Toward a stable hydroxyphosphorane.

Article abstract of DOI:10.1021/ol016941d

[reaction: see text] Hydroxy-pentaoxy-phosphoranes are transient intermediates formed during the hydrolysis of various phosphoesters. Kinetic analyses support the existence of such compounds, although they are not isolated. In an attempt to create a stabl

Full text of DOI:10.1021/ol016941d

ORGANIC
LETTERS
2002
Vol. 4, No. 2
201-203
Toward a Stable Hydroxyphosphorane
Robert E. Hanes, Jr., Vincent M. Lynch, Eric V. Anslyn,* and Kevin N. Dalby*
,†
The Department of Chemistry and Biochemistry and The College of Pharmacy,
DiVision of Medicinal Chemistry, The UniVersity of Texas at Austin,
Austin, Texas 78712
anslyn@ccwf.cc.utexas.edu
Received October 19, 2001
ABSTRACT
Hydroxy-pentaoxy-phosphoranes are transient intermediates formed during the hydrolysis of various phosphoesters. Kinetic analyses support
the existence of such compounds, although they are not isolated. In an attempt to create a stable example, 2 equiv of a ligand possessing
a very high effective molarity were attached to a central phosphorus. Instead of obtaining a hydroxyphosphorane, analysis by 31P and 19F NMR
spectroscopy and X-ray crystallography showed the product to be a phosphotriester. The reason for this is discussed.
Phosphoranes are transient intermediates in the hydrolysis
of phosphodi- and phosphotriesters. While the existence of
these intermediates is beyond dispute, the nature of the
protonation state and/or the lifetime of these compounds
remains controversial.1 Particularly, the stability of a dian-
ionic phosphorane and the pKa of its conjugate acid are
fundamental to determining the role of various functional
groups (metal ions, general acid/base catalysts) in the active
site of an enzyme during RNA and DNA hydrolysis. While
The choice of ligand is crucial to the stability of oxyphos-
Granoth and Martin determined the pKa of hydroxyphos-
phoranes.5 In the case studied by Ramirez, catechol formed
a five-membered ring with phosphorus. This study estab-
lished that the solution concentration of the hydroxyphos-
phorane increased as the temperature decreased and as the
relative solvent basicity increased.
phoranes derived from phosphinates,2 direct measurement
of the pKa of a pentaoxyphosphorane has remained an elusive
goal.3 Ramirez studied the equilibrium between a hydroxy-
phosphorane and its phosphate triester analogue in acetoni-
trile-d3 by NMR spectroscopy and verified the existence of
the phosphorane by trapping it with diazomethane (eq 1).4
To create a stable hydroxy-pentaoxy-phosphorane, we
sought a ligand that would satisfy two criteria: (1) possess
an internal nucleophile with a very high effective molarity
(EM)6 and (2) contain groups to minimize competition from
SN1 loss of phosphate.7 Kirby recently used a high effective
molarity structure to stabilize and thereby isolate a tetrahedral
intermediate analogous to that formed in amide hydrolysis.8
† The College of Pharmacy.
(1) Perreault, D. M.; Anslyn, E. V. Angew. Chem., Int. Ed. Engl. 1997,
36, 432-450 and references therein.
(2) Granoth, I.; Martin, J. C. J. Am. Chem. Soc. 1979, 101, 4618-4622.
(3) Kluger, R.; Covitz, F.; Dennis, E.; Williams, D.; Westheimer, F. H.
J. Am. Chem. Soc. 1969, 91, 6066-6072. Branch, G. E. K.; Calvin, M.
The Theory of Organic Chemistry; Prentice Hall: New York, 1941, pp 183-
270. Guthrie, P. J. Am. Chem. Soc. 1977, 99, 3391-4001. Granoth, I.;
Martin, J. C. J. Am. Chem. Soc. 1978, 101, 5229-5230.
(4) Ramirez, F.; Nowakowski, M.; Marecek, J. F. J. Am. Chem. Soc.
1977, 99, 4515-4517.
(5) Ramirez, F.; Bigler, A. J.; Smith, C. P. Tetrahedron 1968, 24, 5041.
(6) Kirby, A. J. AdV. Phys. Org. Chem. 1980, 17, 183.
(7) Richard, J. P. J. Am. Chem. Soc. 1991, 113, 4588.
10.1021/ol016941d CCC: $22.00 © 2002 American Chemical Society
Published on Web 12/21/2001

In a similar vein, ligand 1 potentially satisfies our design
criteria and was prepared by Dalby and Kirby to study the
hydrolysis of phosphate diesters.9 In their study an EM g
1010 mol L-1 for 1 was estimated from its rate of cyclization
when appended to a phosphate diester (eq 2).9 In addition,
Although the 31P NMR chemical shift is similar to that
reported by Ramirez for his phosphorane, this is not a
completely reliable criterion on which to assign a charac-
terization. Therefore, the solid-state structure was determined
by X-ray crystallography. As shown in Figure 1, the product
the two CF3 groups were included to minimize SN1 loss of
phosphate. Therefore, we set out to determine if 1 would
form a stable phosphorane at ambient temperature via
preparation of the group of geometrical isomers referred to
here as 2 (Scheme 1).
Scheme 1
Figure 1. Solid-state structure of 2B. Displacement ellipsoids are
scaled to the 50% probability level. Hydrogen atoms are drawn to
an arbitrary size. The minor occupancy F atoms were omitted for
clarity.
is phosphotriester 2B. Apparently the high effective molarity
of 1 is not sufficient to create a stable phosphorane. Hence,
we next set out to discover conditions that would shift the
equilibria given in Scheme 1 toward the phosphoranes.
Although Ramirez found a larger population of phospho-
ranes in more polar and basic solvents and at low temper-
ature, for our study of 2 in CD3CN, 31P NMR revealed little
change in the resonance at 223 or at 300 K. However, in
CDCl3 at 223 K the singlet at -23 ppm resolved into two
broad peaks (δ -22 and -24 ppm). Since Ramirez found a
34 ppm difference between phosphotriester and phosphorane,
we interpret this small separation to be indicative of a
dynamic conformational process within 2B alone. We then
attempted to trap the hypothetical phosphoranes.
Synthesis of a phosphorane based upon 1 was envisioned
to be achieved through several possible routes. If ligand 1
meets the two design criteria, we envisioned direct access
through reaction with P(O)Cl3. In THF with K2CO3 as the
base, a complex mixture of quartets was observed in the 19
F
When 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) was
added to 2B in CDCl3, the 31P NMR chemical shift moved
NMR of the crude product from reaction of 1 with P(O)Cl3,
and four singlets in the 31P NMR were observed ranging
from 4 to -23 ppm. When subjected to Kugelrohr distillation
at 135 °C and 1 Torr, a single product was isolated. In the
19F NMR spectrum of the distillate (CDCl3), four quartets
were observed with equal integrals (δ -74.08, -77.04 (J )
9.2 Hz); -74.75, -75.08 (J ) 8.8 Hz)), while in the 31P
NMR spectrum only a singlet at -23 ppm was found. These
spectra could either be interpreted as 2A, 2B, or 2D (Scheme
1), since 2C and 2E possess C2-symmetry and therefore
should only give two quartets in the 19F NMR spectrum.
downfield 10 ppm. Addition of Meirwein’s salt (CH3)3-
-
O+BF4 ) to this solution resulted in a compound whose 31
P
resonance was a quartet at -15 ppm with JPH ) 12 Hz. In
the 19F NMR four quartets in the starting material changed
to two new quartets (δ ) 74.19 and 76.70; J ) 9.6 Hz).
(8) Kirby, A. J.; Komarov, I. V.; Feeder, N. J. Am. Chem. Soc. 1998,
120, 7101-7102. Kirby, A. J.; Komarov, I. V.; Feeder, N. J. Chem. Soc.,
Perkin Trans. 2 2001, 522-529.
(9) Dalby, K. N.; Kirby, A. J.; Hollfelder, F. J. Chem. Soc., Perkin Trans.
2 1993, 1269.
202
Org. Lett., Vol. 4, No. 2, 2002

These data suggested the formation of phosphate triester 3
and an elimination product 4 (eq 3). An authentic sample of
while the second (δ ) -26 ppm) was assigned to 2A.
However, over several days 2A completely isomerized to
2B. This isomerization most likely proceeds via the phos-
phoranes shown in Scheme 1.
One is left wondering why the high effective molarity of
1 did not give a stable phosphorane, let alone why catechol
(eq 1) is even more effective than 1 in stabilizing a
phosphorane structure. The EM of g 1010 mol L-1 was for
the kinetics of a displacement reaction at a phosphorus center
(g14 kcal/mol in rate enhancement, eq 2).9 Although EM
values generally correlate between the kinetics and thermo-
dynamics of reactions, we find here a case where the
“kinetic” EM does not impart sufficient thermodynamic
stability to increase the concentration of a high energy
intermediate enough for its observation or isolation. This is
in contrast to catechol, which undoubtedly has a much lower
kinetic EM but imparts enough thermodynamic stability to
allow observation of the phosphorane at low temperature.
The displacement in eq 2 does not involve the intermediacy
of a phosphorane but instead a “phosphorane-like” transition
state and has entropy assisting the displacement. We postulate
that the reason ligand 1 does not act similarly in our study
compared to the Ramirez study is the fact that a six-
membered ring exists in 2A and 2B and a second is formed
in phosphoranes 2C-2E. In contrast, five-membered rings
are involved when catechol is the ligand. The angle strain
in phosphoester five-membered rings is partially relieved in
the phosphorane structures.4,11 We believe that this relief of
strain stabilizes the phosphorane depicted in eq 1, whereas
there is little difference in strain between the triesters and
phosphoranes depicted in Scheme 1. Therefore their relative
energies are not significantly perturbed from that of standard
triesters and phosphoranes.
phosphate triester 3 was prepared from methyl dichloro-
phosphate, and its NMR spectrum was identical to the
compound resulting from the addition of DBU and Meir-
wein’s salt.
Furthermore, on a preparative scale reaction, after allowing
1 and P(O)Cl3 in MeCN to stir with K2CO3 as the base at
-78 °C, followed by addition of Meirwein’s salt, a bright
yellow color was observed when the reaction temperature
rose above 0 °C, and phosphate triester 3 was isolated as
the product. A HRMS of the product mixture confirmed the
presence of quinone methide 4 as the elimination product
created in the formation of 3. Although the newly added
methyl group is on the phosphoryl oxygen, this does not
necessarily indicate that (CH3)3O+ underwent a reaction with
a phosphorane, since alkylation of the PdO bond of 2B could
lead to 3 and 4. Therefore, these results again provided no
direct evidence for the presence of a phosphorane.
Last, we analyzed a synthetic procedure that would initially
form 2A as a means to check whether 2B was a thermody-
namic sink. Reaction of PCl3 with 5 and then 1, followed
by oxidation with dimethyldioxirane,10 resulted in compound
6 (eq 4). The benzyl group could be easily removed in THF
at room temperature and 1 atm H2 with Pd/C in 10 min.
Acknowledgment. We gratefully acknowledge financial
support for this work from a NSF grant to E.V.A. and the
Welch Foundation.
Supporting Information Available: X-ray crystallo-
graphic data including experimental parameters and unit cell
packing diagram. This material is available free of charge
via the Internet at http://pubs.acs.org.
OL016941D
Analysis of 6 by 31P NMR spectroscopy after deprotection
revealed two resonances. The minor one corresponded to 2B,
(11) Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70. Kluger, R.; Taylor,
S. D. J. Am. Chem. Soc. 1990, 112, 6669. For a solvation viewpoint, see:
Dejeagere, A.; Karplus, M. J. Am. Chem. Soc. 1993, 115, 5316.
(10) Chappell, M. D.; Halcomb, R. L. Tetrahedron Lett. 1999, 40, 1.
Org. Lett., Vol. 4, No. 2, 2002
203