1-Oxo-2-oxa-1-phosphabicyclo[2.2.2]octane: A new mechanistic probe for the basic hydrolysis of phosphate esters

Article abstract of DOI:10.1021/ja9611147

Synthesis of the title compound 2 was accomplished in a multistep sequence starting from hypophosphorous acid. In strong base, the bicyclic phosphinate 2 hydrolyzes 2 orders of magnitude faster than the bicyclic phosphate OP(OCH₂)₃CCH₃ (1a) and the acceleration is entirely enthalpic. This rate enhancement is attributed to the greater ease with which 2 achieves the five-coordinate transition state. The molecular structure of 2, determined by X-ray means, compared with that of 1a reveals no evidence of strain within either bicyclic framework. The observed acceleration does not support the contribution of a stereoelectronic effect in the hydrolysis of six-membered ring phosphates.

Full text of DOI:10.1021/ja9611147

10168
J. Am. Chem. Soc. 1996, 118, 10168-10174
1-Oxo-2-oxa-1-phosphabicyclo[2.2.2]octane: A New
Mechanistic Probe for the Basic Hydrolysis of Phosphate Esters
†
Andrzej E. Wr o´ blewski and John G. Verkade*
Contribution from Gilman Hall, Department of Chemistry, Iowa State UniVersity,
Ames, Iowa 50011-3111
X
ReceiVed April 4, 1996
Abstract: Synthesis of the title compound 2 was accomplished in a multistep sequence starting from hypophosphorous
acid. In strong base, the bicyclic phosphinate 2 hydrolyzes 2 orders of magnitude faster than the bicyclic phosphate
OP(OCH2)3CCH3 (1a) and the acceleration is entirely enthalpic. This rate enhancement is attributed to the greater
ease with which 2 achieves the five-coordinate transition state. The molecular structure of 2, determined by X-ray
means, compared with that of 1a reveals no evidence of strain within either bicyclic framework. The observed
acceleration does not support the contribution of a stereoelectronic effect in the hydrolysis of six-membered ring
phosphates.
3
Introduction
hydrolyzes 0.81 × 10 times faster than triethyl thionophosphate
under the same conditions. The rate enhancements observed
for the bicyclic phosphates with respect to their acyclic analogs
were attributed at least in part to stereoelectronic effects, since
in the bicyclic phosphate systems there are two lone electron
pairs forced to be app to the breaking P-O bond in transition
state A, while this constraint is not operative in the acyclic
The mechanism of phosphate ester hydrolysis has recently
attracted much attention because of the crucial biological
importance of this reaction.1
-4
On the basis of ab initio
2
3
molecular orbital calculations and laboratory experiments,
Gorenstein and co-workers advocated the importance of the role
of kinetic stereoelectronic effects in the reactions of organo-
phosphorus compounds. Rate enhancements observed for cyclic
or bicyclic phosphates compared with their acyclic analogs were
suggested to be due, at least in part, to stereoelectronic effects
that facilitate cleavage of the apical P-O or P-N bonds in
trigonal bipyramidal intermediates. These stereoelectronic
effects presumably arise from antiperiplanar (app) interactions
of the breaking apical bond with electron lone pairs on
equatorially positioned oxygen or nitrogen atoms.
4
analogs.
These studies were frustrated to some extent, however, by
the conformational flexibility of the monocyclic and decalin-
type bicyclic phosphate esters employed. Thus, a conforma-
tionally biased system which also possesses stereoelectronically
favorable electron pairs at the oxygens was introduced by
incorporating the phosphate ester residue into the bicyclo[2.2.2]-
Subsequently we became interested in attempting to devise
further tests of the role of antiperiplanar electron pairs in rate
enhancements of the hydrolysis of bicyclic six-membered ring
phosphates. To this end 1-oxo-2-oxa-1-phosphabicyclo[2.2.2]-
octane (2) was synthesized. This compound has only one
hydrolyzable P-O bond, and in contrast to 1, two ring O atoms
are replaced by two methylene groups, thus providing CH2
groups in the trigonal bipyramidal transition state B instead of
the two O atoms that are crucial to the manifestation of the
stereoelectronic effect in A. A comparison of the rates of the
base-catalyzed hydrolysis of 2 and 1a could be expected to allow
a direct assessment of the rate enhancement in bicyclic 1a by
a stereoelectronic effect. According to the antiperiplanar lone
pair hypothesis (ALPH), 1a is expected to hydrolyze faster than
2. However, we found that the bicyclic phosphinate 2 hydro-
lyzed more than 2 orders of magnitude faster than 1a.
Furthermore, by comparing rates of the base-catalyzed hydroly-
sis of 2 and its acyclic analog, EtOP(O)Et2 (4), we could expect
to estimate the rate enhancement in a bicyclo[2.2.2]octane
system in the absence of a stereoelectronic effect. If the rate
enhancement of 2 over 4 was found to be smaller than that for
4
octane framework of compound 1a. It was observed at pH 14
3
that compound 1a hydrolyzes 5.0 × 10 times faster than triethyl
phosphate (3), while the thiono analogue of 1, namely 1b,
†
Visiting Professor from the Institute of Organic Chemistry, Technical
University, L o´ dz, Poland.
X
Abstract published in AdVance ACS Abstracts, October 1, 1996.
(1) (a) Thatcher, G. R. J.; Kluger, R. AdV. Phys. Org. Chem. 1989, 25,
9
9. (b) Tole, P.; Lim, C. J. Am. Chem. Soc. 1994, 116, 3922. (c) Uchimaru,
T.; Tsuzuki, S.; Storer, J. W.; Tanabe, K.; Taira, K. J. Org. Chem. 1994,
9, 1835 and references cited in these references.
2) Taira, K.; Gorenstein, D. G. J. Am. Chem. Soc. 1984, 106, 7825 and
references cited therein.
3) Farschtschi, N.; Gorenstein, D. G.; Fanni, T.; Taira, K. Phosphorus,
Sulfur, Silicon 1990, 47, 93 and references cited therein.
4) Fanni, T.; Taira, K.; Gorenstein, D. G.; Vaidyanathaswamy, R.;
Verkade, J. G. J. Am. Chem. Soc. 1986, 108, 6311.
1
a over 3, the stereoelectronic effect in 1a would be substanti-
5
(
ated. Somewhat surprisingly, we find that this factor for 2
relative to 4 is more than an order of magnitude greater than
for 1a compared with that of 3. In an effort to identify the
(
q
source of this acceleration, the activation parameters ∆H and
(
q
∆S were estimated from Eyring theory. However, these
S0002-7863(96)01114-6 CCC: $12.00 © 1996 American Chemical Society

Basic Hydrolysis of Phosphate Esters
J. Am. Chem. Soc., Vol. 118, No. 42, 1996 10169
4
h, 0.02 Torr) gave 6 (21.8 g, 87.5%) as a colorless oil. 1H NMR: δ
experiments required the replacement of 1,4-dioxane by 1,2-
dimethoxyethane as an organic solvent. For this and additional
reasons discussed later, we are unable to make direct compari-
sons of our rate constants for 1a and 3 with those reported
2
3
3
2
1
.04 (dtd, CH
.8 Hz), 2.5-2.7 (m, CH
2
P, J(PH) ) 15.2 Hz, J(HH) ) 7.6 Hz, J(HPCH) )
CO), 3.67 (s, CH C), 3.75 (CH OP,
2
O
3 2
3
3
1
3
J(HCOP) ) 11.8 Hz), 7.14 (dt, HP, J(HP) ) 547.8 Hz, J(HPCH) )
1
1
.8 Hz). 13C NMR: δ 23.25 (d, CP, J(CP) ) 95.2 Hz), 25.7 (s, CH
2
-
4
earlier.
2
CO), 51.54 (s, CH
O
3 2
C), 52.32 (d, CH
3
OP, J(COP) ) 5.4 Hz), 171.72
3
31
(d, COO, J(PCCC) ) 12.2 Hz). P NMR: δ 39.76.
Experimental Section
A mixture of crude 6 (21.8 g, 0.131 mol) and methyl acrylate (13.0
Melting points were taken on a Thomas-Hoover capillary melting
mL, 0.144 mol) was added dropwise to a solution of sodium methoxide
(made from 0.65 g, 0.028 mol, of sodium in methanol, 8.2 mL) at 3-5
°C. When the addition was complete, the mixture was allowed to warm
to room temperature and was then diluted with chloroform (100 mL).
The solution was neutralized with acetic acid (1.8 mL, 0.031 mol) and
1
13
point apparatus and are uncorrected. H and C NMR spectra were
measured on a Nicolet NT-300 NMR spectrometer in chloroform-d as
solvent, and chemical shifts are reported in parts per million downfield
1
from tetramethylsilane using chloroform ( H, 7.26 ppm) and chloro-
13
31
form-d ( C, 77.10 ppm) resonances as secondary standards. P NMR
spectra were recorded on a Bruker WM-300 spectrometer for the
chloroform-d solutions, and chemical shifts are externally referenced
then washed with ice-cold water and aqueous NaHCO
washings were extracted with chloroform (3 × 20 mL). The organic
extracts were dried over MgSO , and the solvent was evaporated in
Vacuo (24 h, 0.04 Torr), leaving crude 7 as a yellow oil (31.47 g, 95%).
3
. The aqueous
4
3 4
to 85% H PO with positive values downfield from the standard.
Solvents were reagent grade, predried over molecular sieves, and, when
1
2
3
H NMR: δ 2.05 (dt, CH
2.54-2.65 (m, CH CO), 3.68 (s, CH
) 10.6 Hz). C NMR: δ 22.78 (d, CP, J(CP) ) 93.1 Hz), 26.36 (s,
2
P, J(HP) ) 13.4 Hz, J(HH) ) 7.9 Hz),
3
necessary, distilled from sodium-benzophenone ketyl prior to use.
2
3 2 3
O C), 3.68 (d, CH OP, J(HCOP)
3
1
13
1
P NMR spectra for the kinetic experiments were obtained with a
2
Bruker WM-200 spectrometer at 81.0 MHz, employing the following
parameters: sweep width, 8064 Hz; memory size, 16K; pulse width,
5 µs; relaxation delay, 0.5 s. The number of transients was selected
according to t1/2 values for the reactions and was between 50 and 200
for 1a, depending on temperature, and was 500 for 3 and 4 at all
temperatures. Integrations of these spectra were performed using the
NMR1 program. The processing of the FID included exponential
multiplication with a line-broadening of 1.0 Hz and zero-filling to 64K.
Each signal was fitted to a Lorentzian curve three times using a curve-
fitting routine. Mean integrations were then calculated.
CH
172.28 (d, CO
preparation was contaminated with ca. 6% of 8 (δ( P) 48.65) and ca.
2
CO), 51.05 (d, CH
3
OP, J(CP) ) 5.2 Hz), 51.87 (s, CH
3
O
2
C),
3
31
2
, J(CCCP) ) 14.8 Hz). P NMR: δ 56.11. The
3
1
1
31
4% of an unidentified impurity (δ( P) 45.84).
A solution of sodium tert-amyl oxide (made from sodium hydride,
3.6 g, 0.150 mol, and tert-amyl alcohol, 16.5 mL, 0.150 mol, in benzene,
100 mL) was added to a solution of 7 (31.5 g, 0.125 mol) in benzene
(180 mL) at room temperature. The reaction mixture was warmed,
and a distillate of bp 54-65 °C followed by another up to 80 °C was
slowly collected at atmospheric pressure. The cooled residue was
filtered and washed with benzene, and the solid was added portionwise
to a vigorously stirred aqueous solution of concentrated HCl (16.6 mL)
in 22 mL of water at 5-8 °C. The organic phase was extracted with
Relaxation times (T
1
) were measured with a Bruker WM-200
spectrometer by the inversion-recovery method for at least seven
different τ values. Further calculations were performed using NMR1
3
1
18
-
software. P NMR spectra for the OH -catalyzed hydrolysis of 2
chloroform (5 × 50 mL), dried over MgSO
4
, and concentrated overnight
were obtained with a Varian VXR 300 spectrometer at 121.4 MHz
3824.1 Hz sweep width, 27 008 data points).
in Vacuo, affording crude 8 (24.0 g, 87%) as a brown oil. A sample
of this material was purified on a silica gel column with chloroform-
methanol (40:1, v/v) to give a colorless oil which solidified after a
(
Ethyl diethylphosphinate (4) was prepared from tetraethyldiphosphine
5
,6
1
2
disulfide according to literature procedures, purified by chromatog-
few weeks at room temperature. H NMR: δ 1.98 (dt, CH
2
P, J(PH)
3
raphy on silica gel (chloroform), and distilled twice in Vacuo. Bp:
) 15.6 Hz, J(HH) ) 7.1 Hz), 2.57-2.79 (m, CH
2
P, CCH
C), 12.6 (br s, HO). C
2
C), 3.72 (d,
6
1
3
13
8
6-88 °C, 11 Torr (lit. bp 87-9 °C, 12 Torr. H NMR: δ 1.07 (dt,
CH
3
OP, J(HCOP) ) 10.9 Hz), 3.76 (s, CH
O
3 2
3
3
1
1
CH
3
CP, J(HP) ) 17.6 Hz, J(HH) ) 7.6 Hz), 1.24 (dt, CH
3
COP,
NMR: 21.40 (d, CPC, J(CP) ) 88.9 Hz), 21.65 (d, CPC, J(CP) )
90.4 Hz), 28.09 (d, CH
J(COP) ) 4.2 Hz), 51.82 (s, CH C), 92.59 (s, C-3 based on the
usual lack of two-bond coupling to P), 171.15 (d, C-4 or CO,
3
4
2
2
J(HH) ) 7.1 Hz, J(HP) ) 0.5 Hz), 1.64 (dq, CH
2
P, J(HP) ) 13.9
2
CP, J(CCP) ) 5.9 Hz), 50.63 (d, CH
3
OP,
3
3
3
2
Hz, J(HH) ) 7.6 Hz), 3.98 (dq, CH
2
OP, J(HP) ) 7.1 Hz, J(HH) )
3 2
O
3
1
31
7
.1 Hz). P NMR: δ 59.97.
The bicyclic phosphate 1a was prepared as described previously,7
3
3
J(CCCP) ) 14.4 Hz), 172.04 (d, C-4 or CO, J(CCCP) ) 12.4 Hz).
31
and triethyl phosphate (3) purchased from Aldrich was used without
further purification.
P NMR: δ 48.91. Compound 8 was found to be 100% enolized as
judged from the 13C and H NMR spectra.
1
Although the sequential transformation of H
3
PO
2
to CH
CH
3
OP(O)H
CO Me)
2
A mixture of crude 8 (24.0 g, 0.109 mol) and 0.01 M HCl (25 mL)
was kept at 98 °C for 3 d. After cooling, the solution was saturated
with sodium chloride and extracted with chloroform (10 × 20 mL).
(
(
5) to CH OPH(O)CH CH CO Me (6) to MeOP(O)(CH
3
2
2
2
2
2
2
2
2
2
2
2 2
7) to MeO(O)PCH CH C(OH)dC(CO Me)CH (8) to MeO(O)P(CH -
4
The organic extract was dried over MgSO , concentrated, and distilled
8
CH )
2 2
CdO (9) was reported earlier, the following procedure to obtain
was found to be superior.
-Methoxy-1-oxophosphorinan-4-one (9). A 50% solution of
to give 9 (7.19 g, 40.7%) as a colorless oil which immediately
9
8
crystallized. Bp: 130-135 °C, 0.3 Torr; (lit. bp, 135 °C, 1 Torr).
1
8
1
Mp: 56-57.5 °C (lit. mp 38-40 °C, 51-52 °C hemihydrate).
H
hypophosphorous acid (Aldrich) was dried in Vacuo overnight, leaving
2
3
NMR: δ 2.19 (dt, CH
2
P, J(HCP) ) 16.4 Hz, J(HH) ) 6.7 Hz), 2.67-
9
a solid (9.89 g, 0.150 mol). This crude acid was treated with trimethyl
3
13
2
2
.77 (m, CH
3.68 (d, CH
2 3
CO), 3.82 (d, CH OP, J(HP) ) 10.8 Hz). C NMR: δ
2
orthoformate (18.0 mL, 0.165 mol) at room temperature, and the
solution was stirred for 2 h. The solution was added dropwise to a
mixture of methyl acrylate (13.5 mL, 0.165 mol) and ethyldiisopro-
pylamine (2.6 mL, 0.015 mol) at 5 °C. After the reaction mixture had
been allowed to stand at room temperature for 3 d, chloroform (50
mL) was added and the solution was washed with saturated cold
1
2
P, J(CP) ) 89.2 Hz), 36.75 (d, CH CP, J(CP) ) 4.5
3
2
2
Hz), 50.97 (d, CH
9
3
OP, J(CP) ) 7.6 Hz), 206.81 (d, CdO, J(CP) )
.3 Hz). P NMR: δ 48.29.
-Methoxy-4-methylene-1-oxo-phosphorinane (10). To a suspen-
31
1
10
sion of freshly sublimed potassium tert-butoxide (2.64 g, 23.3 mmol)
in dry ether (40 mL) at room temperature was added portionwise
aqueous NaHCO
3
. The water phase was extracted with chloroform (5
11
methyltriphenylphosphonium iodide (9.41 g, 23.3 mmol). The yellow
×
20 mL). The extract and washings were combined and dried over
slurry was refluxed for 15 min, and then most of the ether was distilled
MgSO
4
. Removal of the solvent and volatile impurities in Vacuo (10
12
off. The ketone 9 (3.3 g, 21 mmol) was then added in portions
followed by benzene (5 mL). The reaction mixture was stirred at 40
°C for 1 h, cooled to room temperature, and partitioned between
chloroform and water. The water phase was extracted with chloroform
(
5) (a) Parshall, G. Organic Syntheses; Wiley: New York, 1973; Collect.
Vol. V, p 1016. (b) Pollard, K. A.; Harwood, H. J. J. Org. Chem. 1962, 27,
444.
6) Cook, R. D.; Diebert, C. E.; Schwarz, W.; Turley, P. C.; Haake, P.
J. Am. Chem. Soc. 1973, 95, 8008.
4
(
(10) Pearson, D. E.; Buehler, C. A. Chem. ReV. 1974, 74, 45.
(11) Lightner, D. A.; Crist, B. V.; Kalyanam, N.; May, L. M.; Jackman,
D. E. J. Org. Chem. 1985, 50, 3867.
(7) Verkade, J. G.; Reynolds, L. T. J. Org. Chem. 1960, 25, 663.
(8) Gallagher, M. J.; Sussman, J. Phosphorus 1975, 5, 91.
(9) Fitch, S. J. J. Am. Chem. Soc. 1964, 86, 61.
(12) Fitjer, L.; Quabeck, U. Synth. Commun. 1985, 15, 855.

10170 J. Am. Chem. Soc., Vol. 118, No. 42, 1996
Wr o´ blewski and Verkade
a
(
4 × 20 mL), and the organic extracts were dried over MgSO
4
. Most
Table 1. b Observed and Calculated Pseudo-First-Order Rate
Constants (s ) for the Hydrolysis of 1a, 2, 3, and 4c
-1
of the triphenylphosphine oxide formed was removed on a silica gel
column with chloroform-methanol (100:1, v/v), and the residue was
temp,
3
3
6
6
distilled to give 10 (2.22 g, 66.6%) as a yellow mobile oil. Bp: 70-
K
10 k (1a)
10 k (2)
10 k (3)
10 k (4)
1
7
CH
3 °C, 0.1 Torr. H NMR: δ 1.7-2.0 (m, CH
2
P), 2.3-2.6 (m, CH
2
-
C
254
7.16 ( 0.18
.31 ( 0.09
13.44 ( 0.26
13.04 ( 0.42
22.17 ( 0.48
23.20 ( 0.32
32.39 ( 0.55
3
13
2
P), 3.70 (d, CH
NMR: δ 27.02 (d, CH
3
OP, J(HP) ) 10.7 Hz), 4.71 (s, H
2
CdC).
7
1
2
P, J(CP) ) 85.9 Hz), 39.22 (d, CH CH
2
2
P,
259
2
2
J(CCP) ) 5.9 Hz), 50.26 (d, CH OP, J(COP) ) 6.0 Hz), 112.52 (s,
31
3
3
CH
0.44.
-Methoxy- and 1-(Hydroxymethyl)-1-oxo-phosphorinane (11
2
dC), 144.80 (d, CdCH
2
, J(CCCP) ) 10.4 Hz). P NMR: δ
264
5
2
2
2
69
1
3
2.92 ( 0.28
74 0.1044 ( 0.0011 53.27 ( 0.98
.0993 ( 0.0007 52.8 ( 1.1
79 0.2082 ( 0.0016
.2112 ( 0.0017
and 12, Respectively). To a solution of 10 (2.20 g, 13.9 mmol) in
THF (13 mL) was added a solution of 1 M borane-THF complex in
THF (16.6 mL, 16.6 mmol, Aldrich) at 6-10 °C followed by stirring
at room temperature for 2 h. Water (2 mL) was then added at 2 °C
followed by a solution of 180 mg of solid NaOH in 4 mL of water.
Finally, hydrogen peroxide (1.67 mL, 30% solution) was added
dropwise, and the mixture was stirred at 10 °C for 2 h. The solvent
was evaporated at room temperature, and the residue was partitioned
between chloroform and water. The organic phase was dried over
0
0
284 0.3271 ( 0.0018
0.3637 ( 0.0030
289 0.6105 ( 0.0089
0
.6450 ( 0.0059
294 1.0570 ( 0.0096
1
1
.063 ( 0.018
.100 ( 0.012
MgSO
column with chloroform-methanol (100:1, v/v) to give a 1:1 mixture
4
, concentrated in vacuo, and chromatographed on a silica gel
298
303
308
11.272 ( 0.074
18.418 ( 0.088
27.26 ( 0.28
42.47 ( 0.20
60.12 ( 0.63
1
2
of isomers of 11 (0.581 g, 23.5%). H NMR: δ 1.2-2.1 (m, CH -
3
CH
2
), 3.1 and 3.25 (2 br s, OH), 3.36 (d, CH
2
O, J(HH) ) 5.9 Hz),
3
3
.42 (d, CH
2
O, J(HH) ) 6.1 Hz), 3.61 and 3.62 (2 d, CH OP,
3
313
3
13
1
J(HCOP) ) 10.7 Hz). C NMR: δ 24.85 (d, CH
2
P, J(CP) ) 86.3
2
1
2
318
323
328
7.21 ( 0.13
10.26 ( 0.16
17.01 ( 0.21
24.68 ( 0.56
30.36 ( 0.35
42.67 ( 0.35
Hz), 24.43 (d, CH
2
P, J(CP) ) 86.7 Hz), 25.23 (d, CH
CP, J(CCP) )
2
3
3.7 Hz), 26.33 (d, CH
2
CP, J(CCP) ) 4.7 Hz), 39.19 (d, CH, J(CCCP)
3
)
7.1 Hz), 39.93 (d, CH, J(CCCP) ) 5.8 Hz), 50.00 (d, CH
3
OP,
OP, J(COP) ) 6.3 Hz), 65.02 and
O). P NMR: δ 52.21 and 53.69. The water phase
was acidified with 4 mL of HCl (1:1 concentrated HCl/H O v/v) and
3
3
3
33
38
43
2
2
J(COP) ) 6.8 Hz), 50.45 (d, CH
3
3
1
6
6.02 (2 s, CH
2
2
refluxed for 12 h. Water was evaporated, and then the residue was
co-evaporated with water twice (2 × 30 mL) followed by drying in
Vacuo over NaOH pellets. The oily residue was extracted with
methanol at room temperature, and the extracts were filtered, evapo-
rated, and left in Vacuo for 2 d to give the crude acid 12 (1.3 g, 57%)
304a 3.08 ( 0.13
628 ( 55
2.39 ( 0.26 19.55 ( 0.29
a
Based on a ln k Vs 1/T relationship. b The error ranges are the
standard deviations. c 0.6 M NaOH in 44% aqueous 1,2-dimethoxy-
ethane.
1
3
1
as a yellow oil. C NMR (CD
3
OD): δ 26.90 (d, CH
2
P, J(CP) )
2
1
00.0 Hz), 27.19 (d, CH CP, J(CCP) ) 3.9 Hz), 40.89 (d, HC,
2
Basic Hydrolysis of 2. A 25 mL flask containing a magnetically
stirred mixture of DME (1.50 mL) and 0.62 M NaOH in D O (1.20
2
mL) was cooled to the required temperature in a well-insulated ethanol
bath using an RK20 thermostat (Brinkmann) equipped with an RKS
control unit. A solution of 2 in D O (47 µL, 68 µmol) was then injected
followed by quenching at various time intervals by injection of glacial
acetic acid (47 µL, 1.1 equiv). At selected temperatures this procedure
was repeated seven to eight times to monitor the progress of hydrolysis
for at least 2 half-lives. After quenching, the slightly acidified solutions
3
31
J(CCCP) ) 5.9 Hz), 68.06 (s, CH
OH): δ 52.23.
2
OH). P NMR (CD
3
3
OD-CH -
1-Oxo-2-oxa-1-phosphabicyclo[2.2.2]octane (2). A mixture of the
crude acid 12 (1.24 g, 7.56 mmol), dicyclohexylcarbodiimide (DCC,
2
1.87 g, 9.06 mmol), 4-(N,N-dimethylamino)pyridine (DMAP, 0.184 mg,
1
.51 mmol), triethylamine (10 mL), and THF (50 mL) was refluxed
for 9 h. Volatiles were then evaporated under vacuum, and the residue
was suspended in benzene and filtered. The oil obtained upon
evaporation of the solvent under vacuum was chromatographed on silica
gel with chloroform to give crude 2 (0.71 g, 64%). Crystallization
31
were transferred to 10 mm NMR tubes, and P NMR spectra were
recorded employing the following parameters: sweep width, 8064 MHz;
memory size, 16K; pulse width, 90° (23.5 µs); relaxation delay, 45 s;
number of transients, 40.
from benzene-hexane afforded 336 mg of pure 2. Mp: 220-223 °C.
1
H NMR: δ 1.73-1.77, 1.9-2.3 (2 m, CH
2
CH
2
CH), 4.51 (dd, CH
2
O,
P, J(CP)
CP, J(PC) ) 5.7 Hz), 26.43 (d, CCCP,
3
3
13
1
J(HP) ) 5.4 Hz, J(HH) ) 1.4 Hz). C NMR: δ 21.83 (CH
2
Integration Correction Factors. It appeared that full relaxation
2
)
81.4 Hz), 25.94 (d, CH
J(CCCP) ) 47.6 Hz), 76.90 (d, CH
NMR: δ 46.05. IR (benzene): ν(PdO) 1271 and 1229 cm . IR
2
of 31P nuclei would be achieved in about 1 min, thus precluding direct
3
2
31
2
OP, J(COP) ) 4.4 Hz).
P
quantitative measurements especially for the reactions having half-lives
of only several minutes. Thus, we decided to acquire P NMR data
under nonequilibrium conditions and later correct for the observed
integrals by multiplication of the ester-to-Na salt ratios by the
appropriate correction factor (Table 2). The validity of this approach
-
1
31
-
1
(
1
1
tetrachloroethylene): ν(PdO) 1262 and 1229 cm
.
MS: m/e
P, 146.04968. MS: m/e 146.0 (100),
31.0 (14), 116.0 (37), 88.0 (23), 68.1 (29.9), 54.0 (92.1). Anal. Calcd
P (found): C, 49.31 (49.34); H, 7.59 (7.46); P, 21.20
6 11 2
46.04956, calcd for C H O
for C
21.24).
Basic Hydrolysis of 3 and 4. In a 10 mm NMR tube, 1,2-
dimethoxyethane (DME) (1.50 mL) and 0.62 M NaOH in D O (1.20
mL) were mixed. The temperature of the solution was sustained for
6 11 2
H O
was confirmed by comparing molar ratios measured gravimetrically
(
31
(for 1a) with those obtained from P NMR integrals. Surprisingly,
the estimated error for three different concentrations was less than 1%.
By contrast, 31P NMR integrals of 2 and the sodium salt of 12 could
be obtained directly from spectra of the quenched mixtures by using
sufficiently long delays.
The correction factors were obtained according to the following
procedure. The solution of the sodium salt (ca. 68 µmol) in DME
2
3
6
0 min in the probe of the spectrometer. Then, compound 3 (11.5 µL,
7.7 µmol) or 4 (10.0 µL, 66.6 µmol) was injected, and the reaction
was monitored at different temperatures (Table 1).
Basic Hydrolysis of 1a. In a 10 mm NMR tube, a solution of 1a
2
(1.50 µL) and D O (1.20 mL) obtained after complete hydrolysis of
(
66.95 µmol) in DME (1.50 mL) was mixed with D
temperature of the solution was sustained for about 1 h in the probe of
the spectrometer. Then, 7.32 M NaOH in D O (100 µL) was injected,
2
O (1.10 mL). The
1a, 3, or 4 with excess base at room temperature was slightly acidified
by adding 1.1 equiv of glacial acetic acid to react with the excess base.
Glacial acetic acid was not added to the sample of hydrolyzed 3, because
its hydrolysis rate at 24 °C is negligibly slow. Then, ca. 34 µmol of
the ester was added in each case, and 31P NMR spectra were collected
2
the sample was shaken well, and the reaction was monitored at five
temperatures (Table 1).

Basic Hydrolysis of Phosphate Esters
J. Am. Chem. Soc., Vol. 118, No. 42, 1996 10171
(s), 31P NMR Chemical Shifts (ppm), and Correction Factors for the Mixtures of Esters and Their Respective
1
Table 2. Relaxation Times T
Na Salts
3
1
31
correction factora
ester
T
1
(δ( P))
Na salt
1
T (δ( P))
1
2
3
4
a
9.4 ( 0.5 (-5.21)
9.0 ( 0.4 (54.52)
13.3 ( 0.1 (-0.10)
10.3 ( 0.06 (66.13)
HOCH
12 (Na)
(EtO) PO
2
CMe(CH
2
O)
2
PO
2
-Na+
5.4 ( 0.1 (-2.86)
1.9 ( 0.2 (38.51)
11.8 ( 0.3 (+0.97)
4.3 ( 0.3 (47.80)
1.585 ( 0.025
not used
1.120 ( 0.013
1.635 ( 0.033
-
Na+
2
2
-
Na+
2 2
Et PO
a
The error is the standard deviation calculated as [[∑ø - (∑ø) /n]|(n - 1)]1/2.
2
2
a
a
b
Scheme 1a
Table 3. Calculated Enthalpies, Entropies, and Free Energies of
Activation for the Hydrolysis of 1a, 2, 3, and 4 in Strong Base
q
∆S , kcal/mol (×10-3)
q
q
∆
H , kcal/mol
∆G 300, kcal/mol
1
2
3
a
17.99 ( 0.34
12.93 ( 0.34
-10.9 ( 1.2
-16.9 ( 1.3
-35.2 ( 2.4
21.3
18.0
25.5
25.1
24
24.2
24.9
23.8
14.92 ( 0.79
c
c
1
1
4.8
4.1
-34.4
d
d
-34
4
15.14 ( 0.35
-30.3 ( 1.1
e
e
1
1
4.3 ( 0.5
-35.2
f
-26.8f
5.8
a
The error ranges are the standard deviations. At 300 K. c Refer-
b
d
ence 26 (in H
in ref 26). Reference 26 (in 50% EtOH/H
DMSO).
2
O). Reference 29b (recalculated values from the data
e
f
2
O). Reference 27 (in 83.3%
a
a
Table 4. Acceleration Factors and Differences in Enthalpies of
q
q
Activation (∆∆H ) and Entropies of Activation (∆∆S ) for Pairs of
a
Conditions: (a) HC(OMe)
Me, MeONa; (d) t-AmONa, benzene; (e) 0.01 M HCl;
; (g) BH ‚THF, then NaOH/H ; (h) DCC/DMAP, THF.
3 2 2 2
; (b) H CdCHCO Me, i-Pr NEt; (c)
Esters
H
2
CdCHCO
(f) Ph PdCH
2
q
∆
∆S ,
3
2
3
2 2
O
pairs of esters
kcal/mol
q
-3
“
faster” “slower”
acceleration
∆∆H , kcal/mol
(×10
)
9.2 ppm, and the following 13C NMR signals: δ 20.34 (d, CH
5
2
P,
CP, J(CP) ) 5.9 Hz), 25.87 (d,
1
2
2
4
2
1
1a
3
4
204 ( 20
8.18 ( 0.90
-5.06 ( 0.68
+0.22 ( 1.1
-6.0 ( 2.5
+4.9 ( 3.5
J(CP) ) 81.7 Hz), 24.87 (d, CH
2
3
2
CCCP, J(CCCP) ) 49.0 Hz), 78.26 (d, CH
2
OP, J(CP) ) 6.0 Hz).
O (0.30 mL, 1.1
3
(32.1 ( 2.9) × 10
-2.21 ( 0.69 +13.4 ( 2.4
After injection of a solution of 0.6025 M NaOH in D
2
3
a
3
(1.29 ( 0.15) × 10
+3.1 ( 1.1
+24.3 ( 3.6
equiv), the following spectral data for the Na salt of 12 were observed.
3
1
13
2
a
P NMR: δ 40.62. C NMR: δ 25.97 (d, CH
CP, J(CP) ) 4.8
The error ranges are the standard deviations.
2
1
3
Hz), 27.07 (d, CH
)
spectra.
2
P, J(CP) ) 86.5 Hz), 39.20 (d, CCCP, J(CCCP)
2
5.4 Hz), 65.65 (s, CH OH). No other signals were detected in these
first with the spectral parameters used in the kinetic studies (pulse width,
5 µs; relaxation delay, 0.5 s) for at least three different numbers of
transients (between 50 and 200 for 1a, between 100 and 600 for 3 and
). Then, P NMR spectra were observed for the same sample using
a 90° pulse and relaxation delays (70 s for 1a and 3 and 52 s for 4) for
1
Crystal Structure Analysis of 2. Crystals of 2 were grown from
a dichloromethane solution layered with hexane. Some difficulty was
encountered in finding a single crystal which yielded accurate cell
constants, apparently because of a large mosaic spread. All nine non-
hydrogen atoms were located by direct methods.14 All of the expected
hydrogen atoms were located in subsequent difference Fourier maps
and were refined with isotropic thermal parameters. One of the
hydrogen atoms would not refine with a reasonable thermal parameter
in full-matrix least squares cycles. A check of F(obs) versus F(calc)
revealed two reflections (indices 750 and 910) for which F(obs) -
F(calc) was greater than 10σ(F). Upon exclusion of these reflections
from the final least squares cycles, all of the hydrogen atoms refined
satisfactorily. Refinement was carried out using the SHELX-76
programs. All calculations were performed on a Digital Equipment
3
1
4
at least three different numbers of transients (between 10 and 40 for
a, between 30 and 300 for 3 and 4). The sequence was completed by
1
once again recording spectra with parameters used in the kinetic
measurements (as in the first step). This sequence allowed us to
calculate at least six different values of a given correction factor by
dividing the ester-to-Na salt molar ratios obtained from spectra under
full relaxation by those ratios found from spectra in which short delays
had to be applied. This three-step sequence was repeated for ca. 1:1
and 2:1 molar ratios of ester to Na salt, thus bringing the number of
correction factor values to at least 18. Mean values and standard
deviations were then calculated (Table 2).
1
5
Corp. MicroVAX II computer using the CAD4-SDP package.
Data Analysis. Pseudo-first-order rate constants at the same base
concentration and the Arrhenius parameters were calculated by least-
Results and Discussion
1
3
squares analysis.
Enthalpies and entropies of activation were
calculated at 304 K in the normal manner using standard Eyring theory.
Pertinent data with their standard deviations are collected in Tables 3
and 4.
Synthesis of 2. The synthesis of 2 was accomplished in an
eight-step sequence (Scheme 1) starting from H3PO2.9 Sequen-
tial Michael-type additions of the P-H subunits in 5 and 6 to
methyl acrylate followed by a Dieckmann cyclization of 7 gave
1
8
2
Hydrolysis of 2 with Na OH in D O. To a solution of 7.32 M
1
8
NaOH in D
a 5 mm NMR tube was added 2 (28.5 mg, 0.195 mmol). After 2 h at
room temperature, the clear solution was diluted with D O (0.6 mL)
2
O (47.2 mg) containing H
2
O (95%, Aldrich, 51.8 mg) in
8
16
enol 8, which after decarboxylation in dilute acid afforded
2
(14) (a) SHELXS-86, G. M. Sheldrick, Institut f u¨ r Anorganische Chemie
der Universit a¨ t, Gottingen, FRG. (b) Sheldrick, G. M. In Computing in
Crystallography; Schenk, H., Olthof-Hazekamp, R., Van Koningsveld, H.,
Bassi, G. C., Eds.; Delft University: Delft, 1978.
3
1
16
and its P NMR spectrum was recorded. Two signals at 40.620 ( -
OP O) and 40.582 ( OP O) ppm in a 49:51 ratio were observed.
1
6
16
18
Product Analysis of Hydrolysis of 2. A solution of 2 (24.0 mg,
(15) Enraf-Nonius: Delft, Holland. Neutral atom scattering factors and
3
1
0
.164 mmol) in D
2
O (0.5 mL) showed a single P NMR resonance at
anomalous scattering corrections were taken from the International Tables
for X-ray Crystallography; The Kynoch Press: Birmingham, England, 1974;
Vol. IV.
(16) Gallagher, M. J.; Honegger, H.; Sussman, J. Austr. J. Chem. 1982,
35, 363.
(
13) Margerison, D. In ComprehensiVe Chemical Kinetics; Bamford, C.
H., Tipper, C. F. H., Eds.; Elsevier Publishing Co.: New York, 1969; Vol.
, pp 360-406.
1

1
2
0172 J. Am. Chem. Soc., Vol. 118, No. 42, 1996
Wr o´ blewski and Verkade
-methoxy-2-oxo-4-phosphorinanone (9) in 29% overall yield.
unexpected that basic hydrolysis of 2 would occur Via nucleo-
2
3
We found this procedure superior to the one which involved
philic attack at the C-O-P carbon atom, this possibility was
the purification of 6, 7, and 8 after each step.8 The P and
NMR spectra of crude 6 revealed the presence of 7, and crude
contained up to 10% 8. Although basic catalysts selected by
31
13
C
tested by carrying out the hydrolysis of 2 with a Na OH/D2O
18
3
1
solution and recording the P NMR spectrum of the reaction
1
6
18
7
mixture. As expected on the basis of the O/ O isotope
8
24
31
Gallagher for the stepwise addition to methyl acrylate worked
quite well in the presence of ethyldiisopropylamine, some 7 was
formed in the first step as shown by 31P NMR spectroscopy,
and sodium methoxide was a strong enough base to promote
the cyclization of 7 to 8 even at room temperature, as was also
shown by 31P NMR spectroscopy.
effect, two P signals separated by 0.038 ppm were observed,
thus clearly indicating exclusive nucleophilic attack at phos-
phorus for this compound. Furthermore, comparisons of the
1
3
C NMR spectra of 2 taken in D2O and chloroform-d solutions
1
3
with the C NMR spectra of 12 formed in the alkaline
hydrolysis of 2 in D2O indicated that 2 is indeed the starting
material and that the Na salt of 12 is the product in our kinetic
studies.
Kinetics. Kinetic studies by NMR spectroscopy can be
properly conducted only when ratios of reacting species can be
extracted quantitatively from the peak areas. This is particularly
important because of the relatively long T1 relaxation times of
1
2
The Wittig reaction of methylenetriphenylphosphorane with
ketone 9 gave 1-methoxy-4-methylene-1-oxophosphorinane (10)
1
7
in 66% yield. Hydroboration of 10 followed by oxidation led
to two products which were identified as a 1:1 mixture of cis/
trans-4-(hydroxymethyl)-1-methoxy-1-oxophosphorinane (11),
and 4-(hydroxymethyl)-1-hydroxy-1-oxophosphorinane (12).
Several attempts to cyclize 11 in the presence of acidic or
basic catalysts failed. However, when crude 12 was treated with
DCC/DMAP18 for 9 h under reflux, the cyclization occurred
and 2 was isolated in 64% yield. We also observed fairly
efficient cyclization to 2 when 12 was warmed in Vacuo.
31
the P nucleus in our case. In order to measure ratios of esters
31
1
a, 2, 3, and 4 to their respective monosodium salts by P NMR
31
spectroscopy in a quantitative manner, measurements of P T1
relaxation times were required (Table 2).
Because an acceleration factor for the basic hydrolysis of the
1
13
31
The structure of 2 was substantiated by H, C, and P NMR
4
pair of esters 1 and 3 was established previously, we attempted
1
spectroscopies. Whereas the H NMR spectrum was straight-
to follow the basic hydrolysis of 2 and 4 under the same
conditions (1,4-dioxane/H2O, 60:40, v/v, at 304 K) and to use
P NMR spectroscopy as we did earlier. This attempt failed
1
3
forward, the C NMR spectrum displayed some intriguing
features. The signal for C-3 was found ca. 11.5 ppm downfield
of chemical shifts of the analogous carbons (CH2OH) in the
monocyclic isomeric esters 11. Furthermore, the methine carbon
in 2 resonated 13 ppm upfield compared with that in 11, and
revealed a very large three-bond coupling (47.6 Hz) to
phosphorus. The latter phenomenon has its precedent in bicyclic
phosphine oxides such as 1-oxo-1-phosphabicyclo[2.2.1]heptane
31
in the case of 2 because only a single resonance at 38.51 ppm
for the sodium salt of 12 was found after ca. 2 min, and no
other signals emerged within the next few hours. When 1/2
equiv of base was used at room temperature, signals at 54.98
and 38.51 ppm, assigned to bicyclic 2 and the sodium salt of
2, respectively, were observed. With this amount of base, we
were able to follow the kinetics of hydrolysis of 2 at 1 °C for
h, at which point the base was exhausted. Unfortunately, the
1
1
3 and 1-oxo-1-phosphabicyclo[2.2.2]octane (14), for which
1
data obtained did not follow the second-order kinetics equation
expected for this compound, which was based on the kinetic
2
5-27
results obtained for its acyclic counterpart 4.
It became
obvious that, under pseudo-first-order conditions, the basic
hydrolysis of 2 is a fast process which at room temperature takes
less than 1 min for completion.
Because kinetic studies of 2 required rate data at temperatures
below 0 °C, dioxane had to be replaced by another organic
solvent; 1,2-dimethoxyethane (DME) was chosen. However,
3
19
J(CCCP) values of 63 and 47 Hz, respectively, were observed.
31
Finally, comparison of the P NMR spectra of 11 and 2 showed
an upfield shift by ca. 7 ppm for the bicyclic compound. Similar
shielding of the phosphorus atom was observed earlier for 1a
the same ratio of D2O to DME, as well as the base and ester
3
1
20
31
(
-
δ( P) -7.97) compared with triethyl phosphate (δ( P)
concentrations, was maintained as reported earlier.4 In all our
1.0). Significant upfield shifts of C-4 in the 13C and P in
2
1
measurements the reactions followed good first-order kinetics
over 2 half-lives. Thus, ln x ) f(t) is a straight line with a
regression coefficient r ) 0.999. For compounds 1a and 2,
the kinetic runs were repeated twice, and the calculated rate
constants agreed within less than (5% for 1a and within less
3
1
the P NMR spectra of 2 compared with 11 may be caused by
the relatively close proximity of these nuclei in 2. It may be
3
that the value of J(C-4, P) in 2 is solely a result of multipath
coupling,19 but through-space spin interactions cannot be ruled
out.
31
than (2% for 2. In a separate P NMR experiment, it was
Mechanism of Hydrolysis in Strong Alkali and Product
Identification. Nucleophilic attack of hydroxide ion at the
phosphorus atom during hydrolysis of 1a and 3 was confirmed
demonstrated that 2 is not hydrolyzed by a small excess of
31
glacial acetic acid. Thus, the P signal of 2 at 54.76 ppm
remained when a solution containing acid-quenched aqueous
4
by Gorenstein et al. by comparing rates of hydrolysis of these
31
base, dioxane, and 2 was subjected to P NMR observation
1
8
esters and their thiono counterparts. Using O-enriched
under the same conditions as in the kinetics experiments. For
compound 4 a single series of experiments at five temperatures
over a 20 °C range was obtained. The correlation coefficient
for the ln k Vs 1/T relationship was found to be 0.9994, and the
water, Haake and co-workers22 showed that basic hydrolysis of
4
involves exclusive attack at phosphorus. Although it is
(
17) Brown, H. C.; Snyder, C.; Subba Rao, B. C.; Zweifel, G.
Tetrahedron 1986, 42, 5505.
(
(
(
(
18) Karanewsky, D. S.; Badia, M. C. Tetrahedron Lett. 1986, 27, 1751.
19) Wetzel, R. B.; Kenyon, G. L. J. Am. Chem. Soc. 1974, 96, 5189.
20) Verkade, J. G.; King, R. W. Inorg. Chem. 1962, 1, 948.
(23) Cox, J. R., Jr.; Ramsay, O. B. Chem. ReV. 1964, 64, 317.
(24) Cullis, P. M. In Phosphorus-31 NMR Spectral Properties in
Compound Characterization and Structural Analysis; Quin, L. D., Verkade,
J. G., Eds.; VCH Publishers: New York, 1994; p 315.
(25) Larsson, L.; Wallerberg, G. Acta Chem. Scand. 1966, 20, 1247.
(26) Aksnes, G.; Bergesen, K. Acta Chem. Scand. 1966, 20, 2508.
(27) Yuan, C.; Li, S.; Liao, X. J. Phys. Org. Chem. 1990, 3, 48.
21) Mark, V.; Dungan, C.; Crutchfield, M.; Van Wazer, J. Topics in
Phosphorus Chemistry; Interscience: 1967; New York, Vol. 5, p 332.
22) Cook, R. D.; Diebert, C. E.; Schwartz, W.; Turley, P. C.; Haake, P.
J. Am. Chem. Soc. 1973, 95, 8088.
(

Basic Hydrolysis of Phosphate Esters
J. Am. Chem. Soc., Vol. 118, No. 42, 1996 10173
q
q
calculated activation parameters ∆H and ∆S (Table 3) based
2
6,27
on these data are in good agreement with literature values.
Because of this good agreement, only a single series of kinetic
runs was conducted for 3.
Comparison of the rate constants extrapolated to 304 K (Table
1
2
) for 2 and 4 shows that hydrolysis of the bicyclic phosphinate
is 32 × 10 times faster than that of its acyclic counterpart 4,
3
while the rate enhancement for the bicyclic triester 1a relative
3
to acyclic 3 is only 1.3 × 10 . Although significant errors can
be introduced through extrapolation, there is no doubt that the
basic hydrolysis acceleration for the phosphinate pair 2/4 is more
than an order of magnitude greater than that for the phosphate
pair 1a/3. Assuming that this type of acceleration is associated
4
mainly with stereoelectronic effects, the present data suggest
that such influences are absent in the base hydrolysis of 1a. A
more definitive test for the nonexistence of stereoelectronic
effects in the hydrolysis of the phosphates investigated here
requires a discussion of activation parameters.
Figure 1. Molecular structure of 2 (PO
2 6 1 1 1
C H11, fw 146.13, P2 2 2 ,
Mechanistic and Structural Considerations. The experi-
mental enthalpies and entropies of activation for the compounds
studied here are collected in Table 3, and the differences in
these parameters for pairs of esters are given in Table 4 along
with their hydrolysis accelerations. Before addressing the
relative rates of the bicyclic esters 1a and 2 and of the acyclic
esters 3 and 4, we address first the acceleration of the hydrolysis
of bicyclic 1a over its acyclic analogue 3 and of bicyclic 2 over
its acyclic counterpart 4.
colorless, a ) 10.147(3) Å, b ) 10.400(2) Å, c ) 6.425(2) Å, Z ) 4,
w
R ) 0.0465, GOF ) 1.68). Ellipsoids are drawn at the 50% probability
level.
compared to the P-O bonds in 1a (1.58 Å) is apparently
compensated by a decrease of the P-C(6)-C(5) and P-C(7)-
C(8) bond angles in 2 to ca. 109° from ca. 115° for the P-O-C
angles in 1a, and an increase of the C(7)-C(8)-C(9) and C(6)-
C(5)-C(4) angles to ca. 113° in 2 Vs 108.7° for the O-C-C
angles in 1a. At the site of cleavage in 2, the P-O(2) bond
length and the P-O(2)-C(3) angle have almost the same values
as their respective counterparts in 1a. Although the O(2)-C(3)
bond in 2 is shorter (1.47 Å) compared with the O-C bonds in
Recent calculations at the HF/3-21+G(d,d) and MP2/6-
1+G* levels28 for (MeO)2P(O)OH and its cyclic analogue
3
(
CH2O)2P(O)OH revealed that although the strain in the ground
29
state of the latter molecule postulated earlier was confirmed
to exist, it did not contribute to the rate acceleration of the cyclic
over the acyclic ester. Moreover, these calculations showed
1
a (1.53-1.56 Å), the former value was previously observed
3
6
in 1,3,2-dioxaphosphorinanes.
Thus, we do not expect a
q
q
significant influence of strain arising from bond angle distortions
that ∆H and ∆G for formation of both trigonal bipyramidal
TBP) transition states (TS) were only negligibly different. It
in 1a or 2.
(
-
was also found that the destabilization of the cyclic TS that
must then be compensating for the ground-state destabilization
caused by strain stems from the impossibility of the cyclic TS
to achieve the most favored conformation adopted by the acyclic
Upon attack of 1a or 2 by an OH to form a five-coordinate
TS, the XPX angle that becomes the Xeq-P-Xeq angle increases
from ca. 103° to about 120° and the OPX angle that becomes
the Oax-P-Xeq angle decreases from 103° to about 90°.
Examination of molecular models reveals that a dominant effect
of these changes is to increase the P-Oax-C angle beyond 115°,
while other interactions already present in the bicyclic frame-
work remain about the same. Although the degree to which a
q
TS. It was concluded from further calculations that ∆G for
the cyclic phosphate ester was more greatly lowered by solvation
q
effects than ∆G for the acyclic analogue, thus accounting for
the rate acceleration of the former over the latter molecule.
1-phosphabicyclo[2.2.2]octane skeleton can accommodate bond
Because of the presence of six-membered rings in 1a and 2,
ground-state strain in these bicyclic molecules is likely to be
considerably less than in the five-membered ring (CH2O)2P-
angle changes while adopting a TBP structure is yet to be
established, we suggest that the transition state in the hydrolysis
of 1a and 2 is a strained trigonal bipyramid whose conformation
is even more unfavorable than that adopted by (CH2O)2P(O)-
3
0-32
(
O)OH discussed above. Indeed, compared with acyclic
3
3,34
and monocyclic
analogues, no significant distortions of the
2
8
OH. This may be particularly true for the TS of 2 in which
there are four pairs of synperiplanar H-H interactions arising
from adjacent CH2-CH2 units, plus four syn-1,3 H-H interac-
tions. By contrast there are no synperiplanar hydrogens in 1a,
and only three syn-1,3 H-H interactions are present. Com-
pensating for the H-H interactions in 2, however, are the lower
energies observed for trigonal bipyramids having two equatorial
C atoms than for comparable trigonal bipyramids containing
O-P-C and C-P-C bond angles are observed for 2 (Figure
). Morevoer, there are no unexpected differences between the
1
35
structural metrics for 2 and 1a. In the phosphorinane ring of
, lengthening of the P-C(6) and P-C(7) bonds (1.78 Å)
2
(
28) (a) Dejaegere, A.; Karplus, M. J. Am. Chem. Soc. 1993, 115, 5316.
b) Dejaegere, A.; Liang, X.; Karplus, M. J. Chem. Soc., Faraday Trans.
994, 90, 1763.
29) (a) Westheimer, F. H. Acc. Chem. Res. 1968, 1, 70. (b) Kluger, R.;
Taylor, S. D. J. Am. Chem. Soc. 1990, 112, 6669.
(
1
(
1
a
two equatorial O atoms. A schematic enthalpy diagram for
the hydrolyses of 1a and 2 is given in Figure 2. In this figure
the assumption that the initial energy states for 1a and 2 are
very close is based on a lack of evidence of strain within either
bicyclic framework. The close proximity of the energy states
for the hydrolysis products is based on the assumption that the
six-membered rings are strain free. It seems likely then that
(
30) Giordano, F.; Ripamonti, A. Acta Crystallogr. 1967, 22, 678.
(31) Schomburg, D.; Stelzer, O.; Weferling, N.; Schmutzler, R.; Sheld-
rick, W. S. Chem. Ber. 1980, 113, 1566.
32) Herceg, M.; Bronzan-Planinic, P.; Meider, H.; Matkovic, B.
Polyhedron 1986, 5, 2013.
33) Yamamoto, H.; Yamamoto, K.; Inokawa, S.; Luger, P. Carbohydr.
Res. 1983, 113, 31.
34) Inokawa, S.; Yamamoto, K.; Kawamoto, H.; Yamamoto, H.;
Yamashita, M.; Luger, P. Carbohydr. Res. 1982, 106, 31.
35) Nimrod, D. M.; Fitzwater, D. R.; Verkade, J. G. J. Am. Chem. Soc.
968, 90, 2780.
(
(
(
(
(36) Patois, C.; Ricard, L.; Savignac, P. J. Chem. Soc., Perkin Trans. 1
1990, 1577.
1

10174 J. Am. Chem. Soc., Vol. 118, No. 42, 1996
Wr o´ blewski and Verkade
interaction. Moreover, the rates of their acyclic counterparts 3
and 4 are quite comparable. Perhaps the lower enthalpy required
to achieve the five-coordinate TS for 1a over 2 is dominated
by the observation that two equatorial carbons in the TS of 2
give rise to a lower energy than two oxygen atoms occupying
1
a
these positions.
Conclusions. The approximately 200-fold observed rate
enhancement in the basic hydrolysis of bicyclic phosphinate 2
over bicyclic phosphate 1a is entirely enthalpic. This result is
interpreted in terms of a lack of a stereoelectronic effect in the
3
hydrolysis of six-membered ring phosphates. The 10 -fold
accelerations in the hydrolyses of bicyclic esters over their
acyclic counterparts appear to be largely due to solvation effects.
Acknowledgment. The authors are grateful to the donors
of the Petroleum Research Fund, administered by the American
Chemical Society, and to the NSF for grant support this research.
The X-ray data collection and structure solution for 2 were
carried out by Dr. Lee Daniels of the Iowa State Molecular
Structure Laboratory. The assistance of Dr. S. Jankowski in
the statistical analysis of the data and in calculating the activation
parameters is gratefully acknowledged.
Figure 2. Schematic enthalpy diagram for the hydrolysis of bicyclic
phosphate 1a and phosphinate 2.
solvation effects are largely responsible for the acceleration by
3
1
0 (Table 4) of the bicyclic esters over their acyclic analogues.
q
q
From the values of ∆H and ∆G in Table 3 it is seen that
Supporting Information Available: Tables of crystal-
lographic data and a computer drawing for 2 (8 pages). See
any current masthead page for ordering and Internet access
instructions.
q
∆
G for the hydrolysis of 1a is more negative than for 2, and
that these reactions are dominated by the enthalpy of activation
q
q
term. While it may be that ∆G solv lowers the overall ∆G for
more than for 1a, it is not clear why this should be so,
particularly because 2 contains fewer polar linkages for solvent
2
JA9611147