H.
R., Jr.; Corcoran, T. W.; de Garavilla, L.; Kauffman, J. A.; Recacha, R.;
Chattopadhyay, D.; Andrade-Gordon, P.; Maryanoff, B. E. J. Am. Chem.
Soc. 2002, 124, 3810.
We have been interested in interfering with the serine
proteases cathepsin G and chymase, which are involved in
inflammatory processes. Our research effort led us to a novel
class of inhibitors that contain a key â-ketophosphonic acid
group.5 In particular, we identified and biologically charac-
terized potent, dual cathepsin G/chymase inhibitor 1.5b During
our studies, we noticed that 1 has a tendency to dephospho-
nylate to 2 under fairly mild conditions. In contrast to the
(6) Kluger, R. J. Org. Chem. 1973, 38, 2721. For an isolated case, see:
Denmark, S. E.; Marlin, J. E. J. Org. Chem. 1991, 56, 1003.
(7) (a) Piettre, S. G.; Girol, C.; Schelcher, C. G. Tetrahedron Lett. 1996,
37, 4711. (b) Thenappan, A.; Burton, D. J. J. Org. Chem. 1991, 56, 273.
(c) Thenappan, A.; Burton, D. J. Tetrahedron Lett. 1989, 30, 6113. (d)
Thenappan, A.; Burton, D. J. Tetrahedron Lett. 1989, 30, 3641. (e) Among
the possibilities in Table 1, Horner-Emmons elimination is least likely
because related substrates have been shown to be thermally stable. See:
Phillion, D. P.; Cleary, D. G. J. Org. Chem. 1992, 57, 2763.
(1) Maryanoff, B. E.; Reitz, A. B. Chem. ReV. 1989, 89, 863.
(2) Wiemer, D. P. Tetrahedron 1997, 53, 16609.
(3) Fields, S. C. Tetrahedron 1999, 55, 12237.
(4) Hwang, J.; Choi, J. Drugs Future 2004, 29, 163.
10.1021/ol060519l CCC: $33.50
© 2006 American Chemical Society
Published on Web 07/04/2006

NMR studies to probe the mechanism of this dephospho-
nylation process. The results from this study are reported
herein.
spontaneous dephosphonylation by elimination of phosphite
(Figure 1, eqs 1-3, respectively). From our results in Table
Aqueous solutions of the bistromethamine salt of 1 proved
to be unstable on standing at room temperature over a wide
pH range. We isolated and identified the major decomposi-
tion product as dephosphonylated material 2. This result
prompted us to investigate this facile dephosphonylation
process in a more systematic manner to gain a better
understanding of the factors that are responsible. Thus, we
exposed the monosodium salts of several â-ketophosphonic
acids to a refluxing solution of water-acetonitrile (1:1) and
calculated reaction half-lives for the dephosphonylation
(Table 1). Substrates 1 and 3-6, each of which has benzene-
Figure 1. Possible mechanisms for dephosphonylation of â-ke-
tophosphonic acids: (1) Horner-Emmons-like decomposition; (2)
addition of water to phosphorus; (3) spontaneous decomposition.
1, it is not possible to discriminate between these mecha-
nisms.
To investigate the mechanism further, and in a more direct
manner, we performed the dephosphonylation on 8 by using
18O-enriched water (water-acetonitrile, 1:1; reflux) and then
characterized the products by 31P NMR (Scheme 1). De-
Table 1. Dephosphonylation of â-Ketophosphonic Acidsa
R1
R2
t1/2 (min)b
Scheme 1. Dephosphonylation of 8 in 18O-Enriched Water
1c
3
4
5
6
7
8
- -
- -
50
22
25
27
19
2-naphthyl
Ph
2-naphthyl
Me
1-naphthyl
1-naphthyl
Ph
Ph
Ph
2-naphthyl
Me
4-NO2C6H4
169
1.5
a Reactions were conducted at 78 °C with a concentration of 0.01 mM
at pH 6.8-7.5. b Half-lives were determined by monitoring the appearance
of dephosphonylated product over time by HPLC. c See text for structure.
phosphonylation of 8 with natural-abundance water served
as a control. After 15 min, the reactions were separated into
organic and aqueous phases. As expected, the major product
isolated from the organic phase was 9. From the reaction
using 18O-enriched water, no 18O isotope shift of the ketone
carbonyl of 9 was observed by 13C NMR, thus ruling out
dephosphonylation via the hydrated ketone carbonyl (Figure
1, eq 1). The major component of the aqueous phase was
18O-labeled sodium phosphate (10), as detected by the 18O
isotope-induced shift of the 31P signal of phosphate by using
31P NMR (Figure 2); no other products from 18O incorpora-
tion were detected. Formation of 18O-labeled sodium phos-
phate indicates that a Horner-Emmons-type elimination
process (Figure 1, eq 2) is unlikely.8 However, a Horner-
Emmons-type elimination process cannot be ruled out for
the reported phosphonate ester examples,7 especially when
one takes into account the high degree of ketone carbonyl
based aromatic groups for R2, exhibit similar half-lives.
However, for 7, which possesses an aliphatic methyl group
for R2, we observed a comparatively slow dephosphonylation
rate. Compound 8, which has an electron-withdrawing nitro
aromatic group for R2, dephosphonylated more readily. These
results, albeit limited in scope, illustrate the role of the R2
substituent and reflect on the importance of an electron-
withdrawing group in facilitating the dephosphonylation. In
contrast, the R1 group appears to exert little influence (cf. 5
vs 6).
In contemplating the mechanistic pathway for this reaction,
three possible alternatives come to mind. There could be (1)
a Horner-Emmons-like process involving the hydrated form
of the ketone carbonyl, (2) direct attack of water on the
phosphorus atom with elimination of phosphate, or (3)
(8) On exposure of the corresponding diethyl phosphonate of 3 to the
conditions of Table 1 for up to 2 h, no dephosphonylation was observed.
3430
Org. Lett., Vol. 8, No. 16, 2006

Dephosphonylation of 3 was examined at pH 3 (refluxing
1:1 water-acetonitrile) and pH 10 (refluxing 1:1 buffer-
acetonitrile), and the product mixtures were analyzed by
HPLC after a 10-min reaction course.11 At pH 3, 7% of the
corresponding ketone was formed; at pH 10, 19% was
formed. This result indicates that there is a slower rate of
dephosphonylation at lower pH. The facility of this reaction
can be further appreciated by the fact that 8 dephosphon-
ylated to the extent of 28% in 5.5 h at 23 °C. A dependency
of reaction rate on pH is consistent with our suggested
mechanisms (eqs 2 and 3).
In summary, we have found that R-substituted â-keto-
phosphonic acids dephosphonylate under relatively mild
conditions. The rate of dephosphonylation appears to be
dependent on the nature of the substituent at the R carbon,
with electron-withdrawing groups facilitating the process.
Our 31P NMR results support two possible mechanisms, one
involving addition of water to the phosphorus atom, followed
by cleavage of the phosphorus-carbon bond to form ketone
and phosphate products (Figure 1, eq 2), and the other
involving spontaneous decomposition (Figure 1, eq 3)
followed by hydration of monomeric metaphosphate.
Figure 2. 31P NMR spectra showing the isotope shift from 18O
(5.3 Hz). (A) Sample taken from the aqueous fraction of the reaction
involving 18O water. (B) Sample from A that was spiked with the
aqueous fraction of the reaction involving 16O water. Spectra are
referenced to 16O phosphate (0.0 ppm, T ) 5 °C).
hydration expected for these substrates.7b Sodium metaphos-
phate is the phosphate-containing species expected to form
by spontaneous decomposition (Figure 1, eq 3). Because
monomeric metaphosphate (eq 3) would be a highly reactive
species,9 it is conceivable that it could hydrate under the
reaction conditions to form sodium phosphate, thus making
the pathways of eqs 2 and 3 indistinguishable.10 Therefore,
the 31P NMR studies do not provide definitive evidence for
a single pathway.
Acknowledgment. We thank Prof. Daniel Comins for
valuable suggestions. We appreciate receiving some cogent
remarks from the referees.
Supporting Information Available: Procedures and
spectral data for all new compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL060519L
(9) Westheimer, F. H. Chem. ReV. 1981, 81, 313.
(10) On heating sodium metaphosphate (City Chemical LLC) in 18O-
enriched water at 78 °C, no formation of 18O-labeled sodium phosphate
was observed by 31P NMR over 2.5 h.
(11) The free phosphonic acid was used at each pH. Rates were compared
at 10 min because the product mixture became complex at pH 10 after 10
min.
Org. Lett., Vol. 8, No. 16, 2006
3431