Mononuclear Co(III)-complex promoted phosphate diester hydrolysis: Dependence of reactivity on the leaving group

Article abstract of DOI:10.1002/poc.770

The TrpnCo(III)(OH)(OH₂)-promoted hydrolysis of a range of methyl aryl phosphate diesters was investigated at 37°C and I = 0.1 M (NaClO₄). The pH-rate profile confirms that the aqua-hydroxy form of the complex is the only kinetically

Full text of DOI:10.1002/poc.770

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2004; 17: 472–477
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.770
Mononuclear Co(III)-complex promoted phosphate
diester hydrolysis: dependence of reactivity on
the leaving groupy
Michela Padovani, Nicholas H. Williams* and Paul Wyman
Centre for Chemical Biology, Department of Chemistry, University of Sheffield, Sheffield S3 7HF, UK
Received 20 November 2003; revised 23 December 2003; accepted 23 December 2003
ABSTRACT: The TrpnCo(III)(OH)(OH2)-promoted hydrolysis of a range of methyl aryl phosphate diesters was
investigated at 37 ꢀC and I ¼ 0:1 M (NaClO4). The pH–rate profile confirms that the aqua-hydroxy form of the
complex is the only kinetically significant ionic form. At pH 6.9, all the reactions are first order in both diester and
Co(III) complex. Plotting the second-order rate constant for Co(III) complex-promoted hydrolysis against the pKa of
the leaving aryloxy group revealed a bent LFER indicating a change in rate-limiting step. This is discussed in terms of
either a change from rate-limiting hydrolysis to rate-limiting binding or the presence of a phosphorane intermediate.
Copyright # 2004 John Wiley & Sons, Ltd.
KEYWORDS: phosphate diester; metal ion catalysis; hydrolysis; linear free energy relationship; phosphorane
INTRODUCTION
mononuclear Co(III) complexes with cis binding sites in
promoting hydrolysis of bis-4-nitrophenyl phosphate.
Despite the similarities in structure, the complexes 1, 2
and 3 showed relative reactivities of 1:50:300. These data
were rationalized by considering the increase in strain at
the metal ion centre as it becomes part of a four-
membered ring including the reacting phosphate diester
(as shown schematically in Fig. 1). The angle ꢀ becomes
more acute as the transition state is formed, and this is
facilitated if the ligand can accommodate an obtuse
complementary bite angle ꢁ.
However, to understand better whether the metal ion-
promoted pathway involves any change in transition state
at the phosphate and to be able to extend these data to
different substrates, systematic study of the effect of
changing the substrate structure is required. To show
that the differences correspond to rate-limiting diester
cleavage for diester substrates, it was demonstrated that
bis-2,4-dinitrophenyl phosphate reacted more rapidly in
each case.7 However, the effect of varying the leaving
group on the transition state structure of the phosphate
diester was not investigated in detail. In this paper, we
describe the TrpnCo (3)-promoted hydrolysis of the
methyl aryl phosphates (4).
Phosphate esters are extremely kinetically stable to
hydrolysis under physiological conditions.1 This means
that the enzymes that catalyse these processes are the most
proficient established to date.2 Many of these enzymes
require metal ion cofactors,3 which are often in intimate
contact with the substrate at the active site. Hence ex-
tensive efforts over the years have been directed at under-
standing the link between these interactions and efficient
catalysis.4 A central question is whether the same transi-
tion state as in solution is stabilized by the contact with the
ions, or whether they significantly change the nature of
the transition state.5 Model complexes provide the most
accessible route to investigating this question in detail for
defined interactions between metal ions and substrates.
Metal ion complexes also form the basis of the most ef-
fective artificial systems that have been created in efforts to
reproduce enzyme-like reactivity. Since the structure of the
ligands coordinating the metal ions are the only means to
control the modes of action and activity of a particular ion,
rational improvement will depend on understanding how
activity and ligand structure are related. Owing to its
defined coordination geometry and relatively slow rates
of ligand exchange, Co(III) has been used extensively to
explore metal ion-promoted phosphate hydrolysis.6
EXPERIMENTAL
A key example is the systematic investigation by Chin
et al.7 on the effect of ligand structure on the reactivity of
Materials
*Correspondence to: N. H. Williams, Centre for Chemical Biology,
Department of Chemistry, Universityof Sheffield, Sheffield S37HF, UK.
E-mail: n.h.williams@sheffield.ac.uk
ySelected paper part of a special issue entitled ‘Biological Applications
of Physical Organic Chemistry dedicated to Prof. William P. Jencks’.
Acetone and tert-butyl alcohol were dried over activated
˚
4A molecular sieves and distilled prior to use. Dichlo-
romethane was distilled from calcium hydride prior to use.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 472–477

Co(III)-PROMOTED PHOSPHATE DIESTER HYDROLYSIS
473
8.09 (1H, d, J ¼ 9:2); ꢂP (101 MHz, D2O) ꢃ 3.62. 4g: ꢂH
(250 MHz, D2O) 3.66 (3H, d, J ¼ 11:6 Hz), 7.69 (1H, dd,
J ¼ 9:1, 0.6 Hz), 8.50 (1H, dd, J ¼ 9:1, 0.6 Hz), 8.91
(1H, dd, J ¼ 2:7, 0.6 Hz); ꢂP (101 MHz, D2O) ꢃ 3.81.
To prepare isotopically labelled dimethyl phosphate,
18O-enriched water (97% 18O) (0.075 ml, 4.16 mmol) was
added to potassium tert-butoxide (0.8 g, 8.33 mmol) in
10 ml of tert-butyl alcohol. Dimethyl chlorophosphate
(0.38 ml, 3.75 mmol) was added and the solution stirred
overnight. The solution was then lyophilized and the
resulting white solid was crystallized from ethanol to
yield dimethyl [18O] phosphate (0.2 g, 32%). ꢂH
(250 MHz, D2O) 3.63 (6H, d, J ¼ 10:7 Hz, CH3); ꢂP
(101 MHz, D2O) 3.56. MS (ES): 126.9 (%TIC: 68.18,
18O-enriched dimethyl phosphate), 124.9 (%TIC: 28.26,
dimethyl [16O]phosphate).
Figure 1. As the transition state for metal hydroxide attack
at the phosphate is reached, the angle ꢀ has to become
acute. This is facilitated if the complementary angle ꢁ can
become obtuse
Kinetic methods
All solutions were made up with doubly distilled, deio-
nized water and AnalaR-grade reagents. Alkaline and
acidic solutions were made up from BDH ConvoL
ampoules. The buffers used in determining the pH–rate
constant profile were MES, MOPS, EPPS, CHES and
CAPS and were freshly prepared for each experiment.
The experiments at pH 6.9 used MOPS buffer. The stock
TrpnCo (3) solution was titrated to the relevant buffer pH
before mixing with the buffer prior to each experiment.
The pH was measured at 37 ꢀC before and after each run
and did not vary in the course of the experiment. All
UV–visible readings were taken on a Varian Cary 1 Bio
UV–visible spectrophotometer and first-order rate con-
stants obtained by fitting the observed changes in absor-
bance to a first-order exponential curve with Cary Win-
UV software. In all experiments, TrpnCo (3) was in
ꢄ10-fold excess over the phosphate substrate. For the
slower reactions (4a and 4b), the reaction progress was
monitored by quenching aliquots using an equal volume
of pH 12.5 phosphate buffer (2 M), analysing them by
HPLC and using an initial rate analysis of the data.
Solvents for HPLC were of HPLC grade, filtered through
Sartolon polyamide 0.2 mm filters and degassed with
helium; the column used was a Phenomenex Luna
C18(2) 5 mm, 250 ꢂ 4:6 mm i.d. reversed-phase column
and the eluent was a 60:40 methanol–20 mM ammonium
phosphate buffer (pH 5.5). Both the UV–visible and
HPLC methods were used with 4c and gave excellent
agreement.
cis-Diaqua-Co(III)-tris(3-aminopropyl)amine was syn-
thesized according to previously published methods.8
Methyl aryl phosphate diesters were synthesized acco-
rding to the following general procedure. Dry imidazole
(0.26 g, 3.7 mmol) was dissolved in dichloromethane
(5 ml) and added slowly to a solution of phenol
(3.7 mmol) in dichloromethane (5 ml). Dimethyl chlor-
ophosphate (0.4 ml, 3.7 mmol) in dichloromethane (2 ml)
was added slowly with stirring and the solution was
stirred at room temperature for 16 h. The resulting pre-
cipitate was removed by filtration, washed with dichlor-
omethane and the filtrate concentrated in vacuo. The
crude triester was purified by flash chromatography on
silica if necessary, then dissolved in dry acetone (5 ml)
and added dropwise to a stirred solution of lithium
chloride (0.16 g, 3.7 mmol) in dry acetone (ꢁ25 ml).
The solution was heated at reflux for 30 min and allowed
to stand overnight. The precipitate was filtered off,
washed with dry acetone (3 ꢂ 10 ml) and air dried to
yield the lithium salt of the methyl aryl phosphate diester.
4a: ꢂH (250 MHz, D2O) 3.70 (3H, d, J ¼ 11 Hz), 7.01–
7.12 (2H, m), 7.15–7.22 (1H, m), 7.29–7.35 (1H, m); ꢂP
(101 MHz, D2O) ꢃ 2.38. 4b: ꢂH (250 MHz, D2O) 3.68
(3H, d, J ¼ 11:3 Hz), 7.65–7.58 (2H, m), 8.11–8.09 (2H,
m); ꢂP (101 MHz, D2O) ꢃ 2.47. 4c: ꢂH (250 MHz, D2O)
3.68 (3H, d, J ¼ 11:3 Hz), 7.50 (2H, d, J ¼ 8:5 Hz), 8.17
(2H, d, J ¼ 8:5 Hz), ꢂP (101 MHz, D2O) ꢃ 2.90. 4d: ꢂH
(250 MHz, D2O) 3.65 (3H, d, J ¼ 11:3 Hz), 8.40 (2H, dd,
J ¼ 2:2, 1.2 Hz), 8.90 (1H, t, J ¼ 2:1 Hz); ꢂP (101 MHz,
D2O) ꢃ 3.01. 4e: ꢂH (250 MHz, D2O) 3.65 (6H, d,
J ¼ 11:3 Hz), 6.78–6.64 (1H, m), 7.27–7.17 (1H, m),
8.17–8.06 (1H, m); ꢂP (101 MHz, D2O) ꢃ 3.42. 4f: ꢂH
(250 MHz, D2O) 3.70 (3H, d, J ¼ 11:3 Hz), 7.50 (1H,
ddd, J ¼ 9:2, 2.5, 0.9 Hz), 7.71 (1H, dd, J ¼ 2:5, 0.6 Hz),
RESULTS
The pH dependence of the hydrolysis reaction was
measured at 37 ꢀC using 4f as the substrate and a TrpnCo
(3) concentration of 1 mM, 10 mM buffer and 0.1 M
NaClO4 (Fig. 2). These data were fitted to a single
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 472–477

474
M. PADOVANI, N. H. WILLIAMS AND P. WYMAN
Table 1. Second-order rate constants k2 for the TrpnCo
(3)-promoted hydrolysis of diesters (4) and pKa values
of the aryloxy leaving groups at 37 ꢀC, in the presence of
10 mM buffer and 0.1 M NaClO4
Substrate
k2 (Mꢃ1 sꢃ1
)
Leaving group pKa
ꢃ6
ꢃ5
4a
4b
4c
4d
4e
4f
(9.9 ꢅ 0.3) ꢂ 10
(6.3 ꢅ 0.3) ꢂ 10
0.69 ꢅ 0.01
2.96 ꢅ 0.03
4.31 ꢅ 0.05
8.0 ꢅ 0.1
8.73 ꢅ 0.03
8.13 ꢅ 0.02
6.83 ꢅ 0.01
6.36 ꢅ 0.01
5.92 ꢅ 0.01
5.11 ꢅ 0.03
3.89 ꢅ 0.01
4g
15.4 ꢅ 0.5
Figure 2. pH--rate constant profile for the hydrolysis of
0.05 mM 4d promoted by 1 mM 3 (37 ꢀC, 10 mM buffer
and 0.1 M NaClO4). The line shown is a non-linear least-
squares fit to a scheme where a single ionisation controls the
pH dependence, and gives a kinetic pKa of 7.11 ꢅ 0.06 and a
maximal rate of (9.0 ꢅ 0.9) ꢂ 10ꢃ3 sꢃ1
ionization event with pKa ¼ 7.11 ꢅ 0.06. None of the
substrates have ionizable groups within the pH range
measured, and all subsequent reactions were followed at
pH 6.9. Using 4c as substrate, the rate of hydrolysis
promoted by 1.2 mM TrpnCo (3) was measured with and
without 10 mM MOPS at pH 6.9 in the presence of 0.1 M
NaClO4 and showed a 10% increase in the observed rate
constant; however, in the absence of added NaClO4, the
rate constants decreased by a similar proportion with the
same change in buffer concentration. Thus, at the buffer
concentrations used, buffer catalysis was ignored as a
significant contribution to the observed reaction rate. All
the subsequent reactions were buffered at pH 6.9 using
10 mM MOPS buffer and the buffering action of the
complex itself, in the presence of 0.1 M NaClO4.
To establish the order of reaction in complex, the
TrpnCo (3) concentration was varied over the range
0.8–3.2 mM for each of substrates 4a–g (maintaining at
least a 10-fold excess of complex over substrate). In each
case, a first-order curve gave an excellent fit to the data,
showing the reaction to be first order in diester. Plotting
these pseudo-first-order-rate constants against TrpnCo (3)
concentration gave good linear plots, confirming that the
reaction is also first order in metal ion complex, and
linear least-squares fitting provided the second-order rate
constants for complex promoted hydrolysis of the diester
(Table 1).
Figure 3. Linear free energy relationship between the sec-
ond-order rate constant for hydrolysis of esters 4 promoted
by 3 and the pKa of the aryl leaving group at 37 ꢀC in the
presence of 10 mM buffer and 0.1 M NaClO4. The solid and
dotted lines are the non-linear least-squares fits to the
reaction schemes described in the text. The dashed line
represents the rate of binding of inorganic phosphate with
3 at 25 ꢀC and pH 7 as reported by Chin et al.7
When dimethyl phosphate was used as substrate (6 mM
in the presence of 2 mM complex), no hydrolysis could be
detected over a period of 3 days at 37 ꢀC, as expected.
Similarly, no change in the isotopic ratio of 18O-enriched
dimethyl phosphate could be detected by ESMS. The
complex does not appear to catalyse exchange of solvent
oxygen into the non-bridging phosphoryl positions.
DISCUSSION
The pH–rate profile is consistent with the assumption that
the only active ionic form of the complex is the aqua-
hydroxy form, as previously described by Chin et al.,7
and has been fitted to this scheme, revealing a kinetic pKa
of 7.1; this is consistent with the previously reported
values for the second pKa of TrpnCo (3). As the associa-
tion constant has been measured as ꢁ1 Mꢃ1,9 the propor-
tion of the substrate that is complexed with the TrpnCo 3
The pKa of each of the phenol leaving groups was
measured by spectrophotometric titration under the ex-
perimental conditions (37 ꢀC, 10 mM buffer and 0.1 M
NaClO4); these are recorded in Table 1. The second-
order rate constants were plotted against these pKas
to generate a linear free energy relationship (LFER)
(Fig. 3).
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 472–477

Co(III)-PROMOTED PHOSPHATE DIESTER HYDROLYSIS
475
the LFER (Fig. 3, dashed line); this difference is a lower
limit due to the 12 ꢀC temperature difference. This
appears to suggest that the change in rate-limiting step
is unlikely to be due to rate-limiting binding.
If this is not the rate-limiting step, then the reaction of
the bound diester would have to involve an intermediate
(pathway B). The likelihood of such an intermediate is
open to debate; in the course of studying many Co(III)-
promoted reactions, Sargeson and colleagues have
frequently invoked pentacoordinate intermediates even
for monoester hydrolysis.13 Recent heavy atom kinetic
isotope studies have led to the suggestion that in dinu-
clear complexes, diesters with such good aryl oxy leaving
groups could form a phosphorane-like intermediate.14
However, related evidence for the reaction between
similar phosphate diester substrates and hydroxide is in
favour of a concerted mechanism.15 Similarly, investiga-
tions of Co(III)-promoted hydrolysis of phosphate die-
sters using heavy atom kinetic isotope effects have been
interpreted in terms of a single step concerted reaction
(pathway A).10 If such a phosphorane intermediate exists
on the reaction pathway, rate-limiting attack at phos-
phorus would be expected to be relatively insensitive to
the nature of the leaving group (assuming binding is
independent of leaving group pKa). As bond fission to the
leaving group becomes rate limiting, ꢃlg increases in
magnitude as expected, reflecting the expected increase
in effective charge development on the O atom in the
rate-limiting transition state.
If a phosphorane intermediate is reversibly formed (i.e.
for diesters with relatively poor leaving groups), then it
opens up the possibility of observing exchange of oxygen
between the solvent and the non-bridging phosphoryl
oxygens of the diester. To explore this, dimethyl phos-
phate enriched in the non-bridging positions with 18O
was incubated with TrpnCo (3) for 3 days. No hydrolysis
was observed over this time-scale, as would be expected
by extrapolating the LFER to a leaving group of
pKa ¼ 15.5. However, no measurable change in the iso-
topic composition of the dimethyl phosphate is observed.
This is perhaps unsurprising, as the exchange process
would require both a proton transfer and pseudo-rotation
step to occur to allow the nucleophilic hydroxide become
incorporated into the diester. Furthermore, the TrpnCo (3)
complex is not stable for long periods in aqueous solu-
tion, and so a decrease in catalyst concentration may also
be occurring on these time-scales. Hence the absence of
exchange does not preclude the presence of such an
intermediate.
Scheme 1
is not significant under our experimental conditions, and
the first-order dependence on both complex and substrate
reflect the presence of only one of each in the rate-
limiting transition state. This is also consistent with
Chin et al.’s7 earlier report and shows that there is no
contribution from an active dimer as observed for 2 at
high concentrations.10
However, the LFER is clearly non-linear, indicating a
change in rate-limiting step when the leaving group has
pKa ꢆ 6; the solid line shown is a non-linear least-squares
fit to this scheme, and indicates that ꢃlg changes from
ꢃ 0.17 to ꢃ 1.05 as the leaving group becomes poorer.
The likely mechanisms are shown in Scheme 1. Under
the conditions used, there is initially a complexation step,
followed by intramolecular attack of the metal-bound
hydroxide which displaces the aryloxy leaving group
either in a single step (pathway A) or via a pentacoordi-
nate intermediate (pathway B).
As Co(III) undergoes ligand exchange slowly, it is
plausible that the substrate binding step will become
rate limiting for highly reactive substrates.11 If this is
the case, then it is known that binding at Co(III) com-
plexes is very insensitive to the nature of the incoming
anion,11 and since the effect of varying the aryl moiety on
the binding properties of the diester will be small in any
case, the rate of the binding step will be constant for all
the compounds 4. Hence, to apply this interpretation, the
rate-limiting binding regime needs to be constrained to a
ꢃbinding value of zero. The resulting non-linear least-
squares fit is shown by the dotted lines on the graph,
and still leads to an acceptable fit to the data (with the
break now occurring at leaving groups with pKa ꢆ 5.7), a
similar ꢃlg for the poorer leaving groups of ꢃ 0.98 and a
The above data do not lead to a definitive distinction
between the two pathways, and need to be examined
more closely. First, there may be a difference between
(monoanionic) diester binding and (dianionic) inorganic
phosphate binding if preliminary ion pair formation is a
key step in the binding event. Assuming that this is the
case, Haim suggested that dianionic ligands may bind
approximately an order of magnitude more rapidly than
limiting rate constant of 13 Mꢃ1 sꢃ1
.
These data can be compared with the rate of com-
plexation of TrpnCo (3) by inorganic phosphate mea-
sured by Chin et al.7 at 25 ꢀC and pH 7, which leads to a
second-order rate constant for binding of 120 Mꢃ1 sꢃ1
,
which is ꢁ10-fold faster than the limiting rate shown by
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 472–477

476
M. PADOVANI, N. H. WILLIAMS AND P. WYMAN
monovalent anions.16 At the pH of Chin et al.’s7 binding
experiment, inorganic phosphate is 60% monoanion and
40% dianion. Thus, if the dianion does bind ꢁ10-fold
more rapidly owing to its higher charge, partitioning the
observed rate constant between the two species leads to
an estimate of 26 Mꢃ1 sꢃ1 for binding by monoanionic
phosphate at 25 ꢀC. This is close to the limit indicated by
the LFER, but anation at Co(III) has a high enthalpy of
activation17 and so the temperature difference will be a
significant factor. Assuming the same entropy of activa-
tion as reported for anation of the aqua-hydroxy form of
bis (diethylenediamine)-Co(III) by phosphate17 leads to
an estimated temperature factor of ꢁ4-fold for the 12 ꢀC
temperature increase, leaving the overall estimate for
binding at around 120 Mꢃ1 sꢃ1 as shown in Fig. 2.
Second, Banaszczyk et al.18 have reported the second-
order rate constant for the binding of sulfate to TrpnCo
(3) at pH 2.9 and 25 ꢀC. Assuming that the hydroxy-aqua
species is the only active form for binding leads to a
second-order rate constant of ꢁ11 Mꢃ1 sꢃ1, in contrast
with the data reported by Chin et al.7 for inorganic
phosphate binding. Nevertheless, the binding of mono-
anionic phosphate is clearly the most relevant data,
although the difference in binding rates (expected to be
independent of anion) is surprising. It is not clear from
the data reported by Banaszczyk et al.18 whether the rates
correspond to rate limiting binding or the rate of chela-
tion of the Co(III) centre by initially monodentate sulfate,
which may explain the discrepancy.
On the other hand, when considering the expected ꢃlg
for rate-limiting attack at the phosphate, it is noteworthy
that even the nucleophilic substitution of phenols by
oxygen nucleophiles with high pKa (which is most likely
to involve a mechanism of rate limiting phosphorane
formation), shows a greater sensitivity to the leaving
group (ꢃlg ꢆ ꢃ 0.35)19 than the very shallow ꢃlg
( ꢃ 0.17) obtained in the LFER reported here. In a similar
vein, the hydroxide-catalysed cyclization of both Up-
OAr20 and UpOR21 esters has been studied at 25 ꢀC with
a range of aryloxy and alkoxy leaving groups, respec-
tively. Interestingly, constructing an LFER from these
two data sets (Fig. 4) also yields a bent plot (with the
break occurring for leaving groups with pKa ꢆ 12.5), but
in comparison with Fig. 3, the shallower ꢃlg is substan-
tially larger in magnitude at ꢃ 0.52. The steeper ꢃlg is
also greater in magnitude than in Fig. 3 ( ꢃ 1.33). In this
case, there is no binding step to consider and the simplest
explanation is in terms of a pentacoordinate intermediate
(although the data may also be explained in terms of a
Hammond effect on a concerted pathway20).
Figure 4. Linear free energy relationship between the sec-
ond-order rate constant for hydroxide catalysed hydrolysis of
UpOAr and UpOR esters and the pKa of the aryloxy or alkoxy
leaving group at 25 ꢀC; data taken from Refs. 20 and 21,
respectively. The solid line is a non-linear least-squares fit to a
two-step reaction mechanism, giving ꢃlg values of ꢃ 0.52
and ꢃ 1.33 for good and poor leaving groups, respectively
ꢃ 1.0, this leads to an estimate of ꢃ 0.26 for the effective
charge on the leaving oxygen in the transition state.
Assuming that the charge reflects bond cleavage, bond
cleavage in the transition state would be ꢁ0.57 along the
reaction coordinate. This is further advanced than the
hydroxide promoted cleavage of the uncomplexed diester
(ꢁ0.37),23 and suggests more substantial bond formation
to the nucleophile in the transition state than for the
intermolecular reaction.
CONCLUSION
Overall, we have to conclude that these data and analyses
cannot unambiguously exclude rate-limiting binding for
the most reactive diesters; this is a conservative inter-
pretation of the data, but is not wholly convincing. If it
can be shown that binding is significantly faster than the
observed rate of hydrolysis for all phosphate diesters
under the conditions of our experiments, then it would
appear that the metal ion diverts the reaction into an
addition–elimination reaction from the concerted process
observed in the intermolecular attack of hydroxide. In
considering the phosphorane intermediate interpretation,
we note that the bound diester will have its negative
charge partially quenched by the metal ion, and so will be
somewhere between a diester and triester in character, so
the proposal of a pentacoordinate intermediate is not
unreasonable. Furthermore, the break in LFER occurs
close to the pKa of the incoming nucleophile, which is
where such a break is expected. At this point, the rate of
return to starting material and breakdown to products will
be in balance, and in a symmetrical system will occur
Regardless of the mechanisms, the LFER does provide
information about the transition state for the poorer
leaving groups. Following the effective charge analysis
of Williams,22 the effective charge on the O-Ar atom in
the starting diester is þ 0.74 and becomes ꢃ 1.0 in the
product aryloxy anion. As the ꢃlg for leaving groups
where the bond breaking step is clearly rate limiting is
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 472–477

Co(III)-PROMOTED PHOSPHATE DIESTER HYDROLYSIS
477
when the nucleophile and leaving group have matching
pKas. In this system, the nucleophile is part of a small
ring, and so its departure might be slightly enhanced
relative to its pKa. The same observation is made for the
transesterification uridyl esters, where the break is ob-
served at around 12.5, close to the pKa of the 20-hydroxyl
nucleophile.
In any case, the metal ion-promoted reaction leads to a
transition state where the leaving oxygen develops more
effective charge ( ꢃ 0.26) than the corresponding inter-
molecular hydroxide-promoted hydrolysis ( þ 0.10),23
suggesting that the transition state is changed and that
the way to more effective catalysts may be focus on
interactions at the leaving group.
4. (a) Chin J. Acc. Chem. Res. 1991; 24: 145–152; (b) Blasko A, Bruice
T. Acc. Chem. Res. 1999; 32: 475–484; (c) Williams NH, Takasaki B,
Wall M, Chin J. Acc. Chem. Res. 1999; 32: 485–493; (d) Molenveld P,
Engberson JFJ, Reinhoudt DN. Chem. Soc. Rev. 2000; 29: 75–86.
5. Kirby AJ, Jencks WP. J. Am. Chem. Soc. 1965; 87: 3209–3216.
6. Hendry P, Sargeson AM. Prog. Inorg. Chem. 1990; 38: 201–
258.
7. Chin J, Banaszczyk M, Jubian V, Zou X. J. Am. Chem. Soc. 1989;
111: 186–190.
8. Connolly JA, Banaszczyk M, Hynes RC, Chin J. Inorg. Chem.
1994; 33: 665–669.
9. Chin J, Zou X. Can. J. Chem. 1987; 65: 1882–1884.
10. Rawlings J, Cleland WW, Hengge AC. J. Inorg. Biochem. 2003;
93: 61–65.
11. Chin J, Zou X. J. Am. Chem. Soc. 1988; 110: 223–225.
12. Langford CH. Inorg. Chem. 1965; 4: 265–266.
13. (a) Jones DR, Lindoy LF, Sargeson AM. J. Am. Chem. Soc. 1983;
105: 7327–7336; (b) Jones DR, Lindoy LF, Sargeson AM. J. Am.
Chem. Soc. 1984; 106: 7807–7819.
14. Humphry T, Forconi M, Williams NH, Hengge AC. J. Am. Chem.
Soc. 2002; 124: 14860–14861.
15. Cassano AG, Anderson VE, Harris ME. J. Am. Chem. Soc. 2002;
124: 10964–10965.
Acknowledgements
16. Haim A. Inorg. Chem. 1970; 9: 426–428.
17. Lincoln SF, Stranks DR. Aust. J. Chem. 1968; 21: 1745–1756.
18. Banaszczyk M, Lee J-J, Menger F. Inorg. Chem. 1991; 30: 1972–
1973.
We thank the EPSRC for a studentship to P.W. and for
funding (GR/M16962) for M.P.
19. Khan SA, Kirby AJ. J. Chem. Soc. B 1970; 1172–1182.
20. Davis AM, Hall AD, Williams A. J. Am. Chem. Soc. 1988; 110:
5105–5108.
REFERENCES
1. (a) Williams NH, Wyman P. J. Chem. Soc., Chem. Commun. 2001;
1268–1269; (b) Lad C, Williams NH, Wolfenden R. Proc. Natl.
Acad. Sci. USA 2003; 100: 5607–5610.
¨
¨
21. Kosonen M, Youseti-Salakdeh E, Stromberg R, Lonnberg H. J.
Chem. Soc., Perkin Trans. 2 1997; 2661–2666.
22. (a) Williams A. Acc. Chem. Res. 1984; 17: 425–430; (b) Williams
A. Chem. Soc. Rev. 1986; 15: 125–140; (c) Bourne N, Williams A.
J. Org. Chem. 1984; 49: 1200–1204.
23. Williams NH, Cheung W, Chin J. J. Am. Chem. Soc. 1998; 120:
8079–8087.
2. Wolfenden R, Snider MJ. Acc. Chem. Res. 2001; 12: 938–945.
¨
3. (a) Strater N, Lipscomb WN, Klabunde T, Krebs B. Angew.
Chem., Int. Ed. Engl. 1996; 35: 2024–2055; (b) Wilcox DE.
Chem. Rev. 1996; 96: 2435–2458.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 472–477