Metal-catalyzed alcoholysis reactions of carboxylate and organophosphorus esters

Metal-catalyzed alcoholysis reactions of carboxylate and organophosphorus esters R. STAN BROWN and ALEXEI A. NEVEROV Department of Chemistry, Queen’s University, Kingston, Ontario K7L 3N6, Canada 1 2 3 4 5

Introduction 271 Background theory 274 Titrations in alcohol 276 Metal ion alcoholysis and titration in alcohol 278 Transition metal ion and Ln3+ catalysts of transesterifications of neutral carboxylate and organophosphate esters 284 Mechanism of alcoholysis of carboxylate esters 288 Mechanism of alcoholysis of neutral phosphate esters 294 6 Transition metal ion and La3+-catalysis of the alcoholysis of phosphate diesters 308 Metal-catalyzed alcoholysis of an RNA model 310 Zn2+ ligand models for dinuclear enzymes promoting the cleavage of RNA 316 Exhalted catalysis of methanolysis of HPNPP promoted by a dinuclear complex in methanol 318 7 Conclusions 324 Acknowledgements 325 References 325

1

Introduction

Metal ion-catalyzed hydrolytic reactions of esters, amides1 and organophosphorus esters2 have been studied for many years. In general these are relatively well-understood processes where the reactions involve catalytically active Mx+(
OH) species with the metal ion serving several possible purposes. These include: (1) decreasing the pKa of the metal associated HOH so that an Mx+-coordinated hydroxide can be formed at near neutral pH; (2) a bifunctional role where the Mx+(
OH) acts as a Lewis acid coordinating transiently to the CQO or PQO group, and then delivering the coordinated hydroxide and (3) possibly promoting accelerated breakdown of any intermediates through coordination of the leaving group, particularly a poor one, to the metal ion. While water is required as a reagent for the hydrolytic processes, in bulk media it solvates the ionic species very heavily so as to impede the interaction of the catalytic species with hydrophobic substrates such as neutral esters. In addition, the high dielectric constant dampens the attractive interaction between species of opposite charge. Not surprisingly, Mx+(
OH) species in water are relatively poor catalysts for acyl or phosphoryl transfer from neutral substrates. It seems reasonable that moving to less polar solvents with lower dielectric constants might ameliorate these problems and allow a more full realization of metal 271 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 42 ISSN 0065-3160 DOI: 10.1016/S0065-3160(07)42006-8

r 2008 Elsevier Inc. All rights reserved

272

R.S. BROWN AND A.A. NEVEROV

ion’s ability to catalyze acyl and phosphoryl transfer reactions, although there seems to be little known about the mechanisms of these processes. Among organic solvents the lower alcohols such as methanol and ethanol, while being close to water in terms of chemical properties, solvation effects and hydrogen-bonding characteristics, should offer opportunities to investigate this thesis. Transition metal or lanthanide metal ion-catalyzed alcoholysis reactions, while being important industrial processes for transesterification3 of carboxylate esters, are less well understood than the analogous hydrolysis reactions and have not been applied to phosphoryl transfer processes except in our labs. Buncel and co-workers4 and Mandolini and co-workers3f–i have extensively investigated the ethanolysis of phosphinate and phosphate esters promoted by alkali metal ethoxides and alkaline earth ethoxides (Ba2+) respectively, but these reactions occur stoichiometrically under highly basic conditions whereas the transition and lanthanide M2+- and M3+-catalyzed alcoholysis reactions we discuss below occur under essentially neutral pH conditions. One might expect that Mx+-promoted alcoholysis reactions will be subject to the same mechanistic generalities as Mx+-promoted hydrolysis reactions. The bulk of our studies show that this is the case although there are important mechanistic and practical differences which in certain ways simplify mechanistic study and circumvent at least two of the problems associated with metal hydroxides. First, in water at pH values above the pKa of the Mx+-coordinated H2O, the active Mx+-hydroxides form oligomeric and polymeric gels or precipitates that complicate the mechanistic analyses. In some cases this can be overcome by complexing the metal ion to ligands that block the oligomerization and prevent the precipitation of the Mx+-hydroxides. However, the latter approach is known to introduce other biases because the ligands employed are generally those which are readily available or easily synthesized and not necessarily the ideal ones for controlling geometry and reactivity of the complexes. Second, Mx+-hydroxides in water are not particularly effective intermolecular catalysts and so the most effective systems are ones where the metal ion is held, through transient or fixed association with some proximal binding group, close to the scissile CQO or PQO linkage. As will be seen later, one of the major stumbling blocks in creating highly active Mx+(
OH) catalysts is the fact that a metal-coordinated hydroxo group is less basic and nucleophilic than free hydroxide, a problem that is difficult to overcome in water but one that seems to be readily overcome in alcohols where the active species involve one or more Mx+(
OR) units. Moving to alcohol solvents substantially solves some of the above problems but requires due consideration of the control and measurement of pH in neat alcohol.5 Bosch and co-workers6 have described simplified methods for determining the ss pH in methanol solution which have greatly facilitated our work. Recently, we have For the designation of pH in non-aqueous solvents, we use the forms described by Bosch and co-

workers6 based on the recommendations of the IUPAC, In Compendium of Analytical Nomenclature. Definitive Rules 1997, 3rd edn, Blackwell, Oxford, UK, 1998. If one calibrates the measuring electrode with aqueous buffers and then measures the pH of an aqueous buffer solution, the term w w pH is used; if the electrode is calibrated in water and the ‘pH’ of the neat buffered methanol solution then measured, the term sw pH is used, and if the electrode is calibrated in the same solvent and the ‘pH’ reading is made, then the term ss pH is used.

METAL-CATALYZED ALCOHOLYSIS

273

shown that one can titrate M2+- and M3+-ions in methanol7 and in ethanol8 with the end result being that now one can perform detailed mechanistic studies of reactions as a function of ss pH in the presence of metal ions in these alcohols under buffered conditions as simply as one can in water. Importantly most of the M2+ and M3+ systems are completely soluble in alcohols throughout the ss pH region where formation of the metal ion alkoxides occurs, obviating the solubility problems encountered in water and in some cases making it unnecessary to employ ligands for successful study. The fact that the autoprotolysis constants for methanol and ethanol are 10
16.77 and 10
19.1 means that neutral ss pH in the two solvents is 8.46 and 9.558 while that in water is, of course, 7.00. These alcohols have additional benefits related to their low dielectric constants and polarity relative to that of water (31.5 (CH3OH), 24.3 (EtOH) vs. 78 (H2O))9 which favor enhanced organic substrate solubility, greater ion pairing between the catalytic metal ion and any anionic or dipolar substrates and greater Lewis acid interaction between neutral substrates and metal ions. This not only enhances binding, but also the rates of the subsequent acyl and phosphoryl transfer reactions. Interestingly, in the case of La3+ catalytically
active dimers, such as La3+ 2 ( OR)1,2,3 y, are formed spontaneously in methanol or ethanol without the need for any binding ligands. As will be shown, many acyl and phosphoryl transfer reactions to alcohol can be greatly accelerated by Mx+(
OR) species in contrast to the hydrolytic reactions in water where the Mx+(
OH) species often have little or no activity. Much of our earlier work on the Mx+-catalysis of acyl or phosphoryl transfer to methanol has been summarized10 and so will only be mentioned here when necessary for a brief background or when relevant to the more recent work. In what follows we will first consider some general background theoretical aspects of the influence of reduced polarity and dielectric constant media on metal ion-catalyzed reactions. Subsequently, a brief introduction is given for titration in alcohol to determine the x+ s (HOR) systems as well as the speciation of s pK a values for ionization of various M the metal ions, and when in the presence of various ligands, the formation constants for the ligand:Mx+ and ligand:Mx+(
OR) species in solution. Subsequently we deal with the effectiveness of some Mx+-ions, notably La3+, Eu3+, and some Zn2+ and Cu2+:ligand systems in promoting the alcoholysis of carboxylate esters and various neutral phosphate, phosphonate, phosphonothioate and phosphorothioate esters. These are cases where the substrates, being neutral but dipolar, are considered to be weakly binding to the metal ion. Thus, the kinetics do not exhibit saturation binding although chemical intuition states that complexes must be formed along the reaction pathway. Almost all these reactions exhibit spectacular rate accelerations when promoted by the Mx+(
OR) catalysts in the neutral ss pH regions relative to the background reactions which are promoted by lyoxide. Where we have studied the effect of varying the leaving group on the rates of transesterification, we now have some fairly detailed mechanistic information. Subsequently, we will deal with some anionic phosphate diester substrates that might be taken as models for RNA and DNA. These bind very tightly to the metal ion and then are subject to rapid metal-catalyzed decomposition at rates that exceed anything so far reported for the analogous reactions in water.

274

2

R.S. BROWN AND A.A. NEVEROV

Background theory

Aside from being fundamentally interesting and industrially important, phosphoryl and acyl transfer reactions are key biological processes. Numerous enzymes containing one or more metal ions (notably Zn2+, but in some cases Mg2+, Ca2+ and Fe2+) catalyze ester and peptide hydrolysis as well as phosphoryl transfer processes.11 It has been stated in the context of short, strong hydrogen bonds, that the ‘active sites of enzymes are non-aqueous, and the effective dielectric constants resemble those in organic solvents rather than that in water’,12a,b and it is perhaps surprising how sparse is the experimental literature concerning the effect of reduced dielectric constant of a solvent on simple reactions of biological relevance. An early example demonstrated that the decarboxylation of pyruvate promoted by 3,4-dimethylthiazolium ion (a model for thiamine pyrophosphate-catalyzed reactions in an enzyme’s interior) occurs 104–105 faster in ethanol than in water.13 From this Lienhard concluded that a large part of the thiamin-dependent enzymes’ catalysis may occur from the hydrophobicity of the active site promoting reactions where charge neutralization of the zwitterionic intermediate is neutralized in the transition state. It is an intriguing possibility that metal ion-catalyzed reactions of the sort we discuss herein, where charge is dissipated in the transition states, might be subject to large rate accelerations in solvents of reduced polarity and dielectric constant. Solvent effects can have important accelerating or decelerating effects on organic transformations depending on the overall solvation of the ground state starting materials and transition states for the rate-limiting steps of the reaction,14 but it is interesting to us how little application of reduced polarity medium effects there is for reactions of biological interest including metal ion-catalyzed processes. The generally accepted process for metal ion-catalyzed reactions of the sort we consider here involves pre-equilibrium binding with the substrate, followed by a reaction of the complex as schematized in Equation (1). Whether the metal ion is free or complexed by ligands, or bears an associated lyate, or whether the substrate is neutral or anionic, these appear to be just the sort of processes one might expect to experience large rate accelerations in passing from water to a medium of reduced dielectric constant such as alcohols or other lower polarity solvents. Lig: MX+ + Subst.

Kb

Lig: MX+:Subst.

kcat

Lig:MX+ + P

(1)

The Debye-Hu¨ckel theory for association of spherical ions in a medium of dielectric constant Dr posits that the electrostatic potential energy of interaction between oppositely charged ions is PE ¼ ðzþ eÞðz
eÞ=ð4pD0 Dr rÞ

(2)

where r is the distance between the centers of the ions, z+e and z
e are their charges in coulombs (e is the proton charge), D0 the permittivity of a vacuum, and Dr is the relative permittivity of dielectric constant of the medium.15 When the electrostatic attraction energy is greater than the thermal kinetic energy of the species in solution (given as its average translational kinetic energy of 3/2kT, where k is the Boltzmann

METAL-CATALYZED ALCOHOLYSIS

275

Table 1 Changes in hypothetical potential energies of attraction of oppositely charged spherical ions based on reductions in dielectric constant in passing from water to methanol and ethanol, and the computed changes in binding constant assuming the changes in potential energy are translated into free energies of bindinga Electrostatic PE (H2O, kJ mol
1)

Electrostatic PE (CH3OH, kJ mol
1)

Increase in Kb H2O methanol

Electrostatic PE (C2H5OH, kJ mol
1)

Increase in Kb H2O - ethanol

10.45 20.9 31.35 41.8 52.25

12.5 158 2  103 2.5  104 3.2  105

13.38 26.75 41.13 53.5 66.9

41 1.7  103 6.9  104 2.8  106 1.24  108

4.18 8.36 12.54 16.72 20.9 a

Kb computations made under assumption that all the P.E. of attraction is reflected in the DG of binding.

constant and T is the Kelvin constant) then the encounter complex lasts long enough for multiple collisions and achieves the status of an ‘ion pair’. A change from water (Dr ¼ 78) to methanol or ethanol (Dr ¼ 31.5, 24.39) increases the potential energy of the attraction for oppositely charged ions by a factor of 2.5 and 3.2, respectively. Ignoring specific changes in solvent effects, the data in Table 1 show the dramatic effect on the calculated binding constant according to log Kb ¼ (
DGH2O+DGROH)/ 2.303RT for a hypothetical process of Mx++Ay
Mx+:Ay
where the elec
1 trostatic potential energy of binding is 4.18–20.9 kJ mol in water, all of which is expressed in the free energy in passing to methanol and ethanol ignoring specific changes in solvent effects. While increasing the extent of binding of the reactants is one way to enhance the rate for a process such as that given in Equation (1), it is absolutely required that the association of the metal ion to the substrate must lower the activation energy for the acyl or phosphoryl transfer reaction, otherwise no increase in rate will be observed. Our earlier10 and more recent studies indicated that metal ion catalysis of the methanolyses of neutral carboxylate esters and activated amides16 and neutral phosphate, phosphorothioate, phosphonate and phosphonothioate esters17 were profoundly accelerated in methanol relative to water, the main accelerating effects of the solvent change being proposed as increased pre-association Mx+/CQO or Mx+/PQO binding, and a changed activity of the metal bound methoxide which is manifested in the actual acyl/phosphoryl transfer. Amis proposed Equation (3)18 for the reaction rate constant involving a limiting case of head-on approach of an ion to a neutral dipolar molecule from electrostatic considerations, as would be the case for metal ion catalysts and neutral CQO or PQO substrates. This expression relates the natural log of the rate constant in a medium of dielectric constant (D) to the charge on the ion (ZA), the dipole moment ln k0D¼D ¼ ln k0D¼1 þ

Z2 m DkTr2

(3)

of the molecule (m), the separation of the ion and head of the dipole (r), Boltzmann’s constant (k) and absolute temperature (T). It indicates that the rate constant of the

276

R.S. BROWN AND A.A. NEVEROV

reaction for an oppositely charged metal ion and negative dipole head will increase with a reduction in the dielectric constant of the medium. A more general, but complex, treatment was developed by Landskroener and Laidler19 describing the effect of dielectric constant on the velocities for ion–dipole, dipole–dipole and ion–ion reactions. The general approaches used for these treatments are highly simplified since they are really pertinent only to cases where electrostatic interactions on reactions are more important than specific solvation and other effects.18,19 Admittedly the catalytic processes generated by the association of the metal ions and neutral or anionic CQO or PQO ester substrates is probably more complicated than can be analyzed quantitatively through application of the simple form of Equation (3), but the general idea that a reduced dielectric constant does lead to a greater interaction energy for bringing the reactants together is well rooted in theory. Up to a certain point, larger Kb binding constants could be a prelude to enhanced rates of the subsequent chemical reactions since the ion-paired intermediate complexes now exist long enough for multiple intra-complex collisions from which a productive chemical reaction could occur. This will be particularly so for a metal-catalyzed reaction where there is charge neutralization of the substrate (as in the reaction of a negatively charged phosphate anion) in the rate-limiting transition state for the reaction. Of course, the attraction energy cannot be so great as to prevent separation of the catalyst and reaction product after the reaction has occurred, else wise true turnover cannot occur.

3

Titrations in alcohol

To conduct meaningful mechanistic and kinetic studies in alcohol media reliable and simple measurement and control of the solution ss pH is essential. Potentiometric titration is the method of choice for obtaining acid dissociation constants or metal ion complex stability constants and in favorable cases the speciation of mixtures of metal-ion-containing complexes in solution can be proposed.20 Titrations in nonaqueous solvents are not nearly as widely reported as those in aqueous media, particularly in cases with metal ions21 and determination of ‘pH’ in a non-aqueous solvent referenced to that solvent is complicated due to the lack of a way to relate the electrode EMF readings to absolute ss pH (see footnote * and ref. 6) so nonaqueous solvents are generally inconvenient to use22 for detailed studies of reaction mechanisms where pH control is required. The measurement of and determination of the ss pK a values for ionization of mono and dibasic acids in methanol6,7 and ethanol8 is now relatively straightforward. deLigny and Rehbach23 empirically determined a method for measuring the ss pH in methanol where one subtracts a correction constant of –2.34 (on the molality scale) from a measured electrode reading. Bosch and coworkers6 subsequently reported a method for determining ss pH on the molarity scale which, for our purposes, is relatively simple: if a glass electrode is calibrated using standard aqueous buffers but the potentiometric measurements are made in a non-aqueous solvent, the values are termed sw pH. Subsequently one computes ss pH ¼ sw pH – d, where d is a correction

METAL-CATALYZED ALCOHOLYSIS

277

factor of –2.24 on the molarity scale for measurements made as above in methanol.6 The measurement in ethanol is done analogously, although the correction factor and autoprotolysis constant are different from, and more uncertain than, those of methanol. Bates et al.24 and Grunwald et al.25 independently determined the d-correction constant which is related to the electrode junction potential between the two solvents (E¯ j ) and the primary medium effect for the solvent ðm gH Þ by the equation d ¼ E¯ j
m gH .26a The practical measurement of ss pH in ethanol is accomplished by subtracting this constant (d ¼
2.91 or
2.3626b on the molality scale for Bates24 and Grunwald25b) from the measured electrode reading (ss pH ¼ sw pH
d) provided that d is determined under ‘ideal conditions’24 of low ionic strength and an intermediate value of ss pH. The Bates and Grunwald d-correction factors differ substantially but there appear to be no criteria to determine which is better so we have chosen for our work to use the mean of these (d ¼
2.54).8 Details about standardization of electrodes and how to conduct the titrations in methanol and ethanol are published6,7,8 and need not be repeated here. Most of our analyses of the titration data (d[titrant]/dss pH) were conducted using the commercially available fitting program Hyperquad 2000 NT27 setting the respective autoprotolysis constants for methanol and ethanol at 10
16.77 and 10
19.1 at 25 1C as described.7,8 Data given in Table 2 are some ss pK a values of some monobasic acids and aminium ions that are useful as buffering agents or ligands which show general trends in the ionization constants in passing from water to methanol and then ethanol. The ss pK a s for carboxylic and phenolic (neutral) acids are significantly higher in ethanol than in water or even methanol, consistent with what is expected for the decreased dielectric constant. The ss pK a values for the cationic aminium ions do not vary greatly in passing from water to ethanol and, depending on the structure of the base, the acidity in ethanol can be higher or lower than in water. Since neutral s s pH in EtOH is 9.55 while in methanol and in water it is 8.39 and 7.0, a species having the same numerical ss pK a is considerably more acidic in ethanol. The trends in ss pK a in passing from water to the less polar alcohols can be explained in part by considering the equilibria for carboxylic acid and aminium ion dissociation shown in Equations (4) and (5). Carboxylic acid dissociation creates two opposite charges, while aminium ion dissociation simply relocates (H+) from the amine to the solvent as ROH+ 2 , which may be R1C(=O)OH + HOR

R1C(O)O− + +H2OR

(4)

R2R3R4NH+ + HOR

R2R3R4N: + +H2OR

(5)

differently solvated in a given medium than the aminium ion depending on the various alkyl groups. Specific solvation that stabilizes charge separated ions is expected to be poorer in less polar solvents than in water, thus accounting for the increased ss pK a s of carboxylic acids and the relative indifference in the ss pK a s of aminium ions. The solvation effects in play comprise more than the dielectric constant and include specific interactions like hydrogen bonding, lone-pair ion

278

R.S. BROWN AND A.A. NEVEROV

Table 2 ss pK a values of various monobasic acids and aminium ions as determined by Hyperquad analysis of titration data at 1  10
3 M in anhydrous ethanol, T ¼ 25.0 1C Acid Acetic acid Benzoic acid 2,4-Dinitrobenzoic acid 3,5-Dinitrobenzoic acid 4-Nitrophenol 2,6-Lutidine 1-Methylimidazole 4-Ethylmorpholine Triethylamine 1,10-Phenanthroline 2,9-Dimethyl-1,10-phenanthroline 1,5,9-Triazacyclododecane

pKa (water) 4.76b 4.21b 1.42c 2.82c 7.15d 6.72f 6.95d 7.76f 11.01d 4.86h,i 6.17h,i 7.49k,i 12.60l

s s pK a

(methanol)a

9.7770.10 9.1270.02 6.4870.03 7.1470.02 11.30e 6.86e 7.60g 8.28 10.78e 5.63j,i 6.43j,i 14.92j,l

s s pK a

(ethanol)a

10.5870.08 10.4070.08 7.2970.05 8.0170.05 11.7970.17 6.3770.06 7.5070.14 7.9870.04 10.2270.05 5.3670.23i 6.1570.01i 6.3270.06i 12.5470.05l

Source: Data from ref. 8. a Errors given are standard deviations of the mean of 3-5 determinations. b Isaacs, N. (1995). Physical Organic Chemistry (2nd edn), pp. 238–239. Pearson Education Ltd., Harlow, Great Britain. c Gluck, S.J., Steele, K.P. and Benko, M.H. (1996). J. Chromatog. A 745, 117. d Handbook of Chemistry and Physics, 48th edn. (1967–68). CRC Press, Cleveland, Ohio, pp. D88–89. e Rived, F., Rose´s, M. and Bosch, E. (1998). Anal. Chim. Acta 374, 309. f Andon, R.J.L., Cox, J.D. and Herington, E.F.G. (1954). Trans. Faraday Soc. 50, 918. g Neverov, A.A. and Brown, R.S. (2001). Inorg. Chem. 40, 3588. h Schilt, A.A. and Smith, G.F. (1956). J. Phys. Chem. 60, 1546. is s pK a 2. j Desloges, W., Neverov, A.A. and Brown, R.S. (2004). Inorg. Chem. 43, 6752. k Bell, T.W., Choi, H.-J., Harte, W. and Drew, M.G.B. (2003). J. Am. Chem. Soc. 126, 12196. ls s s pK a 3. May be a lower limit since s pK a is close to solvent ionization.

interactions, steric effects and enforced ion pairing that can have dramatic effects on the acidity and reactivity.

4

Metal ion alcoholysis and titration in alcohol

The titration of metal ions in alcohol solvents28 follows the same sort of rules as titrations of metal ions in water29 but poses additional problems due to the lower polarity that increases ion pairing and oligomerization of the metal ions. We have performed several such titrations with the analysis of the potentiometric data depending on the level of information one requires. More complete and timeconsuming analyses are reserved for the most effective catalytic metals, namely La3+, and for the transition metal ion Zn2+ and Cu2+ along with some simple complexes of the latter two which we describe a little later. For the other metal ions described in our titration papers,7,8 we only present the data in terms of the

METAL-CATALYZED ALCOHOLYSIS

279

apparent, or conditional, ss pKa values for formation of the metal-bound alkoxides at an [Mx+]total of 1 mM. The data give titration constants30 ss G01 ; ss G02 ; ss G03 . . . ; etc., that are defined as the conditional half-neutralization points determined for a solution containing the metal ion, and are simply interpreted as the solution ss pH at which the [
OR]/[Mx+]total ratio is 0.5, 1.5 and 2.5. The metal ions are generally introduced as triflate or perchlorate salts as these counterions are generally found to have the least propensity for ion pairing. In most cases, knowledge of the halfneutralization values simplifies setting the solution ss pH at a desired value due to the buffering capabilities of the Mx+q(
OR)y species present in solution. However, the conditional half-neutralization values reveal nothing of the intricate details of the acid/base behavior involved and the latter constants are complicated composites of various equilibria that include ion pairing between the Mx+ and solution counter ions (which we generally ignore) and metal dimerizations/oligomerizations that can affect the ionizations of (Mx+)n(HOR)m forms. These details can be worked out with considerable effort which requires computer fittings of the data obtained through titrations at different [Mx+] to various models comprising species for which there is corroborative evidence from other studies such as kinetics or spectroscopic ones. For other cases, such as La3+ where more detail is required about the nature of the species present in solution, titration data can be computer fit to more complicated multi-equilibrium models containing Mx+q(
OR)y forms whose stoichiometry is suggested by information gained from independent spectroscopic or kinetic techniques. One must be mindful of the pitfalls of simply fitting the potentiometric data to complex multi-component models for which there is no independent evidence for the various species. Without some evidence for the species put into the fit, the procedure simply becomes an uncritical mathematical exercise of adding and removing various real and proposed components until the goodness of fit is satisfactory. A case in point concerns the speciation of La3+ in methanol and ethanol which is of interest because of the very large rate enhancements for La3+-catalyzed acyl and phosphoryl transfer to alcohol solvents which will be described later. Earlier kinetic studies of the catalysis of methanolysis of carboxylate esters16a and the activated amide acetyl imidazole31 indicated that the process was largely bimolecular in each
of [La3+] and [
OCH3], suggesting that a dimeric species (La3+ 2 ( OCH3)2) was largely responsible for the great rate accelerations observed. Electrospray MS on solutions of La(OTf)3 in the presence of 1 equivalent of NaOCH3 also indicated the presence of dimeric La3+-containing species containing two and three methoxides as well as three and two, respectively, triflate counterions and varying numbers of solvent methanols.16a Finally the potentiometric titration data7 depended on the concentrations of the La(OTf)3, with higher concentrations driving the apparent s s pK a s to lower values, a phenomenon suggestive of the equilibrium formation of dimers or higher-order aggregates.29a Analysis of these potentiometric titration data according to the model suggested in Equations (6) and (7) using the program Hyperquad27 allows one to compute the various stability constants ðss K n Þ which

280

R.S. BROWN AND A.A. NEVEROV

Table 3 Stability constants for the formation of La2(OR)n species and microscopic ss pK a values in methanol and ethanol (25.0 1C) Equilibrium (Kneq)

log ss K neq a

[La2(
OR)1]/[La]2[
OR] [La2(
OR)2]/[La]2[
OR]2 [La2(
OR)3]/[La]2[
OR]3 [La2(
OR)4]/[La]2[
OR]4 [La2(
OR)5]/[La]2[
OR]5

11.6670.04c 14.8570.06d 20.8670.07c 27.3670.08d 27.5270.09c 38.7770.22d 34.5670.20c 49.8470.16d 39.3270.26c

Microscopic ss pK a b

2 s s pK a

¼ 7.5670.10c ¼ 6.6070.03d 3 c s s pK a ¼ 10.1170.02 ¼ 7.6970.15d 4 c s s pK a ¼ 9.7370.11 d ¼ 8.0370.22 5 c s s pK a ¼ 12.0070.07

a

Derived from fits of the potentiometric titration data in methanol and in ethanol using the program Hyperquad.27 b Defined as
log ss pK a for La2(OR)n(HOR)x La2(OR)n+1(HOR)y
1+H+, calculated from appropriate data in column 2 as ss pK auto (autoprotolysis constant of solvent) – ss pK neq
ss pK n
1 eq where 16.77 is the ss pK auto for pure methanol and 19.1 is the ss pK auto for ethanol. c Computed constants for [La(OTf)3] ¼ 2  10
3 M in methanol, from ref. 7 d Computed constants for [La(OTf)3] ¼ 1  10
3 M in ethanol, from ref. 8.

are given in Table 3 along with the analogous stability constants subsequently determined8 for the − La3+ 2 ( OCH3)n s n s K eq

2La3+ + nOCH3−



n 3þ 2 ¼ ½La3þ 2 ð OCH3 Þn =½La  ½OCH3 

(6) (7)


formation of La3+ 2 ( OEt)1,2,3,4,5 in ethanol. From these can be derived the individ
ual microscopic ss pK a values for the ionizations La3+ 2 ( OR)n(HOR) 3+
+ La2 ( OR)n+1+H : these too are listed in Table 3. From the formation constants derived by fitting the potentiometric titration data
as above can be computed the speciation for the Ln3+ 2 ( OR)1
4 forms by HySS, 27 another program in the Hyperquad suite. The speciation diagram for La3+ at 2  10
3 M in ethanol is shown in Fig. 1. According to this treatment the two dominant dimeric La3+ species have even numbers of attached ethoxides, namely
s La3+ 2 ( OEt)2 between s pH 5.9 and 8.4 (maximum concentration of 48% relative to 3+
s s [La2 ]total at s pH 7.3), and La3+ 2 ( OEt)4 above s pH 7. Species with odd numbers of 3+
3+
ethoxides, La2 ( OEt)1 and La2 ( OEt)3, are also present to a lesser extent (maximum concentrations of 17% and 37% respectively reached at ss pH values of 6.7 and 7.8). The computed speciation for La3+ in ethanol is similar to the calculated speciation diagram shown in Fig. 2 for 2.0  10
3 M La3+ in methanol as a function of a more extended ss pH range.7 The two dominant species again have even numbers
3+
s of attached methoxides, e.g. La3+ 2 ( OMe)2 (80% at s pH 8.9), and La2 ( OMe)4 3+
3+
s (80% at s pH 11). La2 ( OCH3)1 and La2 ( OCH3)3, are also present to a lesser extent (25% in each at respective ss pH values of 7.5 and 10).

METAL-CATALYZED ALCOHOLYSIS

281

0.0006 La2(OEt)4 0.75

La2(OEt)2

0.0004

La2(OEt)3 0.50

0.0003

0.0002

La2(OEt)1

k2obs (dm3mol-1s-1)

[(La3+)2(-OEt)n] (M)

0.0005

0.25

0.0001

0.0000

0.00 5

6

7

8

9

s s pH – Fig. 1 A calculated speciation diagram for La3+ n ¼ 1–4, total 2 ( OEt)n, [La3+] ¼ 2  10
3 M, Overlaid are the observed second-order rate constants (kobs 2 ) for the La3+-catalyzed ethanolysis of paraoxon (1) as a function of ss pH. Reprinted with permission from ref. 8.

0.0010 La2(OCH3)2

La2(OCH3)4

40

30 0.0006

20 0.0004 La2(OCH3)1

La2(OCH3)3

k2obs (dm3mol-1s-1)

[La3+2(-OCH3)n] (M)

0.0008

10

0.0002

0.0000

0 5

6

7

8

9

10

11

12

s s pH

Fig. 2 Speciation diagram for 2.0  10
3 M La3+ in methanol as a function of ss pH. Data superimposed on the figure as () are second-order rate constants for La3+-catalyzed methanolysis of p-nitrophenyl acetate (2) as a function of ss pH.

282

R.S. BROWN AND A.A. NEVEROV

Having established the speciation, we now have a very powerful tool for analyzing the kinetic data for the ss pH dependence of the La3+catalysis of the alcoholysis of various substrates. Included in the Figs 1 and 2 plots are the second-order rate constants for La3+-catalysis of the ethanolysis of paraoxon (1) and the methanolysis of p-nitrophenyl acetate (PNPA, 2) as a function of ss pH in ethanol and methanol,
respectively. The kinetic data mainly follow the rise/fall behavior of the La3+ 2 ( OR)2 3+
3+
species with some involvement of the other species, La2 ( OR)1, La2 ( OR)3 and
La3+ 2 ( OR)4. O EtO P O OEt 1

O NO2

H3C

O EtO P S NO2 OEt

O 2

O NO2

NO2

N

3

4


obs To determine the activities for the various La3+ data 2 ( OR)n we analyze the k2 as a linear combination of individual rate constants (Equation 8), where k2:1 2 , 2:n 3+
k2:2 y k are the second-order rate constants for each La ( OR) promoting 2 2 n 2 ethanolysis and methanolysis of 1 and 2 respectively.

2:1 2:2 3þ
3þ
kobs 2 ¼ ðk2 ½La2 ð ORÞ1  þ k 2 ½La2 ð ORÞ2  þ . . . 3þ
k2:n 2 ½La2 ð ORÞn Þ=½LaðOTfÞ3 t

ð8Þ 8

When fit to this expression, the kinetic data for the ethanolysis of 1 yields the contributions of the relevant La3+-dimers shown in Fig. 3 where the solid line represents the sum of the contributions and the three dotted and dashed lines give the contributions of the individual dimers at each ss pH. The individual species rate constants are determined as 0.5970.05, 2.2470.19 and 1.1170.09 dm3 mol
1 s
1 2:2 2:3 for k2:1 2 , k2 and k2 . 3+ Fits for the La -catalyzed methanolysis of p-nitrophenyl acetate (2),7 paraoxon (1), its thio derivative, O,O-Diethyl S-p-nitrophenyl phosphorothioate (3)10b and N-p-nitrophenyl b-lactam (4)32 have also been computed, with the various k2:n 2 the values being given in Table 4. The catalysis afforded by the La3+ system for the transesterifications of paraoxon in ethanol and methanol is quite spectacular relative to the background reactions that are assumed to be promoted by the lyoxide. The reaction rate constant of ethoxide with paraoxon in ethanol at 5.1  10
3 dm3 mol
1 s
133 is roughly a factor of two lower than the rate constant of methoxide with paraoxon in methanol (1.1  10
2 dm3 mol
1 s
1).17a However a solution 2 mmol dm
3 in total [La3+],
4
1 which contains 1 mmol dm
3 of La3+ s 2 , has a maximum rate constant of 7  10 s for decomposition of 1 in ethanol at s pH of 7.3, and accelerates the rate of ethanolysis of paraoxon by a factor of 4.4  1011-fold relative to the ethoxide reaction at the same ss pH.34 By way of comparison, the acceleration afforded by a 1 mmol dm
3 solution of the La3+ dimer catalyzing the methanolysis of 1 at the 2 maximal ss pH of 8.3 (kobs ¼ 0.0175 s
1) is 109-fold greater than its background methoxide reaction. On this simple basis La3+ in ethanol appears to be catalytically 2 superior to La3+ in methanol, but this stems almost exclusively from the ss pH values 2

METAL-CATALYZED ALCOHOLYSIS

283

0.8 0.7

k2obs (dm3mol-1s-1)

0.6 0.5 0.4 0.3 0.2 0.1 0.0 5

6

7

8

9

s s pH
3+ obs Fig. 3 Plot of the contributions of various La3+ -catalyzed 2 ( OEt)n forms to the k2 for La ethanolysis of paraoxon (1) as a function of ss pH. Solid line gives the combined effects of – various species; (  —  —), contribution of La3+ 2 ( OEt)1; (- - - - - - -), contribution of obs – 3+ – La3+ 2 ( OEt)2; (— — — ), contribution of La2 ( OEt)3; included as (’) are the actual k2 kinetic data. Reprinted with permission from ref. 8.

3+
Table 4 Computed second-order rate constants (k2:n 2 ) for the methoxide and La2 ( OCH3)na species catalyzing methanolysis of 1–4, T ¼ 25 1C

k2/dm3 mol
1 s
1 paraoxon (1)a

k2/dm3 mol
1 s
1 p-nitrophenyl acetate (2)b

k2/dm3 mol
1 s
1 phosphorothioate (3)

k2/dm3 mol
1 s
1 N-p-nitrophenyl b-lactam (4)

kOCH3 1 ¼ 0.011

kOCH3 2 ¼ 2207100c

kOCH3 3 ¼ 0.12

kOCH3 4 ¼ 0.43

k2:1 2 ¼ 15.872.9

d k2:1 2 NA

k2:1 2 ¼ 17.978.7

k2:1 ¼ 0.0670.01 2

k2:2 2 k2:3 2 k2:4 2

k2:2 2 k2:3 2 k2:4 2

k2:2 2 k2:3 2 k2:4 2

¼ 27.971.4

k2:2 2 ¼ 0.9870.11

NAb

k2:3 2 ¼ 1.2370.13

¼ 16.271.8

d k2:4 2 NA

¼ 51.171.1 ¼ 35.676.5 ¼ 49.771.4

¼ 91.878.8 NAd ¼ 38.075.6

a

Data from Ref. 10b. kOCH3 is observed rate constant for the attack of CH3O
on indicated substrate. c At 27.4 1C; Schowen, R.L. and Behn, C.G. (1968). J. Am. Chem. Soc. 90, 5839. d Indicated rate constant not required to obtain satisfactory fits to data. b

where the comparison background reactions are made and the far lower [
OEt] relative to [
OMe]. However, it is clear that in terms of the second-order rate constants the methanolysis reaction is more effective by about 25-fold at the maximum ss pH values for 2:2 the La3+ 2 -catalyzed reactions having the same amount of catalyst. For example k2

284

R.S. BROWN AND A.A. NEVEROV

for (La3+(
OCH3))2 promoted methanolysis of paraoxon is 51.1 dm3 mol
1s
1, while that for (La3+(
OCH2CH3))2 promoted ethanolysis is 2.24 dm3 mol
1s
1. While more study is required to determine the reasons why (La3+(
OCH3))2 is more active than (La3+(
OCH2CH3))2 this probably stems from the fact that coordination to the La2+ cations stabilizes the alkoxide better in ethanol than in methanol, as judged by ðss pK EtOH
ss pK MeOH Þ and a steric effect that may retard the catalytic a a reaction with the latter larger complex. It is important to note that the stabilization of the alkoxides through coordination to two La3+-ions does not render them less reactive than free alkoxide as might be expected from a simple Brønsted relationship. This supports the oft-stated premise that there is a dual role for the Mx+(
OR) species involving Lewis acid activation and delivery of the coordinated alkoxide. While this is rarely observed with Mx+(
OH) species in water, our work shows that this is the norm in alcohols.

5 Transition metal ion and Ln3+ catalysts of transesterifications of neutral carboxylate and organophosphate esters As summarized previously,10 virtually all the transition metal and lanthanide ions we have investigated have some catalytic activity in promoting the transesterification of carboxylate esters and neutral phosphate esters such as paraoxon (1). The active forms of these are routinely generated in situ by adding 1 equivalent of NaOR to an alcohol solution containing 1 equivalent of the metal ion, which gives predominantly species having a net stoichiometry of (Mx+(
OR))1,2.35 For La3+ there is evidence that the most active form for methanolysis and ethanolysis is
3+
the dimer La3+ 2 ( OR)2. The other forms La2 ( OR)n where n ¼ 1, 3, 4 may also 3+
have activity (Table 4), but the La2 ( OR)2 form has the highest activity in the neutral ss pH domains due to its appreciable second-order rate constant and its high prevalence. For other metal ions there are some interesting variations where the Mx+(
OR) monomeric form is the most active, while the dimeric forms are inactive. For example, the Eu3+-catalyzed methanolysis of the esters p-nitrophenyl acetate (2), phenyl acetate (5) and ethyl acetate has been studied as a function of ss pH along with its titration in methanol.16e All the kinetic evidence indicates that the solution behavior is consistent with the species shown in Equations (9–11), with Eu3+(
OCH3)(CH3OH)x
1 being the reactive one (ss pK 1a ¼ 6.33) having a kmax rate 2 constant of 4272 dm3 mol
1 s
1 for methanolysis of 2 and 11.771.5 dm3 mol
1 s
1 for methanolysis of 5. The ss pH rate profile for the Eu3+-catalyzed methanolysis of 2 and 5 exhibit bell-shaped profiles with a second apparent ss pK 2a of 8.02 for conversion of Eu3+(
OCH3) into inactive Eu3+(
OCH3)2 which we believe exists as a heavily methoxy-bridged oligomer of repeating Eu3+(
OCH3)2 subunits as in 6. O H3C

O 5

METAL-CATALYZED ALCOHOLYSIS *

O R

Eu

285

R

R

R

R

O

O

O

O

O

Eu

R

O R

Eu

O

Eu

R

S H3CO P O H3CO 7

x*

NO2 CH3

6 (charges ommitted for clarity)

Ka1

3+

Eu (CH3OH)x

3+

−

Eu3+(CH3O− )(CH3OH)x−1 + H+

Ka2

Eu (CH3O )(CH3OH)x−1 nEu3+(CH3O− )2(CH3OH)x−2

Kolig

Eu3+(CH3O−)2(CH3OH)x−2 + H+

(9)

(10)

[Eu3+(CH3O− )2(CH3OH)y]n + n(CH3OH)(x−2−y)

(11) 2+36

2+37

and Cu there is evidence that the monomeric On the other hand, for Zn forms (M2+(
OR)) are catalytically active toward transesterification of carboxylate esters like 2 and paraoxon (1). The situation is more complicated because the maximum rate constants for the reactions are achieved when the [
OR]/[M2+]total ratio is 0.1–0.3, and then starts to fall at higher ratios. At a given [
OR]/[M2+]total ratio the plot of kobs vs. [M2+]total follows a square root dependence suggestive of a catalytically active monomer in equilibrium with an inactive dimer. The minimal process required to fit the kobs vs. [
OCH3]/[Zn2+]total and kobs vs. [M2+]total behavior for Zn2+-catalyzed methanolysis is shown in Scheme 1. The apparent dimerization constant (Kd) is a complex term that incorporates both dimerization of Zn2+(
OCH3) and an ionization constant that connects Zn2+(
OCH3)2 and (Zn2+(
OCH3)2)2. The key point of this analysis is the requirement to limit the overall [Zn2+(
OCH3)] through the formation of inactive dimeric species (Zn2+(
OCH3))2 and (Zn2+(
OCH3)2)2. Unfortunately, the dimerization constants for the 2(M2+(
OR)) (M2+(
OR))2+(M2+(
OR)2)2 process are large (4 2  105 dm3 mol
1) so the catalysis observed is not very effective even though the monomer is quite active. Given in Table 5 are the second-order rate constants for the Zn2+(
OCH3)- and Cu2+(
OCH3)-catalyzed methanolysis of paraoxon (1) and a closely related phosphorothionate pesticide fenitrothion (7) that are computed from the kobs vs. [M2+(
OCH3)] data.36,37 The complexity exhibited by these transition metal ions can be reduced through the addition of complexing ligands such as phenanthroline (8) and 1,5,9-triazacyclododecane (9), a macrocyclic ligand, the Zn2+-complexes of which have been extensively investigated in water by Kimura.38 Detailed studies36,37 indicate that phenanthroline, when complexed to either Zn2+ or Cu2+ in the presence of 1 equivalent of added –OCH3, produces a kinetically active complex (8:M2+(
OMe)) that still has an appreciable tendency to dimerize to an inactive dimer (10) as in

286

R.S. BROWN AND A.A. NEVEROV Ka1

Zn2+(HOCH3)n

Zn2+(HOCH3)n-1(-OCH3) + H+ [Zn2+(HOCH3)n-1(-OCH3)]2

2 Zn2+(HOCH3)n-1(-OCH3) Kd

[Zn2+(HOCH3)n-1(-OCH3)2]2 + 2H+

Scheme 1 A minimal scheme depicting ionization of a Zn2+(HOCH3)n species followed by formation of inactive Zn2+-dimers.

Table 5 Kinetic constants for the methanolysis of paraoxon 1 and fenitrothion 7 catalyzed by Zn2+ and Cu2+ in the absence and presence of ligands 8 and 9, T ¼ 25 1C Catalyst
OCH3 Cu2+(
OCH3)c Zn2+(
OCH3)d 8:Cu2+:(
OCH3)e 8:Zn2+:(
OCH3)f 9:Cu2+:(
OCH3)g 9:Zn2+:(
OCH3)h
i La3+ 2 ( OCH3)2

Kd (dm3 mol
1)a

k2(1) (dm3 mol
1 s
1)a

k2 (7) (dm3 mol
1 s
1)a

Relative selectivityb

NA 42  105 42  105 42  105 42  105 NA NA NA

1.1  10
2 0.2270U02 1.2170.03 o0.2 2.0770.04 2.7670.17 0.8570.01 47.272.3

(7.270.2)  10
4 0.7970.03 0.1970.001 2.4470.06 0.3270.01 12.270.4 (4.870.2)  10
2 Non-reactive

1 55 2.4 186 2.35 67 0.86 0

a Dimer association constant (Kd) and conditional second-order rate constant (k2(1) or k2(7)) for reaction of monomer with 1 or 7 defined as in text. NA means non-applicable since there is no observable dimerization under the specific conditions. The Kd of 42  105 indicates very strong dimerization and is quoted as an upper limit based on an iterative fitting procedure which provided the lowest standard deviations. Zn2+ results from ref. 36, Cu2+ results from ref. 37. b Defined as (k2(7)/kOCH3 (7))/(k2(1)/kOCH3 (1)). c Based on fits of kobs vs. [Cu2+]total data at [methoxide]/[Cu2+]total ratio of 0.5. d Based on NLLSQ fits of kobs vs. [Zn2+]total data at [methoxide]/[Zn2+]total ratio of 0.3. e Based on fits of kobs vs. [8:Cu2+]total data at [methoxide]/[Cu2+]total ratio of 0.5. f Based on NLLSQ fits of kobs vs. [Zn2+:9]total data at [methoxide]/[Zn2+]total ratio of 0.5. g Based on linear fits of kobs vs. [9:Cu2+:(
OCH3)]total data at methoxide]/[Cu2+]total ratio of 0.5; ss pK a for ionization of 9:Cu2+ to generate 9:Cu2+(
OCH3) is 8.75 70U1 by half neutralization. h Based on linear fits of kobs vs. [9:Zn2+:(
OCH3)]total data at [methoxide]/[Zn2+]total ¼ 1.0. The first and second ss pK a values of 9.1 and 12.9 for 9:Zn2+ to generate 9:Zn2+ (
OCH3) and 9:Zn2+(
OCH3)2. i From ref. 17a.

Equation (12). However, in the case of complexation to 9 there is no dimerization and the catalytically active form is 9:M2+(
OR). H N N

N 8

9

N N H

H

METAL-CATALYZED ALCOHOLYSIS

287

CH3 -O N Zn2+ N

OH3C

N Zn2+ N

CH3 -O k21 = 2.1 dm3mol-1s-1 N P Zn2+

2

N

5

Kd = >2x10 dm3mol-1

8:Zn2+(-OMe)

10 S

CH3

P

N PdII N CH3 CH OSO2CF3 3 11

(12)

O CH3 H

H2C O O O H3C H2C CH3 12

N N

S

N

P H2C O O N O H3C H2C CH3 13

Also included in Table 5 are the kinetic data for methanolysis of 1 and 7 catalyzed
2+
by the La3+ ( OCH3) and 9:M2+(
OCH3) complexes and, for 2 ( OCH3)2, 8:M comparison purposes, the second-order rate constants for methoxide-catalyzed methanolysis. The final column is a ‘relative selectivity (RS) parameter’ that describes the metal ions’ ability to select for a PQS species over the PQO species, defined as (k2 (7)/kOCH3(7))/(k2(1)/kOCH3(1)). Of note is the fact that all the Cu2+-containing species have RS values of 55–186, meaning they strongly prefer the P ¼ S substrate but they will also catalyze methanolysis of the P ¼ O substrate. The Zn2+ complexes are less selective with RS values from 0.9 to 2.5 meaning they do not really discriminate well between the PQO and PQS classes. On the other hand, La3+ has a zero RS value and shows no propensity to react with the PQS substrates at all. In fact, none of the lanthanides we have studied promote the methanolysis or ethanolysis of the PQS derivatives, presumably due to the fact that these are considered ‘hard’ in the Pearson ‘hard/soft’ sense39 while the softer Cu2+ ion selects for the soft PQS derivatives. We have observed that cyclopalladated Pd complexes such as 11 rapidly methanolyze PQS pesticides like 7, diazinon (12) and quinalphos (13) but do not react with the PQO complexes nearly as fast which is also consistent with the ‘soft’ behavior of the Pd.40 These observations are also important from a mechanistic standpoint since they underscore the importance of transient metal ion:substrate binding which preceeds the catalytic events, and in cases where the hard/soft properties of the reaction partners do not match, one can safely predict little or no catalysis will be observed. While the above indicates that Ln3+ and transition metal ions in the presence of at least 1 equi. of –OR promote the alcoholysis of carboxylate and phosphate esters, sometimes by spectacular amounts, we have not presented evidence about the mechanism for the catalytic reactions. So far, the underlying theme is that the most active forms of the lanthanide ions are the Ln3+(
OR) forms, either as a monomer
(as in the case of Eu3+(
OCH3)) or as a dimer (as in the case of La3+ 2 ( OCH3)2). For the transition metal ions the most active forms are those where one face of the

288

R.S. BROWN AND A.A. NEVEROV

metal ion is encapsulated with ligands such as phenanthroline (8) or 1,5,9-triazacyclododecane (9) along with a single alkoxide. MECHANISM OF ALCOHOLYSIS OF CARBOXYLATE ESTERS

The initial study of the La3+-catalyzed methanolysis of carboxylate esters16a re
ported the apparent second-order rate constant for La3+ 2 ( OCH3)2-catalyzed methanolysis of some representative examples of aryl esters (2, 5 and 2,4-dinitrophenyl acetate (14)), phenyl benzoate (15) and three aliphatic esters, ethyl acetate, isopropyl acetate (16) and tert-butyl acetate (17). Given in Table 6 are the rate constants for the La3+ and methoxide-catalyzed methanolysis of these esters along with O

O H3C

NO2 O

14 O2N

O O

H3C

15

CH3 H3C O CH CH3 16

O

CH3 C CH3 O CH3 17

the catalytic rate acceleration relative to the background reaction at an essentially neutral ss pH of 8.5. There are a few features of note, some of which distinguish the methanolysis results from hydrolytic processes. First, simple metal ion-catalyzed hydrolysis of esters, including non-activated ones such as we have here is not, to our knowledge, a well-known phenomenon. Metal ion catalysis of hydrolysis is seen for esters having good leaving groups and particularly where there is an auxiliary binding site which places the metal ion in close proximity to the scissile CQO(LG) unit. Table 6 Maximal second-order rate constants for (La3+)2(CH3O
)2-catalyzed methanolysis and second-order rate constants for methoxide attack on various esters, T ¼ 25 1C Ester 2 14 5 15 Ethyl acetate 16 17

3
1
1 c kester (dm3 mol
1 s
1)a,b kester Accelerationg s ) kester =kester 2 OMe
(dm mol 2 OMe


72 2972 5873 3.070.06 0.143 0.0083 55.7  10
6

2207100c 410d 2.66e 0.33870.006f (5.770.2)  10
2 (7.070.1)  10
3 (2.070.1)  10
4

f f f

0.6–0.225 0.07 21 8.9 2.5 1.2 50.03

243,000 42,000 18,750,000 8,264,000 2,300,000 1,120,000 n.o.

a Methanolysis rates for esters 2, 14, 5, 15 determined by UV kinetics in CH3OH; ethyl acetate, 16, 17 determined by 1H NMR in d4-methanol. bs s pK a set at observed rate plateau for methanolysis of the esters between 8 and 9 through addition of 1:1 La3+/
OCH3. c 27.4 oC; Schowen, R.L. and Behn, C.G. (1968). J. Am. Chem. Soc. 90, 5839. d Machacek, V., Mareckova, S. and Vojeslav, S. (1979). Collect. Czech. Chem. Commun. 44, 1779. e Milton, C.G., Gresser, M. and Schowen, R.L. (1969). J. Am. Chem. Soc. 91, 2047. f Ref. 16a. g Catalytic value at ss pH 8.5, in the presence of 5  10
3 mol dm
3 (La3+)2(CH3O
)2. Background rate s computed from kester OMe
assuming value for solvent reaction at s pH 8.5 is entirely attributed to CH3O
+ester; the autoprotolysis constant of methanol is 10
16.77.hAcceleration computed at ss pH 8.7.

METAL-CATALYZED ALCOHOLYSIS H3C

O

H3C

-

O La3+

La3+ -O

289

+ R

R'

O

O

- O CH

-

O

La3+

La3+

H3C

CH3

O

O R'

3

R

La3+ + HOCH3 3+ O La + -H O -RC=O(OCH3) H3C R + HOCH 3

-HOR' -RC=O(OCH3) H3C

O

La3+

La3+

(poor LG) O-OR' O CH3

O R' R

-OR' H3C

O

(good LG)

La3+ O

H3C

O

La3+

La3+

O-

La3+

H3C

O -

O R'

H3C

O

O R'

R

R

Scheme 2 A proposed mechanism for La3+ (
OCH3)2 catalyzed transesterification of 2 carboxylate esters with good and poor leaving groups.

Second, for the three aryl esters (2, 5, 14) the greatest acceleration relative to the methoxide reaction is seen for the poorest leaving group (phenoxy 4 p-nitrophenoxy 4 2,4-dinitrophenoxy). Third, for the aliphatic esters (ethyl acetate, 16, 17) there is a large steric factor, perhaps superimposed on an electronic one which retards the metal-catalyzed reaction more than the methoxide-catalyzed reaction for the isopropyl and tert-butyl acetates. A proposed mechanism for the reaction is given in
Scheme 2 where the catalytic entity (La3+ 2 ( OCH3)2) is shown as a bis-methoxy bridged dimer. According to the scheme, the CQO unit associates with the metal ion in a typical Lewis fashion for which the equilibrium constant should be greater in alcohol relative to water due to the lower dielectric constant. Since a methoxide bridged between the two metal ions should have limited nucleophilicity, we suggest that one of the methoxy bridges opens to reveal a strong Lewis acid activated CQO unit nearby a methoxide coordinated to the proximal La3+(
OCH3) that subsequently generates a La3+-coordinated tetrahedral intermediate. In the case of good leaving groups the tetrahedral intermediate breaks down probably without assistance via coordination to the metal ion. However for poor leaving groups such as the aliphatic ones, microscopic reversibility suggests that if the formation of the tetrahedral intermediate results from attack of a metal-coordinated methoxy or ethoxy group, then its breakdown must proceed via metal ion assistance of OR0 departure. If the leaving group is very sterically demanding, as well as having a higher ss pK a for its conjugate acid such as in the case of tert-butoxy, then metal ion assistance of departure might be severely retarded to the point that the tetrahedral intermediate cannot break down productively and catalysis will be strongly reduced. The question of whether the reaction proceeds through an intermediate or whether it is a concerted displacement such as has been demonstrated by Ba-Saif and Williams41 for the displacement of the p-nitrophenoxy group from ester 2 (18d) by

290

R.S. BROWN AND A.A. NEVEROV

Table 7 Second-order rate constants for the methanolysis of esters 18, 19 promoted by
methoxide, 9:Zn2+(
OCH3) and La3+ 2 ( OCH3)2 at 25 1C Ester

18a(14) 18b 18c 18d(2) 18e 18f 18g(15) 18h 18i 19a 19b 19c (16) Methyl acetate

s s pK a

of ArOH in MeOH

kOCH3 (dm3 mol
1 s
1)

7.83b 8.84a 12.41b 11.30b 13.59b 14.7c 14.33b 15.04b 15.36b 15.78d 18.42d 19.79d 18.13

9:ZnðOCH Þ

9:ZnðOCH Þ

3 k cat (dm3 mol
1 s
1)h

3 k cat (dm3 mol
1 s
1)i

43.6 23.8 39.8 19.2 16.7 8.6 13.9 4.28 4.30 3.00 0.042 0.006 –

31 14.6 21.9 38.6g 27.7 18.0 29.33g – 4.33 10.0 0.071g 0.004g –

484.8 73.3 49.7 190e 7.04 1.58 2.01 (2.66)f 0.63 0.49 12.64 0.057g 0.007g 0.17e

a

Determined as the ss pH at half neutralization, this work. From refs. 6 and 42. pKa for 18i calculated according to the method provided in ref. (6a); MeOH s 2 O þ 3:56. ¼ 1:12pK H a s pK a c Schowen, R.L. and Lathan, K.S. (1967). J. Am. Chem. Soc. 89, 4677. d 2 O +6.53. Computed from ss pK MeOH ¼ 0.748pK H a a e From ref. 36. f From Milton, C.G., Gresser, M. and Schowen, R.L. (1969). J. Am. Chem. Soc. 91, 2047. g Data from ref. 16a for methoxide and lanthanum reactions; lanthanum reactions determined at pH ¼ 8.86 for 18d (2) and 18g (15); values are computed from the gradient of the plot of kobs vs. [La3+]total. h Computed from the gradient of kobs vs. [9:Zn2+(
OCH3)] plots at ss pH 9.1 under half-neutralization conditions. i Computed from the gradient of kobs vs. [La3+] plots at ss pH 8.74 under buffered conditions. b

phenoxy and probably alkoxy nucleophiles has been addressed16f in the study of the
9:Zn2+(
OCH3) and La3+ 2 ( OCH3)2-catalyzed methanolysis of an extensive series of carboxylate esters, 18 and 19. For the series of esters, the second-order rate constants for the reactions are presented in Table 7 along with the ss pK a values for the ArOH or HOR groups in methanol as reported in the literature6,42 or as determined by means identified in the study.16f From the autoprotolysis constant of methanol (10
16.77 (mol dm
3)2) the ss pK a of MeOH can be computed as 18.13 on the mol dm
3 scale. O H3C

C

O O Ar

18 18a, Ar=2,4-dinitro 18b, Ar=pentafluoro 18c, Ar=3-nitro 18d, Ar=4-nitro 18e, Ar=4-chloro 18f, Ar=4-methoxy 18g, Ar=phenyl 18h, Ar=2,4-dimethyl 18i, Ar=2,3,5-trimethyl

H3C

C

O R

19 19a, R= CH2CF3 19b, R=CH2CH3 19c, R=CH(CH3)2

METAL-CATALYZED ALCOHOLYSIS

291

kcat9:Zn(OCH3)(dm3mol-1s-1)

100 10 1 0.1 0.01 0.001 7.5

10.0 12.5 15.0 17.5 pKa of ROH in methanol

20.0

9:ZnðOCH Þ

3 Fig. 4 A Brønsted plot of log pKa phenol in methanol vs. log kcat for methanolysis of aryl acetates promoted by 9:Zn2+(
OCH3), T ¼ 251C, data in Table 7. Dashed line corresponds to NLLSQ fit of data to Equation (15) encompassing all esters with b1 ¼
0.02370.03 and b2 ¼
0.69070.005 with a breakpoint of pKROH ¼ 14.8. Reproa duced from ref. 16f with permission.

kcatLa(OCH3)(dm3mol-1s-1)

1000 100 10 1 0.1 0.01 0.001 7.5

10.0 12.5 15.0 17.5 pKa of ROH in methanol

20.0

LaðOCH Þ

3 Fig. 5 A Brønsted plot of log pKa phenol in methanol vs. log kcat for methanolysis of 3+
aryl acetates promoted by La ( OCH3), T ¼ 251C, data from Table 7. Dashed line corresponds to NLLSQ fit of data to Equation (15) encompasses all esters with values of b1 ¼ 0.03 70.005 and b2 ¼
0.71570.005 with a breakpoint of pKa ¼ 14.7. Reproduced from ref. 16f with permission.

The Brønsted plots shown in Figs 4 and 5 assist in visualizing the data of the metal-catalyzed reactions which show definite evidence of a break where the rate constants are quite sensitive to phenols or alcohols with high ss pK a values, but almost no sensitivity to the nature of the ROH groups with low ss pK a values. In these cases there is no obvious discrepancy between aliphatic and aryl esters, so all

292

R.S. BROWN AND A.A. NEVEROV

esters are used for the subsequent treatment which assumes that the overall metalcatalyzed processes follow the processes in Equations (13) and (14) involving a fast pre-equilibrium binding followed by reversible creation of an intermediate, the formation and breakdown of which can be rate-limiting depending on the nature of the leaving group. The general kinetic relationship in Equation (15) is derived from a steady state treatment where the

Kb X+ −

M ( OCH3) + RCO2R RCO2R':MX+(− OCH3)

(13)

RCO2R':MX+(− OCH3)

'

k1 k −1

(CH3O-To−):MX+

k2 P

(14)

pKa refers to the ss pK a of the ROH or ArOH leaving groups in methanol and the b terms refer to the binding and kinetic steps. Equations (16) and (17) are derived for limiting cases where formation and breakdown of the tetrahedral addition intermediate ((CH3 O
TO
):Mx+) is rate-limiting. ðbbþb1þb2ÞpKa =ðC
1 10b
1pKa þ C 2 10b2pKa Þ kobs 2 ¼ K b k1 k2 =ðk
1 þ k 2 Þ ¼ C b C 1 C 2 10

(15) ðbbþb1ÞpKa kobs 2 ¼ K b k1 ¼ C b C 1 10

(16)

ðbbþb1
b
1þb2ÞpKa kobs 2 ¼ K b k1 k2 =k
1 ¼ ðC b C 1 C 2 =C
1 Þ10

(17)

The dashed lines in Figs 4 and 5 are computed from the NLLSQ fits of all the data to Equation (15): in Fig. 4 (bb+b1) ¼
0.023 and (bb+b1
b
1+b2) ¼
0.710 with a breakpoint at ss pK a ¼ 14.8; in Fig. 5 (bb+b1) ¼
0.02 and (bb+b1
b
1+b2) ¼
0.710 with a breakpoint at ss pK a ¼ 14.7. Overall the leaving group acidity ranges over 1011-fold while the kinetic data span respective ranges of 5000- and 9000-fold for the two metal ions. That each plot exhibits a pronounced break is consistent with a process that contains at least two steps where formation and breakdown of an intermediate is rate-limiting. In both the cases the descending and plateau b-values are experimentally the same with a plateau range that covers roughly 106-fold change in acidity of the leaving group with virtually no change in catalyzed rate constant. The data in Figs 4 and 5 are analyzed according to the simplified process presented in Scheme 3 where the ester substrate binds reversibly to the metal complex
(9:Zn2+(
OCH3) or La3+ 2 ( OCH3)2) followed by an intramolecular addition of the metal-coordinated methoxide to the bound CQO unit. Unlike the non-metalcatalyzed process of methoxide addition to the esters which must form a highly unstable, and perhaps kinetically unstable (in the case of aryl esters)41 anionic intermediate (TO
), the Mx+(
OCH3)-catalyzed process forms an Mx+-stabilized tetrahedral intermediate (20:TO
: Mx+) with a significant lifetime that partitions between starting materials and products depending on the relative transition state

METAL-CATALYZED ALCOHOLYSIS

Ester

Kb

+ 2+ -

9:Zn ( OCH3)

R O

9:2+Zn

293

-

Zn2+:9 k2 O

k O C CH3 1 H3CO H3CO O OR OR k H3C -1 H3C 9:Zn2+(-OCH3) 20:TO-:Zn2+

Zn2+:9 O H3CO H3C

OR

Scheme 3 A simplified process for 9: Zn2+ (
OCH3) or La3+ (
OCH3)2-promoted meth2 anolysis of a carboxylate ester demonstrating the reversible formation of a metal-ion stabilized tetrahedral intermediate.

energies corresponding to k
1 and k2. In the cases where methoxide addition to the ester is metal-ion assisted, the departure of poor leaving groups such as ethoxy or iso-propoxy must be metal ion assisted suggesting that the k2 step must be partially or entirely rate-limiting and Equation (17) applies. As its ss pK a decreases, eventually the leaving group is sufficiently good that it departs without metal-ion assistance and the rate-limiting step for the reaction must change from breakdown to formation of the intermediate. In both the La3+ and 9:Zn2+ cases this is when the leaving group s s pK a is about 14.7. In the plateau regions of Figs 4 and 5 the computed Brønsted b is 0 signifying little or no dependence on the nature of the leaving group. The observation is explained by the limiting case of Equation (16) as having the formation of the intermediate depending on the opposing effects of the leaving group on both the Kb and k1 steps of Scheme 3 and Equations (13) and (14), i.e. (bb+b1)-0. This is reminiscent of the known insensitivity of the rates of acid-catalyzed hydrolysis of esters to changes in the nature of the leaving group (r0) which Taft43 interpreted as resulting from a counterbalancing of the OR/OAr substituent’s electronic effect on the protonation equilibrium and subsequent nucleophilic attack of water on the protonated CQO. The descending wings of Figs 4 and 5 do have a dependence on the nature of the leaving group b ¼
0.71 which, according to Equation (17), is consistent with significant cleavage of the C–OR bond in the TS. Since the net effect of the substituent on the Kb and k1 steps cancel, this leads to the conclusion that the sum of (
b
1+b2) for the reversal of the intermediate and its forward cleavage is significantly negative. Although the above concentrates mainly on the mechanistic aspects of the metalcatalyzed methanolysis process, it should be clear that there could be important applications for transesterifications in synthetic sequences where esters might be used as protecting groups for sensitive alcohols.10a Since the method works under essentially neutral conditions at ambient temperature, it should be particularly applicable to systems where there are acid or base sensitive structures. For example, as is shown in Scheme 4, deprotection of 6-exo-acetoxybicyclo[2.2.2]octan2-one using (La3+(
OCH3))2 generated in situ by simple addition of 10 mM each of La(OTf)3 and NaOCH3 to a methanol solution, gave a 100% yield of the corresponding 6-exo-hydroxybicyclo[2.2.2]octan-2-one within 3 h with no observed epimerization at the 6-position.44 However, when the deacetylation is performed with NaOCH3 alone, there is rapidly established a 15/85 mixture of the exo- and

294

R.S. BROWN AND A.A. NEVEROV O H3CCO

HO

(La3+(-OCH3))2

O

O

CH3OH, 25oC exo NaOCH3 O

OH

H O

NaOCH3

O

endo

Scheme 4 A La3+-catalyzed deprotection reaction of 6-exo-acetoxybicyclo[2.2.2]octan-2one, where the mild nature of the transesterification conditions does not promote rearrangement of the product (ref. 44).

endo-products formed by base-catalyzed equilibration through the aldehyde as shown in the scheme.

MECHANISM OF ALCOHOLYSIS OF NEUTRAL PHOSPHATE ESTERS

Our initial foray into metal-catalyzed alcoholysis of phosphate esters dealt with the methanolysis of the pesticide paraoxon17a where immediately it became evident that the log k2 vs. ss pH plot for La3+-catalyzed reaction showed a distorted bell-shaped profile (given in Fig. 2 of ref. 17a). Preliminary analysis of the profile indicated that the left ascending wing had a gradient of 1 which is consistent with the active species having a single methoxide, but the situation proved to be far more complex. Detailed analysis of the data required consideration of the La3+ speciation as outlined in section ‘Metal ion alcoholysis and titration in alcohol’, Table 3, which generated the same speciation plot as illustrated in Fig. 2. The kinetic data and speciations were fitted to Equation (8), to compute the plot presented in Fig. 6 which
shows the contribution of each of the La3+ 2 ( OCH3)n species to the observed curve (computed rate constants, along with those for the corresponding thiolate derivative 3 given in Table 8). The mechanism of the reaction was not investigated in detail at the early stages of these studies although it was clear that the acceleration of the decomposition of 1 and 3 was spectacular. For example, a 2 mM solution of total [La3+] ion, which gives a 1 mM solution of La3+ 2 , gives a t1/2 for decomposition of 1 of 20 sec at a ss pH of 8.3, corresponding to an acceleration of 109-fold over the background methoxide reaction at the same ss pH (t1/2 ¼ 600 years). With the La3+-catalyzed methanolysis of phenyl acetate (18g, 5) and p-nitrophenyl acetate (18d) carboxylate esters, the effect of the change in leaving group has only minor effect on the rate constant at ss pH 9 (38 and 29 dm3 mol
1 s
1, Table 7). However, preliminary

METAL-CATALYZED ALCOHOLYSIS

295

25

k2obs (dm3mol-1s-1)

20

15

10

5

0 5

6

7

8

9 s s

10

11

12

pH

Fig. 6 Plot of the predicted kobs vs. ss pH rate profile for La3+-catalyzed methanolysis of 2
paraoxon (1) (solid line) based on the kinetic contributions of (left to right), La3+ 2 ( OCH3)1;
3+
3+
2:1 2:2 2:3 La3+ ( OCH ) ; La ( OCH ) and La ( OCH ) computed from the k , k 2 2 2 3 2 3 3 3 4 2 2 , k2 and 2:4 s k rate constants (Table 4), and their speciation as a function of s pH. Reproduced from ref. 10b with permission.

3+
Table 8 Computed second-order rate constants (k2:n 2 ) for the La2 ( OCH3)n-catalyzed at each methanolysis of paraoxon 1 and its thioate derivative 3 obtained through fits of kobs 2 a s s pH to Equation (8)

Paraoxonb 1

O,O-diethyl-p-nitrophenyl phosphorothioatec 3

a


La3+ 2 ( OCH3)n species

3
1
1 k2:n s 2 /dm mol


OCH3 3+
La2 ( OCH3)1
La3+ 2 ( OCH3)2 3+
La2 ( OCH3)3
La3+ 2 ( OCH3)4
OCH3
La3+ 2 ( OCH3)1 3+
La2 ( OCH3)2
La3+ 2 ( OCH3)4

kOCH3 ¼ 0:011 k2:1 2 ¼ 15.872.9 k2:2 2 ¼ 51.171.1 k2:3 2 ¼ 35.676.5 k2:4 2 ¼ 49.771.4 kOCH3 ¼ 0:12 k2:1 2 ¼ 11.675.3 k2:2 2 ¼ 28.471.2 k2:4 2 ¼ 16.171.6

Errors computed from the average % deviation in the fitted numbers calculated by Equation (8) from the actual kinetic data. b Paraoxon data from ref. 17b. c DATA from ref. 17b

296

R.S. BROWN AND A.A. NEVEROV

study17b of the La3+-catalyzed methanolysis process for diethyl phenyl phosphate and 1 showed a very great sensitivity to the leaving group (3.5  10
3 and 51 dm3 mol
1s
1 respectively)17b suggesting the mechanisms for carboxylate and phosphate ester methanolysis might be different. Subsequent work on the La3+- and 9:Zn2+(
OCH3)-catalyzed methanolysis of expanded sets of neutral OP substrates including phosphates and phosphorothioates,17c phosphonates17e and phosphonothioates17g led to the conclusion that all these materials probably react by a common mechanism without the formation of a kinetically detectable intermediate as is detailed below. Phosphates and phosphorothioates,17c phosphonates17e and phosphonothioates17g The metal ion promoted methanolysis of a series of aryl phosphates 1, 20a–f and aryl phosphorothioates 21a–f (in which 21b is the previously described 3) was studied in the presence of La3+ and 9:Zn2+(
OCH3) in greater detail at 25 1C at ss pH values that correspond to the maximum for the metal-catalyzed reaction. In the case of La3+, the reaction ss pH was buffered by N-ethylmorpholine at 9.1 while for 9:Zn2+(
OCH3) the operational ss pH was set at the ss pK a of 9.1 for ionization of 9:Zn2+(HOCH3) through half neutralization. The second-order rate constants for the metal-catalyzed reactions at this ss pH were determined from the gradients of the plots of the pseudo-first-order rate constants for methanolysis of series 20 and 21 vs. [catalyst]. These are presented in Tables 9 and 10 along with those for the methoxide reactions at 25 oC.17c The Brønsted plots shown in Figs 7 and 8 and their gradients are computed from all data except for the lowest ss pK a member in each plot which can be seen to fall significantly below the lines. O EtO

P

O OAr

OEt 20 a, Ar = phenyl b, Ar = pentafluorophenyl c, Ar = 4-chloro-2-nitrophenyl d, Ar = 3-nitrophenyl e, Ar = 4-chlorophenyl f, Ar = 4-methoxyphenyl (1) Ar = 4-nitrophenyl

EtO

P

SAr

OEt 21 a, Ar = phenyl b, (3) Ar = 4-nitrophenyl c, Ar = 4-chlorophenyl d, Ar = 3,5-dichlorophenyl e, Ar = 4-fluorophenyl f, Ar = 4-methoxyphenyl

The Brønsted plots of the log kOMe for methoxide reactions of the phosphate and 2 phosphorothioate esters vs. the ss pK a values45,46 (not shown here) provide reasonable linear correlations, the respective blg values being –0.7070.05 and –0.7670.08. As is the case for the metal-catalyzed reactions shown in Figs 7 and 8, the methoxide reactions also show that the derivatives having the lowest ss pK a values of the leaving phenol/thiophenol (20b and 21b (3)) are less reactive than predicted on the basis of the equilibrium ss pK a data. Cursory analysis might suggest a curvature in all the

METAL-CATALYZED ALCOHOLYSIS

297

Table 9 Second-order rate constants (k2) for the methanolysis of phosphates 20 and 1 promoted by methoxide, La3+ and 9:Zn2+(
OCH3) in methanol solvent, T ¼ 25 1C k2 (OMe) (dm3 mol
1 s
1)

3
1
1 s ) kLa 2 (dm mol

Aryloxyphosphate

pHa a

b s s pHa

Pentafluoro (20b) 4-chloro-2-nitro (20c) p-nitro (1) m-nitro (20d) p-chloro (20e) p-H (20a) p-methoxy (20f)

5.53 6.32

8.84 10.64

0.20170.002 (6.470.1)  10
2

1070740 18577

23.070.6 11.470.4

7.14 8.39 9.38 10.0 10.20

11.30 12.41 13.59 14.33 14.7

(1.0270.03)  10
2 (6.070.1)  10
3 (6.370.1)  10
4 1.4  10
4d (6.5070.01)  10
5e

23.270.9d 2.4270.07 (1.9870.07)  10
2 1.97  10
3d (2.270.1)  10
4e

1.3 0.5870.01 (8.0470.04)710
3

k29:Zn(OMe) (dm3 mol
1 s
1)

(2.2570.05)  10
4e

a

Values of the pK a in water from ref. 42.
3 Values of the ss pK a in methanol from refs. 16f, 6a and 6b.cValues of kLa 2 determined in a 17 mmol dm N-ethylmorpholine buffer at ss pH 9.1. d Values of k2 from ref. 17b.eObtained from duplicate initial rate measurements monitored by 1H NMR in CD3OD as described in ref. 17c. b

Table 10 Second-order rate constants (k2) for the methanolysis of phosphorothioates 21 and 3 promoted by methoxide, La3+ and 9:Zn2+(
OCH3) in methanol solvent, T ¼ 25 1C Aryl Sphosphorothioate

pKaa

b s s pK a

p-nitro (21b, 3) 3,5-dichloro (21d) p-chloro (21c) p-fluoro (21e) p-H (21a) p-methoxy (21f)

4.61 5.07 5.97 6.54 6.68 6.95

8.4c 8.9 10.1 10.7 10.9c 11.2

a

kOMe 2 (dm3 mol
1 s
1) 0.12d 0.15270.001 (1.8870.03)  10
2 (1.1170.02)  10
2 4.8  10
3d (2.2270.03)  10
3

3
1
1 s ) kLa 2 (dm mol

12.4d 14.070.5 1.2370.05 0.4670.01 0.48d (9.370.2)  10
2

k29:Zn(OMe) (dm3 mol
1 s
1) 0.8470.01 0.9770.01 (11.670.01)  10
2 (5.3470.06)  10
2 (4.270.1)  10
2 (1.4670.04)  10
2

Aqueous pKa values from Hong, S.-B. and Rauchel, F.M. (1996). Biochemistry 35, 10904. Values of the ss pK a in methanol computed from 2-point linear regression pK(MeOH) ¼ 1.2 (pK(waa a

b

ter))+2.83. c

The experimental valuses ss pK a from Clare, B.W., Cook, D., Ko, E.C.F., Mac, Y.C. and Parker, A.J. (1966). J. Am. Chem. Soc. 88, 1911. d The value of kLa 2 from ref. 17b.

plots that is indicative of a change in the rate-limiting step commencing with the lowest ss pK a compounds. However, more detailed analysis indicates this is probably not so since non-linear Hammett and Brønsted behaviors have been observed before for the hydroxide reactions of substituted aryl benzoates47 and acetates.48 In addition Schowen49 reported that log k2 values for methoxide reactions of aryl acetates and carbonates in methanol are not linearly related to the ss pK a values for ionization of the corresponding phenols. This sort of curvature in the Brønsted plots does not result from a change in mechanism or rate-limiting step but from a greater importance of the resonance and inductive interactions in the equilibrium acid dissociation constants (on which the Hammett and ss pK a values are based) than in the kinetic processes where less charge development occurs in the rate-limiting TS.47–49 Since

298

R.S. BROWN AND A.A. NEVEROV

2.5 (dm3 mol-1s-1)

log k2La(OMe) or log k29:Zn(OMe)

5.0

La 0.0

Zn

-2.5

-5.0 8

9

10

11

12 13 ArOH

14

15

s s pKa of

Fig. 7 Brønsted plot of the log second-order rate constant for La3+- and 9:Zn2+(
OCH3)catalyzed methanolysis of phosphates 20 (including 1) vs. the ss pK a values for the corresponding phenols; linear regressions through the La3+ and 9:Zn2+(
OCH3) data (excluding 20b) give gradients of –(1.4370.08) (solid line, &) and
(1.1270.13) (dashed line, J), respectively. Note that points on lower right for the p-methoxy derivative (20f) are coincident. Reproduced from ref. 17c with permission.

log k2La or log k29:Zn(OMe) (dm3mol-1s-1)

1 La 0 Zn -1

-2

8

9 s pK s a

10 of ArSH

11

Fig. 8 Brønsted plot of the log second-order rate constant for La3+- and 9:Zn2+(
OCH3)catalyzed methanolysis of phosphorothioates 21 (including 3) vs. the ss pK a values for the corresponding thiophenols; linear regressions through the La3+ and 9:Zn2+(
OCH3) data (excluding that for 3) give gradients of
0.8770.10 (solid line, J) and
0.7470.06 (dashed line, J), respectively. Reproduced from ref. 17c with permission.

METAL-CATALYZED ALCOHOLYSIS

299

the visible curvature in the Brønsted plots for methanolysis and metal-catalyzed methanolysis seems to occur at the same point, a plot of the log second-order rate constants for the two processes should be a straight line if the curvature is an artefact of the ss pK a scale. Indeed, as is shown in Fig. 12, these plots for the phosphates and phosphorothioates are linear, indicating that nucleophilic attack of the metal-coordinated alkoxide and methoxide behave the same way in response to varying the leaving group without a change in rate-limiting step over the series. The above gives a mechanism based justification for excluding the lower ss pK a datum for linearization of the Brønsted plots for 20b and 21b for each of the methoxide and metal-catalyzed methanolysis reactions. The blg values of –0.70 and –0.76 for the methoxide reaction of the phosphate triesters and phosphorothioates, respectively, can be compared with the corresponding blg values of
0.43 and –0.44 for hydroxide attack on diethyl aryl phosphate triesters45,50 and –0.42 for hydroxide attack on diethyl S-aryl phosphorothioates.45 The relatively low blg values of 
0.4 obtained with these nucleophiles is consistent with little cleavage of the P–OAr bond in the TS, and supports a two-step mechanism where there is a rate-limiting k1 step to form a pentacoordinate intermediate with preferential breakdown of the intermediate to product. In the case of the HO
reacting with diethyl aryloxy phosphates Ba-Saif and Williams50 judged that HO
displacement of aryloxy leaving groups was probably concerted although with little cleavage of the ArO–P bond. This conclusion is supported by the 18O-phenoxy kinetic isotope effect of 1.006 for hydroxide-promoted cleavage of paraoxon (1) which was interpreted51 as being consistent with a bond order of 0.75 for the P–OAr bond in the ‘SN2-like transition state of an associative mechanism with concerted, asynchronous departure of the leaving group’.

Methoxide reaction. For the methoxide reactions of the aryl phosphates 1 and 20 and the diethyl S-aryl phosphorothioates 21, the blg values are more negative by 0.3 unit than is the case for the hydroxide reactions.45,50 Such might suggest that there is more cleavage of the P–XAr bond in the transition state than is the case for the hydroxide reactions. Applying the ‘effective charge treatment’ described by Jencks52 and Williams53 to the Brønsted blg value suggests a process for the aryloxy phosphate triesters where the rate-limiting transition state has appreciable changes in the P–OAr bond. This could be result from a two-step process with CH3O
attack being largely rate-limiting due to the fact that the methoxide nucleophile is a far poorer leaving group and better nucleophile than any of the aryloxy anions,54 or just as likely with a concerted process as shown in Scheme 5. It is instructive to consider the progress of the P–OAr bond cleavage in the TS in terms of the Leffler parameter, a, which measures the ratio of the Brønsted blg for the TS relative to the beq for equilibrium transfers of acyl or phosphoryl groups between oxyanion nucleophiles. In the case of the transfer of the (EtO)2P ¼ O group, the beq value is
1.8750 which comes about because the O–Ar oxygen in the starting material has a net effective charge of +0.87. When methoxide is the nucleophile, the Leffler parameter of blg/beq ¼ 0.37 suggests that the P–OAr bond change has progressed 37% of the way from starting material to product.

300

R.S. BROWN AND A.A. NEVEROV

CH3O- +

O α=0.37

O

-1

+0.87 EtO

P

CH3O

OAr βlg = −0.70

EtO

EtO

P

+ -+

OAr

OEt

O CH3O EtO

P

-1 + ArO OEt

Scheme 5 A proposed concerted mechanism for methoxide promoted methanolysis of carboxylate esters with displacement of aryloxy leaving groups.

For the methoxide reaction of the phosphorothioates, a similar sort of analysis for the reaction proceeding through a two-step or concerted process can be invoked. In this case there is no reported value for the effective charge on the S-atom in the ArS–P(QO) unit but based on comparison of the known effective charges on the S- and O-atoms of ArS–C(QO)CH3 and ArO–C(QO)CH3 of 0.4 and 0.7,53a one might expect that S is less positive than O in the case of the ArX–P(QO) unit. Assuming the effective charge on S is 0.5–0.6, the Leffler a of 0.45–0.50 for a concerted P–SAr cleavage suggests that the change in the P–S bond has progressed about 50% of the way from starting material to product. Unfortunately, the data are not sufficient that one can unambiguously tell the difference between a concerted reaction and one proceeding through a pentacoordinate intermediate. The situation might be diagrammatically shown by the MoreO’Ferrall Jencks diagram as in Fig. 9. This diagram is predicated on the expectation that the Q-corner, involving full cleavage of the P–XAr bond prior to methoxide attack, is sufficiently high in energy that any concerted process moves substantially toward the S-corner describing the pentacoordinate intermediate. If the intermediate is in fact formed, then either its formation (with a late transition state) or its breakdown (with an early transition state) could be rate-limiting, with the latter probably better fitting Leffler index data that shows the substantial progress in the P–XAr bonding. Metal– methoxide reactions of phosphates, phosphorothioates, phosphonates and phosphonothioates. The rate constant data for the (La3+(
OCH3))2 and 9:Zn2+(
OCH3)-catalyzed methanolysis of the phosphates and phosphorothioates are given in Tables 9 and 10, respectively, while those for the phosphonates (22a–f)17e and phosphonothioates (23a–e)17g are given in Tables 11 and 12 along with the corresponding methoxide data. The latter two series data are displayed in Figs 10 and 11 as Brønsted plots. O EtO

P

OAr

CH3 22 O EtO P SAr CH3 23

22 a Ar = (4-Cl-2-NO2)phenyl b Ar = (4-NO2)phenyl c Ar = (3-NO2)phenyl d Ar = (4-Cl)phenyl e Ar = phenyl f Ar = (4-OCH3)phenyl 23 a Ar = 3,5-dichlorophenyl b Ar = 4-chlorophenyl c Ar = 4-fluorophenyl d Ar = phenyl e Ar = 4-methoxyphenyl

METAL-CATALYZED ALCOHOLYSIS

301

Q CH3O- + (EtO)2P+(=O) +-XAr P---XAr dist.

P

CH3O-P(=O)(OEt)2 +-XAr

* *

CH3O- + (EtO)P(=O)XAr

* R

OCH3O-P(OEt)2

CH3O---P dist.

S

XAr

Fig. 9 A hypothetical More-O’Ferrall Jencks diagram for the attack of methoxide on O-aryl phosphate triesters (20) and S-aryl phosphorothioates (21). Note that the diagram for attack of a metal-coordinated methoxide would be similar, but Mx+-coordination would push the TS toward the S-corner, possibly stabilizing the pentacoordinated intermediate to the point that the reaction occurs stepwise with the likely rate-limiting step being breakdown. Table 11 Acid dissociation constants for the phenols in water and methanol, as well as second-order rate constants for the various methanolysis reactions of phosphonates 22a–e promoted by methoxide, La3+ and 9:Zn2+(
OCH3) Aryloxyphosphonate 22a 22b 22c 22d 22e 22f a

pKa

s s pK a

kOMe (mol dm
3 s
1) 2


3
1 kLa s ) 2 (mol dm

k29:Zn(OMe) (mol dm
3 s
1)

6.32 7.14 8.39 9.38 10.00 10.20

10.64 11.30 12.41 13.59 14.33 14.70

14.370.1 (2171)a 2.7070.02 1.4470.02 0.11870.001 0.021770.0003 0.008470.0001

(2.6070.20)  104 (1.4570.03)  103 (2.6070.20)  102 3.9770.03 (4.070.1)  10
1 (1.4070.02)  10
1

51773 (510720)a 46.870.5 2271 0.4570.01 0.07270.002 0.01470.002

In DOCH3, determined at a 9:Zn2+(
OCH3) ratio of 1:1:0.5.

The Brønsted plots for the (La3+(
OCH3))2 and 9:Zn2+(
OCH3)-catalyzed methanolyses of 20 and 21 shown in Figs 7 and 8 exhibit gradients that are much steeper than the corresponding Brønsted plots for the methoxide reactions. This is easily visualized in the log/log plot of the second-order rate constants of the metalcatalyzed reactions vs. the methoxide reaction which is shown in Fig. 12 where the gradient of the La3+ plot is 1.9470.10 while that for the 9:Zn2+(
OMe) plot is 1.4970.11. The situation for the phosphorothioates is not nearly so pronounced since the similar plots for series 21 (not shown) has a gradient for the La3+ plot of 1.1570.10 while that for the 9:Zn2+(
OMe) plot is 0.9970.06.17c δ− OAr

O MX+

H3C

P O-

MX+

H3C

OAr P

OEt

OEt 24c

O

O

OEt

OEt 24i

302

R.S. BROWN AND A.A. NEVEROV

H

α = 0.60

N

OEt

O

N H

OAr

P OEt

Zn2+ -

OCH3

N H

25

The large negative blg values for the metal ion-catalyzed methanolysis of the phosphate esters suggests a process where there is considerable cleavage of the leaving group in the transition state, far more so than is the case for the methoxide reaction. In the cases of the carboxylate esters that we have reviewed in section ‘Mechanism of alcoholysis of carboxylate esters’, it appeared that the 9:Zn2+(
OCH3)- and (La3+(
OCH3))2-catalyzed reactions produce an intermediate whose formation and breakdown could be rate-limiting depending on whether the leaving OR/OAr group was actually stabilized by metal ion coordination. However, the situation with the present phosphate esters is clearly different because there is a large dependence of the rate on the leaving group in the ss pK a region where blg is zero for the reaction of the carboxylate esters. Accordingly we suggest that, for the metal-catalyzed methanolysis of these phosphates, there is little evidence for a

Table 12 Second-order rate constants for the methanolysis of phosphonates 23a–e catalyzed by methoxide, (La3+(
OCH3))2, and 9:Zn2+(
OCH3) in methanol at T ¼ 25oC Phosphonothioate 23a 23b 23c 23d 23e Et)(CH3)P( ¼ O)SCH2CH2NEt2 a

of thiol (methanol)a

kOMe 2 (mol
1 dm3 s
1)

k29:Zn(OMe) (mol
1 dm3 s
1)b

kLa 2 (mol
1 dm3 s
1)c

9.0870.04

2.1770.03 (1.9670.03)d 0.6070.01 0.1670.01 0.05470.002 0.04070.006

95.271.4 (75.774.6)d 7.8570.25 4.8370.11 2.6070.06 1.1670.03 38.0270.04f

(4.8470.09)  103

s s pK a

10.4770.01 11.0770.03 11.2870.14 11.9870.08 9.5470.04

e

(4.2570.12)  102 (1.6470.07)  102 (1.1070.02)  102 (3.5570.17)  101 (2.0670.09)  103f

Experimental values from 2 mmol dm
3 solutions titrated in methanol according to procedures in refs. 6 and 7. b Catalyst prepared in situ by adding 1 equivalent of each of Zn(OTf)2 and the triaza ligand along with 0.5 equivalent of tetrabutylammonium hydroxide in methanol to form solution at ss pH 9.1; kinetics determined in duplicate at 5 [catalyst] ranging from 0.4 to 2.0 mmol dm
3. c Catalyst prepared in situ by adding a stock solution of La(OTf)3 in methanol to a 17 mmol dm
3 Nethylmorpholine with perchloric acid in a 4:1 ratio to solution at ss pH 9.1; kinetics determined in duplicate at 5 [La3+]total from 0.4 to 2.4 mmol dm
3. d In d1-methanol. e The value predicted is 2  10
3 mol
1 dm3 s
1 under highly basic conditions based upon a comparison with data for for O-ethyl S-ethyl methylphosphonothioate given in Yang, Y.-C., Berg, F.J., Szafraniec, L.L., Beaudry, W.T., Bunton, C.A. and Kumar, A. (1997). J. Chem. Soc., Perkin Trans. 2, 607. f Predicted kcatalyst computed using Equations (32) and (33). 2

METAL-CATALYZED ALCOHOLYSIS

303

log k2 (dm3mol-1 s-1)

5.0

2.5

0.0

-2.5 10

11

12 s s pKa

13

14

15

phenol

Fig. 10 Brønsted plots of the second-order rate constants for methoxide and metal ioncatalyzed methanolysis of 22a–f at 25 1C. (La3+(
OCH3))2 (m, gradient ¼
1.2670.06), 9:Zn2+(
OCH3) (&, gradient ¼ +
1.0670.09),
OCH3 (., gradient ¼
0.7670.06). Reproduced with permission from ref. 17e.

logk2 (dm3mol-1 s-1)

5.0

2.5

0.0

-2.5

9

10 s s pKa

11

12

of ArSH

Fig. 11 Brønsted plots for log kcatalyst vs. ss pK a of aryl thiol for the methanolysis of 23a–e, 2 3+
K, (La ( OCH3))2, gradient ¼
0.7570.01; m, 9:Zn2+(
OCH3), gradient ¼
0.6670.04; ~,
OCH3, gradient ¼
0.6570.10. Redrawn from ref. 17g.

change in rate-limiting step for these aryloxy leaving groups. This leaves open the possibility that the reactions with the metal ions are proceeding via a hypothetical concerted TS 24c, but it cannot exclude the possibility that the reaction is stepwise with a metal stabilized pentacoordinated phosphorus intermediate (24i), the formation or breakdown of which could be rate-limiting throughout the series. This possibility is shown diagrammatically in Fig. 9. Although we have never observed evidence of saturation kinetics in any of the metal-catalyzed reactions of the neutral phosphorus esters,17 it is difficult to envision

304

R.S. BROWN AND A.A. NEVEROV

3

log k2La or log k29:Zn(OMe) (dm3mol-1s-1)

2 1 0 -1 -2 -3 -4 -4

-3

-2

-1

0

log k2OMe (dm3mol-1s-1) 9:Zn(OMe) Fig. 12 Plots of log kOMe vs. log kLa (J) for the methanolysis of 2 2 (&) or log k2 phosphate esters (20, including paraoxon 1) at 25 1C: gradient for La3+ plot is 1.9470.10; slope for 9:Zn(OMe) plot is 1.4970.11 all data included. Note data points for the p-methoxy derivative (20f) in lower left corner are coincident. Reproduced with permission from ref. 17c.

a mechanism where the ion does not bind to the phosphate to provide Lewis activation toward attack. Indeed, phosphate complexation of lanthanides and actinides is well known55 and structures are reported for Zn2+-complexes of phosphine oxides56 and tritoluoyl phosphate57 where coordination does occur through the PQO unit. Two of the most likely mechanistic possibilities are shown in the equations below and involve a preequilibrium binding of the metal ion to the PQO unit, Equation (18), followed by a concerted displacement Equation (19), or a stepwise formation of a metal-coordinated pentacoordinated intermediate, Equation (20). In fact, derivation of the kinetic expression corresponding to the binding and two-step process of Equations (18) plus (20) gives the identical expressions as derived in Equations (15–17) for the carboxylate esters, while the concerted mechanism corresponding to Equations (18) plus (19) has the identical kinetic expression as given in Equation (16). In what follows we will deal with each of the possibilities. Kb MX+-(− OCH3):(EtO)2P(=O)XAr MX+-(−OCH3) + (EtO)2P(=O)-XAr (18) MX+-(−OCH3):(EtO)2P(=O)XAr

k1

P

(19)

METAL-CATALYZED ALCOHOLYSIS

MX+-(−OCH3):(EtO)2P(=O)XAr

305

k1 k-1

MX+- − OP(OCH3)(OEt)2XAr k2 MX+ + (EtO)2P(=O)(OCH3) + − XAr

(20) Concerted mechanism?. For the concerted mechanism, the kinetic expression (16) is used from which it can be seen that the overall Brønsted relationship is log kobs 2 ¼ flog C b þ log C 1 g þ ðbb þ b1 ÞpK a

(21)

where bb and b1 are the Brønsted b-values for the binding step and kinetic step and pKa represents the ss pK a values for the phenol or thiophenol in methanol. The actual linear regressions are: s HOAr log kLa 2 ðphosphates 20Þ ¼ ð17:60  1:07Þ
ð1:43  0:05Þs pK a

(22)

s HOAr log k9:ZnðOMeÞ ðphosphates 20Þ ¼ ð13:05  1:59Þ
ð1:12  0:13Þs pK a 2

(23)

s HSAr log kLa 2 ðphosphorothioates 21Þ ¼ ð8:93  1:01Þ
ð0:87  0:10Þs pK a

(24)

s HSAr log k9:ZnðOMeÞ ðphosphorothioates 21Þ ¼ ð6:64  0:69Þ
ð0:74  0:06Þs pK a 2

(25)

from which it can be seen that (bb+b1) assumes large negative values which correspond to a composite measure of the influence of the leaving group on the preequilibrium binding step and the kinetic step. It is difficult to predict an exact value for the bb and what few data there are available in the literature58 suggest that this should be (+), but probably not large because electron donors are expected to promote the binding, but these are somewhat far away from the PQO binding site. Rackham reported that the europium shift reagent Eu3+(2,2,6,6-tetramethylheptane-3,5-dione
)3 binds trimethyl phosphate and triphenyl phosphate with constants of 348 and 23.2 dm3 mol
1, and suggested that the fall in the latter’s binding constant can be attributed to steric bulk and the inductive withdrawal of the phenoxy group relative to the methoxy group.59 Du Preez and Preston have reported that the extraction into toluene of ScIII, YtIII and the trivalent LnIII ions from aqueous nitrate solutions by coordination to neutral organophosphorus (P ¼ O) compounds correlates with the Taft s*-values of the substituents.55d These considerations indicate that for the La3+ or 9:Zn2+(
OCH3) promoted methanolysis of the phosphates, the observed blg may be a lower limit measure of b1 because it will be offset by the positive bb. The general mechanism for the (La3+(
OCH3))2-catalyzed concerted process is given in Scheme 6 in which, for the sake of visual clarity, we have omitted the methanols of solvation on each La3+ as well as any associated counterions. As was proposed for the carboxylate esters, a methoxy group bound between two La3+ ions will not be sufficiently nucleophilic to attack the phosphate,60 so we propose that the

306

R.S. BROWN AND A.A. NEVEROV CH3

CH3

O

O+

La3+

La3+

O-

Kb

O-

EtO

P

O-

EtO 20, 21

CH3

XAr

La3+

La3+

XAr

P

O CH3 complex

OEt

OEt

+HOCH3 -H+ -phosphate CH3

CH3 O-

La3+ O

La3+

O

P

-XAr-

O

k1

La3+ P

O H3C

CH3

La3+ α=0.76

O-

XAr

O-

La3+ O

La3+ O-

P

EtO OEt H3C

EtO TS

OEt

CH3

EtO

XAr

OEt

open complex

Scheme 6 A proposed mechanism for the concerted La3+ (
OCH3)2-catalyzed methanolysis 2 reaction of phosphate triesters with XAr leaving groups.

pre-equilibrium binding of phosphate and metal ion to form the complex induces opening of one of the methoxy bridges to reveal a kinetically active open complex. Since the Leffler parameter, a, for the La3+-catalyzed is blg/beq ¼
1.43/
1.87 ¼ 0.76, the transition state for this reaction (24c) has extensive cleavage of the P–OAr bond which would be consistent with a concerted reaction within the complex as drawn in Scheme 6. Catalytic turnover requires a final dissociation of the diethyl methyl phosphate with the reformation of (La3+(
OCH3))2. In the case of the 9:Zn2+(
OCH3)-catalyzed reaction a similar concerted mechanism might be envisioned but this time the transition structure (shown above as structure 25) will be five-coordinate at Zn and P with a Leffler a of blg/beq ¼
1.12/
1.87 ¼ 0.60. This also signifies extensive dissociation of the P–OAr bond in the transition state, but less so than in the case of La3+ catalysis. In the case of La3+- and Zn2+-catalyzed methanolysis of the phosphorothioate esters the observed blg values of
0.87 and
0.74 also signify an associative mechanism with some departure of the leaving group, but it is difficult to assign the extent of the bond cleavage since the beq value is not known for the phosphoryl transfer between thiol and oxygen nucleophiles. Stepwise mechanism?. The possibility exists that the metal-catalyzed reaction really proceeds as described in Equations (18) plus (20), where the actual kinetic steps involve formation of a pentacoordinate phosphorus intermediate, suggested to be akin to 24i. When viewed within the context of the More-O’Ferrall Jencks

METAL-CATALYZED ALCOHOLYSIS

307

diagram of Fig. 9, the S-corner of the diagram should be stabilized by association of the anionic intermediate with the electropositive metal ion driving the reaction further to a stepwise one. From the general steady state expression of Equation (15) can be deduced two limiting Brønsted relationships given in Equations (21) and (26) log kobs 2 ¼ flog C b þ log C 1 þ log C 2
log C
1 g þ ðbb þ b1
b
1 þ b2 ÞpK a

ð26Þ

where the rate-limiting steps are, respectively, formation of the pentacoordinate intermediate and its breakdown. In the former case, all the arguments for blg ¼ bb+b1 are essentially the same as for the concerted mechanism, but with the large negative blg now referring largely to the kinetic step of formation of the metal-coordinated pentacoordinate intermediate. In this case the Leffler index blg/beq ¼
1.43/
1.87 ¼ 0.76 is difficult to rationalize as it implies a very extensive loosening of the P–XAr bond. However, under the likelihood that the rate-limiting step is the breakdown of the pentacoordinate intermediate, then blg ¼ (bb+b1
b
1+b2) and the Leffler index makes more sense since the blg now explicitly includes a term relating to the step that involves cleavage of the intermediate. The (La3+(
OCH3))2 and 9:Zn2+(
OCH3) promoted methanolyses of the phosphonates (22)17e and phosphonothioates (23)17g generally follow the same sort of trends as the phosphates and phosphorothioates discussed above so they need not be discussed in great detail here. Analysis of the linear Brønsted plots for the phosphonates 22 gives the relationships shown in Equations (27)–(29) which shows the common trend that the blg observed for the metal-catalyzed reactions are greater than that of the methoxide reactions. Since the Leffler parameter, a, for the La3+-catalyzed ¼ ð9:24  0:81Þ
ð0:76  0:06Þss pK HOAr ; log kOMe 2 a s HOAr ; log kLa 2 ¼ ð17:78  0:84Þ
ð1:26  0:06Þs pK a

r2 ¼ 0:9896; 6 data

(27)

r2 ¼ 0:9716; 6 data

(28)

¼ ð14:04  1:17Þ
ð1:06  0:09Þss pK HOAr ; log k9:ZnðOMeÞ a 2 17e

r2 ¼ 0:9734; 6 data

(29)

process is
blg/beq ¼ 1.26/1.5 ¼ 0.84, the transition state has extensive cleavage of the P–OAr bond which is best analyzed at this stage in terms of the concerted process or a two-step mechanism where breakdown of the Mx+-pentacoordinate intermediate is rate-limiting. In the case of the 9:Zn2+(
OCH3)-catalyzed reaction a similar mechanism is envisioned but this time the transition structure will involve five-coordinate Zn2+ with a Leffler a of
blg/beq ¼ 1.06/1.5 ¼ 0.7. Catalysis of the cleavage of the phosphonothioates (23) also follows linear Brønsted relationships (Equations (30–32)) although the gradients of the metalcatalyzed processes are not nearly as steep as those of the phosphonates discussed above. This trend of steeper gradients for the OP esters bearing OAr leaving groups relative to SAr leaving groups seems to be a general one. The reasons for this are not immediately clear although it may be related to the fact that the O–Ar oxygen of the starting ester may be significantly more positively charged than the S–Ar sulphur, i.e. beq (OAr)obeq (SAr) so that the lower Brønsted blg values still represent a very

308

R.S. BROWN AND A.A. NEVEROV

significant cleavage of the P–Ar bond in the transition state. log kOMe ¼ ð6:39  1:12Þ
ð0:65  0:10Þ ss pK HSAr 2 a s HSAr log kLa 2 ¼ ð10:51  0:10Þ
ð0:75  0:01Þ s pK a

r2 ¼ 0:930

(30)

r2 ¼ 0:999

(31)

¼ ð7:89  0:41Þ
ð0:66  0:04Þ ss pK HSAr log k9:ZnðOMeÞ a 2

r2 ¼ 0:990

(32)

Relative to the reactions of the corresponding O,O-diethyl O-aryl phosphate triesters with the same leaving groups,17c the phosphonates are about 100-fold more reactive with toward methanolysis promoted by methoxide, La3+ and the 9:Zn2+(
OCH3) systems. Nevertheless, the reaction rates are quite spectacular considering that the catalysts are maximally operative at ss pH values near neutrality (8.34) in methanol. For the most reactive substrate (22a), a solution containing
2+
1 mmol dm
3 of catalyst (La3+ ( OCH3)) accelerates the meth2 ( OCH3)2 or 9:Zn anolysis relative to the background methoxide reaction at ss pH optimum values of 9.1 by 8.5  107 and 1.7  106 times, leading to t1/2 values of 0.026 and 1.33 seconds, respectively. Compounds of this general structure can be used as simulants for OP chemical warfare agents such as sarin (26), soman (27) and the V-agents such as VX (28). The fact that the metal-catalyzed methanolysis reactions of the neutral OP esters described above is so efficient suggests that this methodology might prove effective in decomposition of OP CW agents, particularly in situations where neutral conditions and ambient temperature is desirable for decontamination of sensitive equipment. While we have not yet tested these systems on live agent, we can use the data here to predict the reactivity of the particularly noxious VX agent. To this end, we can use the Brønsted relationships of Equations (31) and (32) determined for the catalyzed reactions of the phosphonothioates (23) to predict the rate constants shown in O

O (CH3)2CHOP

F

CH3 26 3+

(CH3)2CHCHOP F CH3 CH3 27

O EtO

P

SCH2CH2N(CH(CH3)2)2

CH3 28

2+

Table 12 for the La and 9:Zn -catalyzed methanolysis of VX. Based on the experimental first ss pK a of 9.54 for ionization of HSCH2CH2NEt2, which we feel is a good model for the HSCH2CH2N(CH(CH3)2)2 in 28, the predicted half-times for methanolysis of VX are 18 and 0.33 seconds in the presence of 1 mmol dm
3 of 9:Zn2+(
OCH3) or (La3+(
OCH3))2.

6 Transition metal ion and La3+-catalysis of the alcoholysis of phosphate diesters Phosphate diesters of general structure 29 are among the most resistant species to hydrolysis due the reluctance of hydroxide or even water to attack an anionic

METAL-CATALYZED ALCOHOLYSIS

309

61 (RO)2PO
On the other hand, hydrolysis is achievable under highly acidic 2 species. conditions once the phosphate is neutralized by protonation. In principle, metalcatalyzed hydrolysis should be an

O O-

O

O

ROPO

nucleobase

P OR'

RO

30 X=H, DNA 31 X=OH, RNA

-O

O

29 O

P -O

X OR

effective strategy if the role of the metal ion is to complex the phosphate in a Lewis acid/base fashion thereby neutralizing its negative charge and permitting attack of an anionic or metal-coordinated HO
. In several reports where metal-catalyzed hydrolysis of phosphate diesters has been discussed, a dual role of the metal ion was proposed, acting both as a Lewis acid activator and deliverer of an intramolecularly bound
OH group to the metal-bound OQP unit.62,63 Metal ion catalysis is not a universally effective for phosphate diester hydrolysis in water although a large number of studies have been reported where these do catalyze the hydrolysis of selected phosphate diesters (mainly p-nitrophenyl substituted).62 In aqueous solution, due to the solvation effects of H2O on both the metal ion (complex) and phosphate diester, their binding is not particularly effective unless special efforts are made to enhance this.64 Owing to their great functional group stability phosphodiesters make up the important link in the RNA and DNA biomolecules (general structures 31 and 30) that are responsible for the storage of genetic information.11,65–68 While the respective half-times for hydrolysis of RNA and DNA at pH 7 and 25 1C are 110 and up to 100 billion years, Nature provides enzymes that promote the cleavage by up to a factor of 1015–16 thus affording some of the most spectacular rate enhancements known. Many of these enzymes have active sites comprising two or more metal ions (usually Zn2+ and in some cases Mg2+, Ca2+ and Fe2+) as exemplified by phosphodiesterases such as ribonuclease H from HIV reverse transcriptase,66 30 ,50 -exonuclease from DNA polymerase I,67 the P1 nucleases68 and phospholipase C.11,65 In order to cast some light on the metal ion-mediated biological phosphoryl transfer reactions, intense research was directed at understanding the origins of catalysis of phosphate diester cleavage provided by metal ion systems.69–71 From that earlier work and more recent reports72–78 it is seen that dinuclear complexes are usually more reactive than the mononuclear counterparts, and there are four proposed roles by which the dinuclear catalysts promote the phosphoryl transfer reactions: (1) by double Lewis acid activation of a phosphate diester through M2+–
OP(OR)(OR)O–M2+ coordination; (2) through bifunctional catalysis whereby the metal ion activates the bound phosphate and delivers a metal-coordinated hydroxide, alkoxide or oxide that serves as a nucleophile or base; (3) by electrophilically assisting the departure of the phosphate’s leaving group through metal-coordination; and (4) as an

310

R.S. BROWN AND A.A. NEVEROV

electrostatic reservoir of (+)-charge to interact favorably with the anionic phosphate to stabilize the transition state for the phosphoryl transfer reaction. As important as all these effects are, there must be an additional factor which has heretofore not been demonstrated experimentally, as none of the Zn2+-based model systems achieves a catalysis in water that even remotely approaches enzymatic rates. However, as will be shown below, metal-catalyzed alcoholysis reactions shed important light on the effectiveness of a reduced dielectric constant/polarity medium as a possible way that Nature could accelerate these reactions. As will be seen, there are very profound increases in the strength of phosphate diester binding to metal ions in alcohols relative to water, and once bound the cleavage of these exhibits spectacular rate accelerations over the background reactions.

METAL-CATALYZED ALCOHOLYSIS OF AN RNA MODEL

The phosphate diester 2-hydroxypropyl-p-nitrophenyl phosphate79 (32, HPNPP) has been employed in numerous studies as an easily obtained and studied surrogate for RNA. The general mechanism for its reaction is depicted in Scheme 7 involves an intramolecular attack of the 2-hydroxy group on the phosphorus with expulsion of the leaving group to produce a cyclic phosphate (33) which undergoes subsequent ring opening to give 34a and 34b. The intramolecular cyclization confers a kinetic reactivity toward base promoted methanolysis that is some 3000-fold larger than a comparable phosphate diester that lacks this group (e.g. methyl p-nitrophenyl phosphate).80 The alcoholysis of this sort of phosphate can be directly compared with the situation in water since the cleavage step leading to the cyclic phosphate does not incorporate solvent during the departure of the leaving group which is the observable step in the catalyzed reaction. Our initial studies of the La3+-catalyzed methanolysis of HPNPP81 immediately indicated that there were important features of the reaction which are not seen in the O NO2

HO

O

P O O

HO

O

P

34a

O-

O

k1

O

P O

O

O + O-

33

32

NO2 O

+ OCH3 HO

k2(k2')

O 34b

P

OCH3

O-

Scheme 7 Generalized mechanism for the methanolysis reaction of 32, HPNPP proceeding via intramolecular cylcization to form a 5-membered cyclic phosphate which undergoes subsequent opening via attach of methoxide.

METAL-CATALYZED ALCOHOLYSIS

311

80

5

10 kobs, sec-1

60

40

20

0 0

1

2

3

104La(OTf)3,

4

5

M

Fig. 13 Plots of kobs vs. [La(OTf)3] for the transesterification of 32 (2.02  10
5 M) in the low s s s pH regime at three values, T ¼ 25 1C. s pH ¼ 5.0 (&), 5.6 (’) and 6.4 (K). Lines through the data computed from fits to Equation (5) of ref. 81. Reproduced from ref. 81 with permission.

aqueous studies of the La3+-catalyzed hydrolysis.82 At low ss pH between 4.6 and 7.6, one observes that plots of the kobs vs. [La3+] for the methanolysis of 32 exhibit very pronounced saturation curves as shown in Fig. 13 indicative of strong binding to the metal ion. Fitting of the data to a strong binding equation83 gives the pseudofirst-order rate constants (kcat) at each ss pH as well as lower limits for the Kd dissociation constant of 10
6 mol dm
3. The binding in methanol is at least 104-fold larger than the reported value of 1.4  10
2 mol dm
3 for La3+ and 32 in aqueous solution at pH 6.8582 and is a consequence of the reduced dielectric constant of the alcoholic medium which strongly favors ion association. The plot of the kobs data vs. ss pH (not shown) is linear (slope ¼ 0.9670.07) and yields a value of 7 3
1
1 kOMe s for the second-order rate constant for OCH
cat ¼ 2.65  10 dm mol 3 pro3+ moted cyclization of La -coordinated 32. As remarkable as this number seems, the actual mechanism is more complicated and, as is demonstrated by the data shown in Fig. 14, actually proceeds through dimers which are simply formulated as (La3+(32))2. Fitting of the kobs vs. [La3+:32] kinetic data to a standard one-site binding model84 gives values for the dimerization constant (Kdimer) and maximum 3 3
1
4
1 catalytic rate constant (kmax s cat ) of (9.170.5)  10 dm mol ; (3.3070.07)  10 3 3
1
2
1 s s at s pH 5.0 and (7.170.9)  10 dm mol ; (1.0270.06)  10 s at s pH 6.7. Interestingly, the calculated saturation curves for the dimers shown in Fig. 14 indicate that at very low concentrations, the rate constant becomes negligible which indicates that the La3+:32 monomer is far-less reactive (at least 100-fold) than the (La3+(32))2 dimer. The fact that the binding constant does not change with ss pH indicates

312

R.S. BROWN AND A.A. NEVEROV 80

5

70 4

50

3

40 2

30

104 kobs s-1

104 kobs s-1

60

20 1 10 0

0 0

1 2 3 104 ([NaHPNPP] = [La(OTf)3]), M

Fig. 14 Plots of observed pseudo-first-order rate constants for the methanolysis of increasing and equimolar [La3+] ¼ [32, HPNPP] at 25 1C and ss pH 5.0 (N,N-dimethylaniline buffer, &, right axis) or ss pH 6.7 (2,6-lutidine buffer, ’, left axis). Lines through the data computed from fits to a standard one-site binding model. Reproduced from ref. 81 with permission.

the dimerization does not require a methoxide, but since the log kmax value cat increases linearly with ss pH, the cyclization of the phosphate within the dimer requires 1 equivalent of (base) in a specific base-catalyzed process. The secondorder rate constant for the methoxide-induced cyclization within the dimer is 1.18  108 dm3 mol
1 s
1and relative to simple methoxide-promoted cyclization of 33 ((2.670.2)  10
3 dm3 mol
1 s
1) the dimer provides an acceleration of 4.6  1010-fold in the low ss pH regime. It is interesting that the rate enhancement we see in methanol is 3 million-fold larger than that for La3+:32 cyclization in water, for which a rate enhancement of 14,000-fold over the simple hydroxide-promoted cyclization has been reported.82 This points to a significant solvent effect of methanol favoring formation of reactive dimers with two metal ions, the dinuclear metal ion core being a known motif employed by Nature in promoting such reactions in the active sites of enzymes. While the acceleration afforded to the cyclization of 32 by La3+ in methanol is certainly spectacular, this is not a biologically relevant metal ion and its charge exceeds that of the natural metal ion Zn2+. Very recent investigations of Zn2+catalysis of the methanolysis and ethanolysis of 32 indicated that there were indeed interesting catalytic effects, and that the situation in pure ethanol is quite different.85 Shown in Figs 15 and 16 are plots of the pseudo-first-order rate constant (kobs) for ethanolysis and methanolysis of HPNPP (32) as a function of [Zn2+]total when the [–OR]/[Zn2+] ratio is 0.5. This ratio was chosen to buffer the system at the half neutralization ss pH of 7 in ethanol8 and 9.5 in methanol at [Zn2+]total ¼ 1–2 mM7

METAL-CATALYZED ALCOHOLYSIS

313

kobs, s-1

2.0×10-2

1.0×10-2

0 0

2.0×10-4

4.0×10-4 6.0×10-4 [Zn2+]t, M

8.0×10-4

Fig. 15 Plot of kobs vs. [Zn2+]total for the decomposition of 32 (8  10
6 mol dm
3) in anhydrous ethanol at [
OEt]/[Zn2+]total ¼ 0.5. Fitting the data to a universal binding equation gives an apparent dissociation constant Kd for Zn2+:32 of (6.270.1)  10
5 mol dm
3 and a kmax of 1.88  10
2 s
1; r2 ¼ 0.9811. Reproduced with permission from ref. 85.

3.5×10-3 3.0×10-3

kobs, s-1

2.5×10-3 2.0×10-3 1.5×10-3 1.0×10-3 5.0×10-4 0 0

1.0×10-3 2.0×10-3 3.0×10-3 4.0×10-3 5.0×10-3 [Zn2+], M

Fig. 16 A plot of kobs vs. [Zn2+]total for the cyclization of 32 (4  10
5 M) in anhydrous methanol at [
OMe]/[Zn2+]total ¼ 0.5. Fitting the data to a universal binding equation gives an apparent dissociation constant Kd for Zn2+:32 of (4.2770.03)  10
4 mol dm
3 and a kmax of 3.52  10
3 s
1; r2 ¼ 0.9967. Reproduced with permission from ref. 85.

and in each case the active Zn2+-species contains 1 equivalent of alkoxide. The data appear to exhibit saturation behavior and can be fit by the universal binding equation83,85 where Kd refers to the dissociation constant for Zn2+:32 Zn2++32. Although the fits to the data shown on Figs 15 and 16 appear satisfactory, we know the situation is more complicated than can be assessed by this simple one-site binding model as we need to consider also the various equilibria of Zn2+-association with alkoxide and its oligomerization that produces dimers and higher-order

314

R.S. BROWN AND A.A. NEVEROV

aggregates. Nevertheless, what can be established is that at [Zn2+]total 4 0.4 mmol dm
3 and 5 mmol dm
3 in ethanol and methanol, respectively, essentially all the HPNPP is complexed to Zn2+. This binding is very much stronger than seen in water where it has been reported38b that Zn2+-complexes bind only weakly to phosphate diesters, the reported Kb values being o0.5 dm3 mol
1. The Zn2+-catalyzed process in ethanol is more complicated than in methanol because there is evidence for a higher than first-order dependence on [32]. For example, if we determine the kobs for the reaction starting with an initial [32] ¼ [Zn2+:0.5(NaOEt)] ¼ 0.4 mM, concentrations where all the phosphate is bound as a 1:1 complex, and then increase the concentration of each keeping the [32]/[Zn2+:0.5(NaOEt)] ratio at unity there is additional saturation shown in Fig. 17 indicative of a process where two (Zn2+:32) complexes come together to form a kinetically active dimeric species (Zn2+:32)2. This behavior is not seen in analogous experiments with 32 and Zn2+:0.5(NaOMe) in methanol, nor is it reported in water86 suggesting that those solvents do not engender the formation of kinetically active spontaneously formed dimers, at least at the concentrations employed. The similarity in the La3+-promoted cleavage of 32 in methanol that was reported above to what is observed with Zn2+ in ethanol suggests that the reactive dimers are probably doubly activated as (Zn2+:32)2 as shown in Scheme 8 where the actual catalytic role of the second phosphate in the dimer is to act as a template in stabilizing the dinuclear Zn2+-core. If the template role is operative, any phosphate

kobs, s-1

2

1

0 0.0000

0.0005

0.0010

0.0015

0.0020

0.0025

[Zn2+]bound (M)

Fig. 17 A plot of kobs vs. increasing [Zn2+]bound for the cyclization of 32 in anhydrous ethanol with [HPNPP]/[Zn2+]total/NaOEt ¼ 1:1:0.5. The [Zn2+]bound data used are corrected for the amount of unbound Zn2+ using the stability constant Kd of (6.270.1)  10
5 M as described in ref. The dissociation constant Kdimer for the presumed (32:Zn2+)2 complex is (6.270.5)  10
4 M and the maximum kmax for reaction of the dimeric complex is cat 2.9270.06 s
1, r2 ¼ 0.9586. The measured ss pH at 2 mM is 7.1. Reproduced with permission from ref. 85.

METAL-CATALYZED ALCOHOLYSIS O OAr H O P OEt -O Zn2+ +

HO 32:Zn

2+

315

O

Kd

O O P OAr OEt -O 32:Zn+(-OEt) Zn2+ HO

H P OAr OEt O -O 2+ Zn Zn2+ EtO- -O O P ArO O OH HO

O

O-

H P OAr OEt O -O

H Zn2+ Zn2+ O -O O Et P ArO O - HOAr OH

kcatOEt O O O 2+

HOAr

H OEt

P -O

Zn

+ EtO- -O

Zn P O

O

O

2+

O kcat2OEt

O

P O

H O Et

2+

-O

Zn

Zn2+

-

-O O P ArO O

OEt OH

Scheme 8 Zn2+(
OEt) promoted cleavage of 32 where two of the initially formed 32:Zn2+(
OEt) complexes dimerize and subsequently undergo and intramolecular cyclization.

monoanion should be capable of performing this role which is verified by the observations85 that low concentrations of diphenyl phosphate (1  10
6
2.5  10
4 M) exert a profound accelerating effect on the cyclization reaction of (Zn2+:32), the second-order rate constant being 790 dm3 mol
1 s
1 which is far too fast for any general base role. There are additional considerations that cast some important light on the chemical process proposed in Scheme 8. First, although the kinetics for appearance of p-nitrophenol are strictly first order, each dimer releases two equivalents of p-nitrophenol in sequential steps, so the second release must be at least as fast or faster than OEt the first, i.e. kOEt cat pkcat2. The phosphate products of the HPNPP cyclization and opening with ethoxide can also fulfill a template role so that in principle the product of the reaction actually serves to catalyze the loss of HPNPP by maintaining the dimeric form of the catalyst. Second, the observed pseudo-first-order rate constant
1 (kmax at ss pH 7.1 at 2 mM [Zn2+:0.5(NaOEt)]. Consideration of the rates cat ) is 2.9 s excludes an external –OEt acting as a base since the observed process is at least 100 times faster than permitted by a diffusion limited abstraction of the 2-hydroxypropyl hydrogen by free (ethoxide) at the concentrations set at ss pH 7.1.87 The lack of general base assistance to the cyclization of 1 was noted before in processes not nearly as fast as that presented here on the bases of: (1) no observed general catalysis of cyclization of 32 by the strong base piperidine in the presence of metal ions;79 (2) no observed primary deuterium kinetic isotope effect on the cyclization of 32 when catalyzed by a dinuclear Zn2+-complex;75 and (3) no observed buffer catalysis of the cyclization of 32 promoted by a mononuclear Zn2+-complex where a specific base-catalyzed process was proposed.64 Thus, the proposed mechanism in Scheme 8

316

R.S. BROWN AND A.A. NEVEROV

involves two relatively non-reactive HPNPP:Zn2+ complexes (one suggested to have a coordinated ethoxide or the kinetic equivalent of an O-deprotonated hydroxypropyl group and another to have coordinated ethanols). These associate to form a reactive (HPNPP:Zn2+)2 dimer with a Zn2+-coordinated ethoxide (or its kinetic equivalent) which acts as an internal base to deprotonate the HPNPP. This process would be facilitated by the dimeric nature of the complex where the large positive charge on the Zn2+-core stabilizes the anionic deprotonated propanolate75 which subsequently cyclizes with the expulsion of p-nitrophenolate. The latter in turn acquires a proton to recreate the Zn2+-
OEt in the nascent complex. The second equivalent of p-nitrophenol is released from the complex in a similar two-step pathway summarized as kOEt cat2. ZN2+ LIGAND MODELS FOR DINUCLEAR ENZYMES PROMOTING THE CLEAVAGE OF RNA

As stated above, Zn2+ in alcohol in the presence of alkoxide forms dimers and oligomers which can sometimes complicate the kinetic analysis, particularly in the cases of weakly binding substrates. However, complexation to the triaza ligand 938 simplifies the speciation as only 9:Zn2+(HOR) and its deprotonated form 9:Zn2+(
OR) (where RQCH3, CH2CH3) are present in solution. The cyclization of 32 mediated by 9:Zn2+(
OH) in water has been reported88 to have a second-order rate constant of 0.018 dm3 mol
1 s
1which is about three times less than the rate constant for the hydroxide promoted cyclization (0.065 dm3 mol
1 s
1).72 However, when the reactions with 9:Zn2+(
OR) are conducted in methanol and in ethanol, additional facets of the reaction are revealed. Presented in Fig. 1885 is a plot of the kobs for the cyclization reaction of 32 in methanol vs. [9:Zn2+]total under conditions where the [
OCH3]/[Zn2+]total ¼ 0.5 which sets the measured ss pH at 9.2–9.3. The plot shows upward curvature consistent with a process bimolecular in [9:Zn2+]total. Non-linear least squares (NLLSQ) fitting of the data to the expression kobs ¼ k1 [9:Zn2+]total+ 3 kobs [9:Zn2+]2total, gives k1 ¼ 18.9 dm3 mol
1 s
1 and kobs 3 3 ¼ (1.870.4)  10 3
1 2
1 (dm mol ) s , where the best fit line through the data is shown in Fig. 18. Similar upward curving plots are found when the [
OCH3]/[Zn2+]total ratio is 0.3 and 0.7, which is consistent with the participation of both the (CH3O
)Zn2+:9 and (CH3OH)Zn2+:9 forms of the catalyst, but when the ratio is 1.0 (ss pH 11.09), the plot (not shown) is linear, indicative of the involvement of only one molecule of the 3
1
1 (CH3O
)Zn2+:9 form, with a gradient kobs s . In no case do 2 ¼ 16.970.7 dm mol we observe evidence of saturation binding in methanol up to a [9:Zn2+]total of 5  10
3 mol dm
3, so the Kdis for dissociation of any 9:Zn2+:32 complex must be at least five times higher. The situation in ethanol is similar but the upward curvature is manifested at lower [9:Zn2+] as demonstrated in Fig. 18 by the plot of kobs for the cyclization of 32 vs. [9:Zn2+]total under conditions where the [
OEt]/[Zn2+]total ¼ 0.5 which sets the measured ss pH at 8.170.2. At progressively higher concentration this plot85 shows a downward curvature suggestive of a saturation phenomenon which is not seen in

METAL-CATALYZED ALCOHOLYSIS

317

0.15

0.30

0.25

0.15

0.05

kobs, s-1

0.20

kobs, s-1

0.10

0.10

0.05

0.00

0.00 0

1

2

3

4

5

6

[9:Zn2+]total

Fig. 18 Plot of kobs vs. total [9:Zn2+]total for the cyclization of 32 (8  10
5 dm mol
1) in ethanol (’, left y-axis) and methanol (J, right y-axis) under conditions where the [
OR]/ [Zn2+]total is 0.5. Dotted line corresponds to process first-order in [9:Zn2+], while curved line is the fit in methanol to the expression kobs ¼ k1 [9:Zn2+] + kobs [9:Zn2+]2. Heavy solid line 3 through the solid squares is the fit of the data to Equation (34). Reproduced with permission from ref. 85.

methanol nor water at these concentrations. That this behavior does not come from an aggregation of the 9:Zn2+ species or from inhibition by triflate counterion is verified by the fact that the plot of the kobs for ethanolysis of the weak-binding substrate p-nitrophenyl acetate (2) at [
OEt]/[Zn2+]total ¼ 0.5 is linear throughout the concentration range of 0.1 and 6 mmol dm
3 with a gradient of 0.8170.01 dm3 mol
1 s
1. Additional studies indicated that the active species is (EtO
)Zn2+:989 which is consistent the situation in methanol36 where the secondorder rate constant for (CH3O
)Zn2+:9 promoted methanolysis of paraoxon is 0.84 dm3 mol
1 s
1. Overall the behavior in both methanol and ethanol is consistent with a process where there are two 9:Zn2+ complexes in the transition state for methanolysis of HPNPP given in Scheme 9. Notably, a sigmoidal plot was observed before for the hydrolysis of ethyl p-nitrophenyl phosphate catalyzed by Co(III)–cyclen (cyclen ¼ 1,4,7,10-tetraazacyclododecane) and analyzed in terms of a process analogous to that in Scheme 9.90 Since the methanol data show the bimolecular process involves a k1 process that is first-order in 9:Zn2+, and another process dependent on [(CH3O
)Zn2+:9] and [(CH3OH)Zn2+:9], we can speculate that, for the k2 process, one of the units binds the substrate, and the other serves to deprotonate it perhaps leading to a ternary complex of 32
:[(CH3OH)Zn2+:9]2. For technical reasons relating to the inability to stabilize the ss pH at a [
OEt]/[Zn2+]total ¼ 1.0, we have not specifically tested to see whether the bimolecular behavior in ethanol depends on the

318

R.S. BROWN AND A.A. NEVEROV

2+

2+

-

32 + (Zn :9(HOEt)) + (Zn :9( OEt))

Kb

32:(Zn2+:9(HOEt):(Zn2+:9(-OEt) 32-:(Zn2+:9(HOEt):(Zn2+:9(HOEt)

k1

kcat

P

P

Scheme 9 A minimal scheme for the reaction of 32 with 9: Zn2+(
OEt) and 9: Zn2+(HOEt) accounting for the reaction kinetics which are both first order and second order in [catalyst].

presence of both [(EtO
)Zn2+:9] and [(EtO
)Zn2+:9] but based on the results in methanol, there is no reason to suggest it does not. Fitting of the Fig. 18 data to the expression derived for the process of Scheme 9 in Equation (33) where [A] and [B] are [(EtO
)Zn2+:9] and [(EtO
)Zn2+:9] (equal concentrations) yields k1 ¼ 3.69 dm3 mol
1 s
1, Kb ¼ 2.85  10
8 (dm3 mol
1)2 and kcat for the fully bound complex ¼ 0.13 s
1. kobs ¼ kcat K b ½A½B=ð1 þ K b ½A½BÞ þ k1 ½A

(33)

EXHALTED CATALYSIS OF METHANOLYSIS OF HPNPP PROMOTED BY A DINUCLEAR COMPLEX IN METHANOL

The interesting ability of ethanol and methanol to recruit two independent 9:Zn2+ complexes together in the transition state for an enhanced rate of cleavage of HPNPP suggests that tethering the CH3OH and CH3O
forms of the 9:Zn2+ moieties together would change a formally trimolecular process to a bimolecular one with a significant rate enhancement anticipated if the two forms react cooperatively. The dinucleating ligand 3591 was prepared by Kim and Lim92 who studied the catalytic effect of its bis-Zn2+-complex (35:2Zn(II)) for hydrolysis of two phosphate diesters. Unfortunately, in water this complex is only 1.2 and 4.4 times more effective than 9:Zn2+ in promoting the hydrolysis of bis-p-nitrophenyl phosphate and p-nitrophenyl phosphate at pH 7.92 It appears that this complex is not particularly stable in water and may lose a Zn2+-ion in solution and cannot form a stable 2:2 ligand:Zn2+-complex. This behavior is akin to what is observed with the Cu2+- and Zn2+-complexes of the analogous bis-9[ane]N3 ligand 36.93,94 On the other hand, this complex displays a remarkable activity for the cyclization of HPNPP in methanol.95 We have determined recently the structure of the dinuclear complex 35:2Zn(II) by H

N 2+ N Zn

N 2+ N Zn

N

N

H

H 35:2Zn(II)

H N 2+ Cu N N

H 2 H

H -Cu2+

2+ N N Cu N

(36:Cu(II))2 H

36:2Cu(II)

X-ray diffraction.96 The structure, shown in Fig. 19, has the two [12]aneN3 rings facing each other and binding two Zn2+-ions which are bridged by a single

METAL-CATALYZED ALCOHOLYSIS

319

Fig. 19 The structure of 35:2Zn(II):OH
(CF3SO
3 )3:CH3OH as determined by X-ray diffraction. For clarity, the triflate counter ions and methanol solvate are not shown.

hydroxide anion: also in the unit cell are three triflates for charge neutrality along with a single molecule of methanol. We assume that the lyoxide-bridged form is the one that is present when the catalytically active form is generated in situ through the sequential addition of 1 equivalent of 35 and NaOCH3, followed by 2 equivalent of Zn(OTf)2. It is important to note that the in situ formation of the active complex takes about 40–50 min, as judged by the fact that the catalytic activity of the solution continues to rise for that period of time, after which it assumes a constant value for several hours. Such an unusually slow complex formation can be attributed to initial formation of the mononuclear complex 35:Zn(II) where Zn2+ ion is ‘sandwiched’ between two triazamacrocycles within a single ligand molecule as was reported for the 36:Zn(II) complex.94 Intramolecular dissociation of such a complex and coordination of the second Zn2+-ion is probably responsible for the relatively slow formation of the thermodynamically stable and kinetically active di Zn(II)-complex which we formulate as 35:2Zn(II):(
OCH3). Shown in Fig. 20 is a plot of kobs for the cyclization of 32 vs. [CH3O
] ¼ [35: 2Zn(II)] determined at a measured ss pH ¼ 9.570.1. The primary data (dashed line) show downward curvature, although this not due to HPNPP binding but rather to a specific ion effect of the triflate anions that suppresses the rate (since each 35:2Zn(II)-complex carries with it 4
OTf counterions). The methanolysis of paraoxon (1) as a function of [35:2Zn(II):(
OCH3)] which also exhibits a downward curvature as shown in Fig. 21. That the weakly binding substrate (1) exhibits the identical curvature in the plot as does the potentially strong binding 32 suggests that the curvature is independent of the substrate, and is more likely dependent on the presence of anions that accompany the Zn2+-complex. Indeed, we have observed

320

R.S. BROWN AND A.A. NEVEROV 300

kobs,s-1

200

100

0 0.00

0.25

0.50

0.75

1.00

1.25

1.50

[CH3O-]=[35:2Zn(II)], mmol dm-3

Fig. 20 A plot of the observed pseudo-first-order rate constants (kobs) for the methanolysis of HPNPP (4  10
5 mol dm
3) as a function of [35:2Zn(II)] in the presence of 1 equivalent of added CH3O
per complex giving ss pH ¼ 9.5, T ¼ 2570.1 1C. Dotted line is presented as a visual aid directed through all actual data collected at 280 nm (&) or 320 nm (J) which are the wavelengths for disappearance of HPNPP and appearance of p-nitrophenol; solid line is a linear fit of the data corrected for inhibition by triflate counterions at 280 nm (’) or 320 nm (K). Reproduced with permission from ref. 95.

4.5 4.0 104kobs, s-1

3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.0

0.5

1.0

1.5

2.0

[CH3O-]=[35:2Zn(II)], mmol dm-3

Fig. 21 A plot of the observed pseudo-first-order rate constants (kobs) for methanolysis of 2.5  10
5 M paraoxon vs. [35:2Zn(II)] in the presence of 1 equivalent of added CH3O
per complex, ss pH ¼ 9.5, T ¼ 2570.1 1C. Dotted line is presented as a visual aid directed through all actual data (J); solid line is a linear fit through the data (K) corrected for inhibition by triflate counterions. Reproduced with permission from ref. 95.

that the catalytic activity toward methanolysis of paraoxon of a solution containing 0.4 mM of the [(CH3O
):Zn2([12]aneN3)2] complex does decrease as a function of increasing [Bu4N+(
OTf)]. The kinetic data for this inhibition, when fit to a one-site binding model shown in Equation (34) give an inhibition constant of 14.9 mM. This

METAL-CATALYZED ALCOHOLYSIS

321

constant was used to calculate the free kobs ¼ kmax K inhib =ð½Bu4 NOTf þ K inhib Þ

(34)

[catalyst] under the kinetic conditions and when the kinetic data of Figs 20 and 21 are corrected for triflate inhibition by plotting the kobs vs. free [catalyst], linear correlations are observed, consistent with no saturation binding of either substrate. The gradient of the linear plot in Fig. 21 for paraoxon is 0.31470.006 dm3 mol
1 s
1, while that in Fig. 20 is (2.7570.10)  105 dm3 mol
1 s
1. It is important to note that the latter rate constant is the largest second-order rate constant reported for the cyclization of HPNPP by any catalytic system reported to date. It is evident that there is a very large cooperative effect for the two Zn2+-ions which is brought about by their complexation to 35 and the medium effect in methanol, since the catalyst is 1.1  108-fold more reactive toward HPNPP than is methoxide (kMeO
¼ 2.56  10
3 dm3 mol
1s
1).81 A 1 mM solution of the 2
35:2Zn(II):( OCH3) catalyst at ss pH of 9.5 provides a t1/2 for decomposition of HPNPP of 2.5 ms, providing a remarkable acceleration of 2  1012 over the background methoxide reaction at that ss pH. Furthermore, the exalted catalytic ability of 35:2Zn(II):(
OCH3) is not limited to cyclization of HPNPP, but is also evident for the cleavage of a DNA model, methyl p-nitrophenyl phosphate (MNPP) The kinetic data are shown graphically in Fig. 22 which is analyzed95 to afford a Kb value of 0.37 mmol dm
3 and a kmax value of 4.1  10
2 s
1 for decomposition of the MNPP:35:2Zn(II):(
OCH3) complex at ss pH 9.5. Relative to the background methoxide reaction with MNPP (kOMe ¼ (7.970.6)  10
7 dm3 mol
1s
1) this constitutes 2 12 an acceleration of 10 -fold at that ss pH. In terms of comparison of the second-order

102kobs, s-1

3

2

1

0 0.00

0.25

0.50

0.75

1.00

1.25

[35:2Zn(II):(-OCH3)], mmol dm-3

Fig. 22 A plot of kobs for methanolysis of 4  10
5 M methyl p-nitrophenyl phosphate (MNPP) vs. [35:2Zn(II)] in the presence of 1 equivalent of CH3O
per ligand showing a saturation behavior, ss pH ¼ 9.5, T ¼ 2570.1 1C. Line through the data calculated by NLLSQ fits to a Michaelis–Mentin equation corrected for complex dissociation95 giving a binding constant of KM ¼ 0.37 mmol dm
3 and a maximum rate constant for reaction of the MNPP:[(CH3O–):35:2Zn(II)] complex of kmax ¼ (4.170.1)  10
2 s
1. Reproduced with permission from ref. 95.

322

R.S. BROWN AND A.A. NEVEROV

rate constants, defined as kmax/Kb/kOMe , the 35:2Zn(II):(
OCH3) catalyst is 2 8 1.4  10 -fold better than methoxide for promoting the methanolysis of MNPP. These accelerations exceed any previously reported catalysis of HPNPP cyclization or phosphate diester cleavage by several orders of magnitude, now approaching the realm of enzymatic catalysis. It is important to emphasize that this sort of catalysis is not seen with dinuclear Zn2+-containing catalysts in water, including 35:2Zn(II),92 and must be a consequence of a medium effect that optimizes substrate binding and catalytic transformation within a substrate/catalyst complex. Several dinucleating complexes76,97 of [9]aneN3 are known, but their Zn2+-complexes in water are, relative to the situation we observe here, only weakly catalytic toward the hydrolysis of phosphate diesters and toward the cyclization of HPNPP. Richard and Morrow76 have evaluated several of these and determined that only the di Zn2+-complex of 1,3-bis-N1(1,3,7-triazacyclononyl)propan-2-ol (37) had appreciable activity (120-fold) over its Zn:[9]aneN3 mononuclear analogue. Several other bis-complexes of the triazacyclononyl system showed only a 3–5-fold larger activity than the mononuclear complex. The di Zn2+-complex 38 is reported97 to promote the hydrolysis of bis-p-nitrophenyl phosphate (BNPP) with observed N Zn2+ N N

O

N Zn N

37

2+

N

N Zn2+ N N

N Zn2+ N N N N 38

H

saturation kinetics at pH 9.2, 35 1C and a Kb and kcat of 1.2  10
2 mol dm
3 and 2.24  10
6 s
1 for an overall second-order rate constant of kcat/Kb ¼ 1.87  10
4 dm3 mol
1 s
1. However, this value is only 7.5 times that of hydroxide98 indicating that very little exhalted catalysis is evident in water. Although HPNPP and MNPP are two closely related phosphate diesters in terms of charge, substitution and size, they exhibit far different appearances for the kobs vs. [35:2Zn(II)] plots as is evidenced by the data in Figs 20 and 22. Of course HPNPP is far more reactive than MNPP, but it should still require binding to the catalyst to achieve the very large accelerations observed for its cleavage by the dinuclear complex. The slower-reacting phosphate diester follows the expected pattern of saturation binding followed by a rate-limiting phosphate cleavage within the complex, but there is no visible saturation kinetics observed with HPNPP. The simplest explanation of the lack of observed saturation behavior in Fig. 20 is that the chemical step of cyclization of HPNPP is not rate-limiting, but some prior step is, such as productive binding. Shown in Scheme 10 is a working model for the binding and reaction of phosphate diesters promoted by 35:2Zn(II). Our current explanation is that there are two binding events, a rapid and reversible first binding event, perhaps to form a transient complex with a single Zn2+-ion in the complex, followed by a rearrangement to form a doubly activated substrate, suggested to involve coordination of both Zn2+-ions to the phosphate. The latter mode of coordination has been shown to optimize metal ion catalysis of the cleavage of phosphate diesters2b and was also proposed in the La3+-catalyzed cleavage of (La3+:HPNPP)281 and the

METAL-CATALYZED ALCOHOLYSIS

S+

k1

Zn 35 Zn

S

k -1

Zn

Zn 35

323 k3

Zn

k2

35

S Zn

k-2

CH3O-

Product + Zn2:35

Scheme 10 A proposed simplified scheme for the three step reaction of HPNPP or MNPP with 35:2Zn(II) with two initial binding steps followed by a chemical step which releases p-nitrophenol. 60

+ H+

5.0

+ -OCH3

kobs, s-1

40 30

2.5

20

102kobs, s-1

50

10 0

0.0 0

1

2

[CH3O-]/[Zn2([12]aneN3)2 ratio

Fig. 23 A plot of the observed pseudo-first-order rate constant for the methanolysis of 0.04 mM HPNPP (’, left axis) catalyzed by 0.2 mM 35:2Zn(II) or 0.04 mM methyl p-nitrophenyl phosphate (J, right axis) catalyzed by 0.4 mM 35:Zn(II) as a function of the [CH3O
]/ [35:Zn(II)] ratio at 2570.1 1C. Experiments done by ‘pH jump’ method starting at a [CH3O
]/ [35:Zn(II)] ratio of 1.0 (vertical dashed line, ss pH ¼ 9:5) and adding acid (left) or base (right). Reproduced with permission from ref. 95.

Zn2+-catalyzed cleavage of (Zn2+:HPNPP)2 in ethanol discussed above.85 For the slower reacting MNPP, the chemical cleavage step (represented by k3 and requiring a methoxide which is probably coordinated to one or both of the metal ions in 35:2Zn(II)) is relatively slow, so that both the pre-equilibrium steps are established and typical Michaelis–Menten behavior is observed with saturation at higher [35:2Zn(II)]. On the other hand, with the far more reactive HPNPP the chemical cyclization step, k3, is proposed to be faster than the k
2 step in the concentration range of 35:2Zn(II) used here. In this event, the observed kinetics would be linear in [35:2Zn(II)] as is the case in Fig. 20, with kobs ¼ k1[35:2Zn(II)]k2/(k
1+k2). This mechanism is also consistent with the base dependent behavior of the kinetics demonstrated in Fig. 23 for the reaction of MNPP and HPNPP catalyzed by 35:2Zn(II). The unusual increase in rate constant for the HPNPP reaction that accompanies addition of acid is not consistent with the cyclization step (k3) being rate-limiting. Rather, the increase in rate with added acid is more consistent with a process depending on binding the HPNPP to a greater amount of a complex devoid of an associated methoxide with its higher net positive charge attracting the negatively charged HPNPP. This requires that the chemical step of methoxide-dependent cyclization would be faster than the rearrangement step throughout most of the

324

R.S. BROWN AND A.A. NEVEROV

plot. At some point where the ss pH falls below a critical value and the complex contains little or no methoxide (or deprotonated HPNPP), the cyclization step slows and becomes rate-limiting which accounts for the discontinuity in the plot at the low [CH3O
]/[ 35:Zn(II)] ratio. The simplified process described in Scheme 9 introduces another aspect of the medium effect that accelerates the chemical step of HPNPP cleavage to the extent that it is no longer rate-limiting with this catalyst. The reaction seems to be limited by an aspect of substrate binding, a phenomenon often observed in enzymatic systems where the catalytic steps are competitive with binding steps. While the secondorder rate constant of 2.75  105 dm3 mol
1 s
1 for the HPNPP reaction seems somewhat slow for a binding step involving simple ligand exchange on Zn2+ (known to be 107–108 dm3 mol
1 s
1), it is well within the range for more complex ions in which an internal rearrangement or breaking of intramolecular hydrogen bonds occurs, for example in salicylate (
O2C–Ar–OH similar to
O2P(OR0 )O–R–OH in the HPNPP anion) binding to Zn2+ where the rate constant for ligand exchange in water is reported to be 1.4  105 dm3 mol
1 s
1.99 Of course, this thesis remains to be tested and probably can be confirmed or rejected through study of a series of close derivatives of HPNPP where the aryloxy group is varied which is currently underway in our laboratories.100

7

Conclusions

Metal ion-catalyzed hydrolytic processes have been studied for a long time, and many interesting systems have been explored which give valuable information about catalysis. However, with very few exceptions the catalysis afforded by these systems in water is disappointing when compared with enzymatic systems where a metal ion cofactor activates a substrate and a nucleophilic or basic group in an acyl or phosphoryl transfer process. It has been noted that bulk water may not be a good medium to approximate the medium inside the active site of an enzyme where it is now known that the effective dielectric constants resemble those of organic solvents rather than water. Our early studies of metal-catalyzed acyl transfer reactions were predicated on the idea the a reduced polarity/dielectric constant medium would allow one to better realize the catalytic potential of the metal ion by reducing the tightness of the solvation shell around the metal ion and its constituents, as well as allow a stronger interaction energy with substrate. The first demonstrations of the pronounced catalysis of acyl transfers were from activated, and not particularly challenging, substrates such as acetyl imidazole which led to important understandings of the metal species in solution.10 These formed the basis for several additional studies of metal-catalyzed transesterifications of a wide variety of neutral and unactivated carboxylate esters and aryloxy phosphorus-based esters for which metal-catalyzed hydrolytic reactions are notoriously inefficient. It now seems possible that this general area of study might have far-reaching implications for metal-catalyzed

METAL-CATALYZED ALCOHOLYSIS

325

transesterification processes and biological ones which involve metal ions as cofactors for acyl and phosphoryl transfer reactions. Despite a great deal of effort that has led to an increased understanding of how enzymatic catalysis might occur, it is generally held that ‘none of several models so far described approaches the enormous catalytic efficiency of natural enzymes’.88 It is a venerable hypothesis that one mode by which enzymes could achieve exhalted efficiency is to utilize low polarity active sites that are specially tailored for the catalytic task at hand. The above account indicates that very simple systems comprising metal complexes and a medium effect engendered by the lower alcohols does give rate accelerations for acyl and phosphoryl transfers approaching those of enzymes. It remains to be seen whether these systems will work for more biologically relevant substrates. It is our hope that the ideas contained in this report will spur further work using the multiple effects of structure and medium to bring us closer to understanding the ways in which Nature performs such transformations.

Acknowledgements The authors are grateful to the myriad of undergraduate and graduate students, postdoctoral and summer researchers whose names appear on the publications from this laboratory. In addition, they acknowledge the financial support of the Natural Sciences and Engineering Research Council of Canada, the Canada Foundation for Innovation, The United States Department of the Army, Army Research Office, Grant No. W911NF-04-1-0057 and the Defense Threat Reduction Agency, Joint Science and Technology Office (06012384BP) for financial support of this work. Finally, they are indebted to Professor J. P. Guthrie (University of Western Ontario) and Professor Andrew Williams (retired from the University Chemical Laboratories, Canterbury, England) for helpful discussions on various aspects of this work.

References 1. (a) Przystas, T.J. and Fife, T.H. (1990). J. Chem. Soc., Perkin Trans. 2, 393 and references there in; (b) Suh, J. (1992). Acc. Chem. Res. 25, 273 and references therein; (c) Tecilla, P., Tonellato, U., Veronese, F., Fellugg, F. and Scrimin, P. (1992). J. Org. Chem. 62, 7621; (d) Chin, J., Jubian, V. and Mrejen, K. (1990). J. Chem. Soc., Chem. Commun. 1326; (e) Meriwether, L. and Westheimer, F. (1956). J. Am. Chem. Soc. 78, 5119; (f) Grant, T.J. and Hay, R.W. (1965). Aust. J. Chem. 18, 1189; (g) Groves, J.T. and Dias, R.M. (1979). J. Am. Chem. Soc. 101, 1033; (h) Groves, J.T. and Chambers, R.R. (1984). J. Am. Chem. Soc. 106, 630; (i) Groves, J.T. and Olson, J.R (1985). Inorg. Chem. 24, 2715; (j) Sayre, L., Reddy, K.V., Jacobson, A.R. and Tang, W. (1992). Inorg. Chem. 31, 935; (k) Fife, T.H. (1993). Acc. Chem. Res. 26, 325; (l) Fife, T.H. and Bembi, R. (1993). J. Am. Chem. Soc. 115, 11358; (m) Parac, T.N. and Kostic´, N.M. (1996). J. Am. Chem. Soc. 118, 51; (n) Singh, L. and Ram, R.N. (1994). J. Org. Chem. 59, 710; (o) Sutton, P.A. and Buckingham, D.A. (1987). Acc. Chem. Res. 20, 357; (p) Kimura, E., Nakamura, I., Koike, T., Shionoya, M., Kodama, Y., Ikeda, T. and Shiro, M. (1994). J. Am. Chem. Soc. 116, 4764; (q) Parac, T.N., Ullman, G.M. and

326

2.

3.

4.

5.

R.S. BROWN AND A.A. NEVEROV Kostic´, N.M. (1999). J. Am. Chem. Soc. 121, 3127 and references therein; (r) Ridder, A. M. and Kellogg, R. M. (1996). Models for zinc-containing enzymes. In Comprehensive Supramolecular Chemistry Vol. 4: Supramolecular Reactivity and Transport: Bioorganic Systems, Murakami, Y. (ed.), Chapter 11, Elsevier Science Ltd., Oxford; (s) Chin, J. (1991). Acc. Chem. Res. 24, 145 and references therein (a) Blasko, A. and Bruice, T.C. (1999). Acc. Chem. Res. 32, 475 and references therein; (b) Williams, N.H., Tasaki, B., Wall, M. and Chin, J. (1999). Acc. Chem. Res. 32, 485 and references therein; (c) Moss, R.A., Park, B.D., Scrimin, P. and Ghirlanda, G. (1995). J. Chem. Soc., Chem. Comm. 1627; (d) Moss, R.A., Zhang, J. and Bracken, K. (1997). J. Chem. Soc., Chem. Comm. 1639; (e) Sumaoka, J., Miyama, S. and Komiyama, M. (1994). J. Chem. Soc., Chem. Comm. 1755; (f) Morrow, J.R., Buttrey, L.A. and Berback, K.A. (1992). Inorg. Chem. 31, 16; (g) Morrow, J.R., Buttrey, L.A., Shelton, V.M. and Berback, K.A. (1992). J. Am. Chem. Soc. 114, 1903; (h) Breslow, R. and Zhang, B. (1994). J. Am. Chem. Soc. 116, 7893; (i) Takeda, N., Irisawa, M. and Komiyama, M. (1994). J. Chem. Soc., Chem. Comm. 2773; (j) Hay, R.W. and Govan, N. (1990). J. Chem. Soc., Chem. Commun. 714; (k) Schneider, H.-J., Rammo, J. and Hettich, R. (1993). Angew. Chem. Int. Ed. Engl. 32, 1716; (l) Ragunathan, K.G. and Schneider, H.-J. (1996). Angew. Chem. Int. Ed. Engl. 35, 1219; (m) Go´mez-Tagle, P. and A.K. Yatsimirsky (1998). J. Chem. Soc., Dalton Trans. 2957; (n) Roigk, A., Hettich, R. and Schneider, H.-J. (1998). Inorg. Chem. 37, 751 and references therein; (o) Go´mez-Tagle, P. and Yatsimirski, A.K. (2001). J. Chem. Soc., Dalton Trans. 2663; (p) Jurek, P.E., Jurek, A.M. and Martell, A.E. (2000). Inorg. Chem. 39, 1016 (a) Parshall, G.W. and Ittel, S.D. (1992). Homogeneous Catalysis: The Applications and Chemistry of Catalysis by Soluble Transition Metal Complexes (2nd edn), pp. 269–296. Wiley Interscience, New York; (b) Stanton, M.G., Allen, C.B., Kissling, R.M., Lincoln, A.L. and Gagne´, M.R. (1998). J. Am. Chem. Soc. 120, 5981 and references therein; (c) Otera, J. (1993). Chem. Rev. 93, 1449; (d) Okano, T., Miyamoto, K. and Kiji, J. (1995). Chem. Lett. 246; (e) Ranu, B.C., Dutta, P. and Sarkar, A. (1998). J. Org. Chem. 63, 6027; (f) Baldini, L., Bracchini, C., Cacciapaglia, R., Casnati, A., Mandolini, L. and Ungaro, R. (2000). Chem. Eur. J. 6, 1322; (g) Cacciapaglia, R., Di Stephano, S., Kelderman, E. and Mandolini, L. (1999). Angew. Chem. Int. Ed. 38, 349; (h) Cacciapaglia, R., Mandolini, L., Arnecke, R., Bo¨hmer, V. and Vogt, W. (1998). J. Chem. Soc., Perkin Trans. 2, 419; (i) Breccia, P., Cacciapaglia, R., Mandolini, L. and Scorsini, C. (1998). J. Chem. Soc., Perkin Trans. 2, 1257; (j) Stevels, W.M., Ankone´, M.J.K., Dijkstra, P.J. and Feijen, J. (1996). Macromolecules 29, 8296; (k) Hinde, N.J. and Hall, C.D. (1998). J. Chem. Soc., Perkin Trans. 2, 1249 and references therein; (l) Otton, J., Ratton, S., Vasnev, V.A., Markova, G.D., Nametov, K.M., Bakhmutov, V.I., Komarova, L.I., Vinogradova, S.V. and Korshak, V.V. (1988). J. Polym. Sci., Part A: Polym. Chem. 26, 2199; (m) Cacciapaglia, R., Casnati, A., Mandolini, L., Reinhoudt, D.N., Salvio, R., Sartori, A. and Ungaro, R. (2005). J. Org. Chem. 70, 5398; (n) Cacciapaglia, R., Di Stefano, S., Fahrenkrug, F., Luening, U. and Mandolini, L. (2004). J. Phys. Org. Chem. 17, 350 (a) Balakrishnan, V.K., Dust, J.M., vanLoon, G.W. and Buncel, E. (2001). Can. J. Chem. 79, 157; (b) Nagelkerke, R., Pregel, M.J., Dunn, E.J., Thatcher, G.R.J. and Buncel, E. (1995). Org. React. 29, 11; (c) Dunn, E.J. and Buncel, E. (1989). Can. J. Chem. 67, 1440; (d) Dunn, E.J., Moir, R.Y., Buncel, E., Purdon, J.G. and Bannard, R.A.B. (1990). Can. J. Chem. 68, 1837; (e) Pregel, M.J. and Buncel, E. (1991). J. Chem. Soc., Perkin Trans. 2, 307; (f) Pregel, M.J., Dunn, E.J. and Buncel, E. (1991). J. Am. Chem. Soc. 113, 3545; (g) Pregel, M.J., Dunn, E.J., Nagelkerke, R., Thatcher, G.R.J. and Buncel, E. (1995). Chem. Soc. Rev. 449. Bates, R.G. (1969). Medium effects and pH in non-aqueous solvents. In Solute– Solvent Interactions, Coetzee, J.F. and Ritchie, C.D. (eds), 45. Marcel Dekker, New York

METAL-CATALYZED ALCOHOLYSIS

327

6. (a) Bosch, E., Rived, F., Roses, M. and Sales, J. (1999). J. Chem. Soc., Perkin Trans. 2, 1953; (b) Rived, F., Rose´s, M. and Bosch, E. (1998). Anal. Chim. Acta 374, 309; (c) Bosch, E., Bou, P., Allemann, H. and Rose´s, M. (1996). Anal. Chem. 68, 3651; (d) Canals, I., Portal, J.A., Bosch, E. and Rose´s, M. (2000). Anal. Chem. 72, 1802 7. Gibson, G.T.T., Neverov, A.A. and Brown, R.S. (2003). Can. J. Chem. 81, 495 8. Gibson, G.T.T., Mohamed, M.F., Neverov, A.A. and Brown, R.S. (2006). Inorg. Chem. 45, 7891 9. Harned, H.S. and Owen, B.B. (1957). The Physical Chemistry of Electrolytic Solutions. ACS Monograph Series 137 (3rd edn), p. 161. Reinhold Publishing, New York. 10. (a) Brown, R.S. and Neverov, A. (2002). J. Chem. Soc., Perkin Trans. 2, 1039; (b) Brown, R.S., Neverov, A.A., Tsang, J.S.W., Gibson, G.T.T. and Montoya-Pela´ez, P. (2004). J. Can. J. Chem. 82, 1791 11. (a) Lipscomb, W. and Stra¨ter, N. (1996). Chem. Rev. 35, 2024; (b) Wilcox, D.E. (1996). Chem. Rev. 96, 2435; (c) Coleman, J.E. (1998). Curr. Opin. Chem. Biol. 2, 222; (d) Parkin, G. (2004). Chem. Rev. 104, 699; (e) Cowan, J.A. (1998). Chem. Rev. 98, 1067; (f) Weston, J. (2005). Chem. Rev. 105, 2151 12. (a) Cleland, W.W., Frey, P.A. and Gerlt, J.A. (1998). J. Biol. Chem. 273, 25529; (b) Simonson, T., Carlsson, F. and Case, D.A. (2004). J. Am. Chem. Soc. 126, 4167 and references therein 13. (a) Crosby, J. and Lienhard, G.E. (1970). J. Am. Chem. Soc. 92, 5707; (b) Crosby, J., Stone, R. and Lienhard, G.E. (1970). J. Am. Chem. Soc. 92, 2891 14. Lowry, T.H. and Richardson, K.S. (1987). Mechanism and Theory in Organic Chemistry (3rd edn), pp. 361–367, 589–590, 600. Harper and Row, New York 15. Levine, I.N. (1978). Physical Chemistry (4th edn). pp. 276–281, McGraw-Hill, USA. 16. (a) Neverov, A.A., McDonald, T., Gibson, G. and Brown, R.S. (2001). Can. J. Chem. 79, 1704; (b) Neverov, A.A. and Brown, R.S. (2001). Inorg. Chem. 40, 3588; (c) Neverov, A.A. and Brown, R.S. (2000). Can. J. Chem. 78, 1247; (d) MontoyaPelaez, P.J. and Brown, R.S. (2002). Inorg. Chem. 41, 309; (e) Neverov, A.A., Gibson, G. and Brown, R.S. (2003). Inorg. Chem. 42, 228; (f) Neverov, A.A., Sunderland, N.E. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 65 17. (a) Tsang, J.S., Neverov, A.A. and Brown, R.S. (2003). J. Am. Chem. Soc. 125, 7602; (b) Tsang, J.S.W., Neverov, A.A. and Brown, R.S. (2004). Org. Biomol. Chem. 2, 3457; (c) Liu, T., Neverov, A.A., Tsang, J.S.W. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 1525; (d) Lu, Z.-L., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 3379; (e) Lewis, R.E., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 4082; (f) Maxwell, C., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 4329; (g) Melnychuk, S.A., Neverov, A.A. and Brown, R.S. (2006). Angew. Chem. Int. Ed. 45, 1767 18. (a) Amis, E.S. (1955). Anal. Chem. 27, 1672; (b) Amis, E.S. (1953). J. Chem. Educ. 30, 351 19. (a) Landskroener, P.A. and Laidler, K.J. (1954). Effects of Solvation of Hydrolysis. Ph.D Dissertation, Catholic University of America, Catholic University of America Press: Washington; (b) Laidler, K.J. and Landskroener, P.A. (1956). Trans. Faraday Soc. 52, 200 20. Martell, A.E. and Motekaitis, R.J. (1988). Determination and Use of Stability Constants. VCH Publishers, New York 21. (a) Moeller, T., Martin, D.F., Thompson, L.C., Ferru´s, R., Feistel, G.R. and Randall, W. (1965). J. Chem. Rev. 65, 1; (b) Burgess, J. (1978). Metal Ions in Solution. John Wiley and Sons, New York 22. (a) Fritz, J.S. (1973). Acid– Base Titrations in Nonaqueous Solvents. Allyn and Bacon, Boston, MA; (b) Huber, W. (1967). Titrations in Nonaqueous Solvents. Academic Press, New York; (c) Gyenes, I. (1967). Titration in Non-Aqueous Media. D. Van Nostrand Company, New Jersey

328

R.S. BROWN AND A.A. NEVEROV

23. deLigny, C.L. and Rehbach, M. (1960). Recl. Trav. Chim. 79, 727 24. Bates, R.G., Paabo, M. and Robinson, R.A. (1963). J. Phys. Chem. 67, 1833 25. (a) Grunwald, E. and Berkowitz, B.J. (1951). J. Am. Chem. Soc. 73, 4939; (b) Gutbezahl, B. and Grunwald, E. (1953). J. Am. Chem. Soc. 75, 559–565 26. (a) Bates, R.G. (1964). Determination of pH: Theory and Practice (2nd edn). p. 244, John Wiley and Sons, New York; (b) This value was calculated by Bates26a using the equation d ¼ E¯ j
logm gH taking values for logm gH and E¯ j (termed log l) from Grunwald25 27. Gans, P., Sabatini, A. and Vacca, A. (1996). Talanta 43, 1739 28. (a) Moeller, T., Martin, D.F., Thompson, L.C., Ferru´s, R., Feistel, G.R. and Randall, W.J. (1965). Chem. Rev. 65, 1; (b) Burgess, J. (1978). Metal Ions in Solution. John Wiley and Sons, New York 29. (a) Baes Jr., C.F. and Mesmer, R.E. (1976). The Hydrolysis of Cations. John Wiley and Sons, New York; (b) Martell, A.E. and Motekaitis, R.J. (1988). Determination and Use of Stability Constants. VCH Publishers, New York; (c) Martell, A.E. and Smith, R.M. (1974). Critical Stability Constants, Vols 1–6, Plenum Press, New York 30. Simms, H.S. (1926). J. Am. Chem. Soc. 48, 1239 31. Neverov, A.A., Montoya-Pela´ez, P. and Brown, R.S. (2001). J. Am. Chem. Soc. 123, 210 32. Montoya-Pela´ez, P., Gibson, G.T.T., Neverov, A.A. and Brown, R.S. (2003). Inorg. Chem. 42, 8624 33. Liu, T., Neverov, A.A. and Brown, R.S. unpublished observations 34. Given that the autoprotolysis constants of ethanol and methanol are 10
19.1 and 10
16.77, respectively, the background kobs values for the lyoxide-catalyzed ethanolysis and methanolysis reactions of paraoxon at ss pH 7.3 (ethanol) and 8.3 (methanol) are 8.1  10
15 and 3.5  10
11 s–1, respectively 35. Okano, T., Miyamoto, K. and Kiji, J. (1995). Chem. Lett. 246 reported the attempted transesterification of various esters promoted by Ln(OiPr)3 alkoxides which showed exceedingly poor activity even at higher temperatures and prolonged reaction times. From our work, the trialkoxy forms would be the least active, while the monoalkoxy forms would be the most active 36. Desloges, W., Neverov, A.A. and Brown, R.S. (2004). Inorg. Chem. 43, 6752 37. Neverov, A.A. and Brown, R.S. (2004). Org. Biomol. Chem. 2, 2245 38. (a) Kimura, E., Shiota, T., Koike, T., Shiro, M. and Kodama, M. (1990). J. Am. Chem. Soc. 112, 5805; (b) Koike, T. and Kimura, E. (1991). J. Am. Chem. Soc. 113, 8935 39. Smith, M.B. and March, J. Advanced Organic Chemistry (5th edn), pp. 338–342. Wiley Interscience, New York, and references therein 40. Lu, Z.-L., Neverov, A.A. and Brown, R.S. (2005). Org. Biomol. Chem. 3, 3379 41. Ba-Saif, S., Luthra, A.K. and Williams, A. (1987). J. Am. Chem. Soc. 109, 6362 42. (a) Palm, V.A. (1975). Tables of Rate and Equilibrium Constants of Heterocyclic Reactions (Suppl. Vol. 1 (1)), Proizbodstvenno-Izdatelckii Kombinat Biniti, Moscow; (b) Palm, V.A. (1985). Tables of Rate and Equilibrium Constants of Heterocyclic Reactions (Vol. 1, Issue 3). Tartuskii gosudarsvennii Universitet, Tartu 43. (a) Taft Jr., R.W. (1952). J. Am. Chem. Soc. 74, 2729–3120; (b) Taft Jr., R.W. (1953). J. Am. Chem. Soc. 75, 4231 44. Di Stephano, S., Leonelli, F., Garofolo, B., Mandolini, L., Bettolo, B. and Migneco, M. (2002). Org. Lett. 4, 2783 45. Hong, S.-B. and Rauchel, F.M. (1996). Biochemistry 35, 10904 aquenous pKa values for Table 10 46. Clare, B.W., Cook, D., Ko, E.C.F., Mac, Y.C. and Parker, A.J. (1966). J. Am. Chem. Soc. 88, 1911 experimental pKa values for Table 10 47. (a) Chaw, Z.S., Fischer, A. and Happer, D.A.R. (1971). J. Chem. Soc. (B) 1818; (b) Kirsch, J., Clewell, A. and Simon, A. (1968). J. Org. Chem. 33, 127; (c) Humffray, A.A. and Ryan, J.J. (1967). J. Chem. Soc. (B) 468

METAL-CATALYZED ALCOHOLYSIS

329

48. (a) Ryan, J.J. and Humffray, A.A. (1966). J. Chem. Soc. (B) 842; (b) Bruice, T.C. and Mayahi, M. (1960). J. Am. Chem. Soc. 82, 3067; (c) Kirsch, J.F. and Jencks, W.P. (1964). J. Am. Chem. Soc. 86, 837 49. Mitton, C.G., Schowen, R.L., Gresser, M. and Shapley, J. (1969). J. Am. Chem. Soc. 91, 2036 50. Ba-Saif, S.A. and Williams, A. (1988). J. Org. Chem. 53, 2204 51. Caldwell, S.R., Raushel, F.M., Weiss, P.M. and Cleland, W.W. (1991). Biochemistry 30, 7444 52. (a) Jencks, W.P. and Gilchrist, M. (1968). J. Am. Chem. Soc. 90, 2622; (b) Hupe, D.J. and Jencks, W.P. (1977). J. Am. Chem. Soc. 99, 451; (c) Sayer, J.M. and Jencks, W.P. (1977). J. Am. Chem. Soc. 99, 464; (d) Gresser, M.J. and Jencks, W.P. (1977). J. Am. Chem. Soc. 99, 6963 53. (a) Thea, S. and Williams, A. (1986). Chem. Soc. Rev. 15, 125; (b) Williams, A. (1984). Acc. Chem. Res. 17, 425; (c) Williams, A. (2000). Concerted Organic and Bio-Organic Mechanisms. CRC Press, Boca Raton 54. Given that the autoprotolysis constant of methanol is 10
16.77 (mol dm–3)2 (ref. 6) one can compute a ss pK a of methanol of 18.13 on the mol dm
3 scale 55. (a) Petrova, J., Momchilova, S., Haupt, E.T.K., Kopf, J. and Eggers, G. (2002). Phosphorus, Sulfur Silicon Relat. Elem. 177, 1337; (b) Peterman, D.R., Fox, R.V., and Rollins, H.W. (2002). Abstract of Papers, 223rd ACS National Meeting, Orlando, FL, April 7–11, NUCL-131; (c) Ferraro, J.R., Herlinger, A.W. and Chiarizia, R. (1998). Sol. Extract Ion Exch. 16, 775; (d) Du Preez, R. and Preston, J.S.S. (1986). S. Afr. J. Chem. 39, 89–137; (e) Peppard, D.F., Mason, G.W., Driscoll, W.J. and McCarty, S.J. (1959). J. Inorg. Nucl. Chem. 12, 141; (f) Lebedeva, E.N., Zaitseva, M.G., Galaktionova, O.V., Bystrov, L.V. and Korovin, S.S. (1981). Koordinatsionnaya Khimiya 7, 870 CAN 95:87058; (g) Galaktionova, O.V., Lebedeva, E.N., Yastrebov, V.V. and Korovin, S.S. (1980). Zh. Neorg. Khim. 25, 2660 CAN 93:226562; (h) Pyartman, A.K., Kopyrin, A.A., Puzikov, E.A. and Bogatov, K.B. (1996). Zh. Neorg. Khim. 41, 347 CAN 125:124905; (i) Pyartman, A.K., Kopyrin, A.A., Puzikov, E.A. and Bogatov, K.B. (1996). Zh. Neorg. Khim. 41, 686 CAN 126:11927; (j) Pyartman, A.K., Keskinov, V.A., Kovalev, S.V. and Kopyrin, A.A. (1997). Radiochemistry (Moscow) (Translation of Radiokhimiya) 39, 142 CAN 127:210889; AN 1997:471873 56. Kosky, C.A., Gayda, J.-P., Gibson, J.F., Jones, S.F. and Williams, D.J. (1982). Inorg. Chem. 21, 3173 57. Mikulski, C.M., Pytlewski, L.L. and Karagannis, N.M. (1979). Inorg. Chim. Acta 32, 263 58. Schurhammer, R., Erhart, V., Troxler, L. and Wipf, G. (1999). J. Chem. Soc., Perkin Trans. 2, 2423 and references therein. 59. Rackham, D.M. (1980). Spectrosc. Lett. 13, 513 60. (a) Williams, N.H., Cheung, W. and Chin, J. (1998). J. Am. Chem. Soc. 120, 8079; (b) Wahnon, D., Lebuis, A.-M. and Chin, J. (1995). Angew. Chem. Int. Ed. Engl. 34, 2412 61. Williams, N.H. and Wyman, P. (2001). J. Chem. Soc., Chem. Commun. 1268 62. (a) Blasko, A. and Bruice, T.C. (1999). Acc. Chem. Res. 32, 475 and references therein; (b) Williams, N.H., Tasaki, B., Wall, M. and Chin, J. (1999). Acc. Chem. Res. 32, 485 and references therein; (c) Moss, R.A., Park, B.D., Scrimin, P. and Ghirlanda, G. (1995). J. Chem. Soc., Chem. Commun. 1627; (d) Moss, R.A., Zhang, J. and Bracken, K. (1997). J. Chem. Soc., Chem. Commun. 1639; (e) Sumaoka, J., Miyama, S. and Komiyama, M. (1994). J. Chem. Soc., Chem. Commun. 1755; (f) Morrow, J.R., Buttrey, L.A. and Berback, K.A. (1992). Inorg. Chem. 31, 16; (g) Morrow, J.R., Buttrey, L.A., Shelton, V.M. and Berback, K.A. (1992). J. Am. Chem. Soc. 114, 1903; (h) Breslow, R. and Zhang, B. (1994). J. Am. Chem. Soc. 116, 7893; (i) Takeda, N., Irisawa, M. and Komiyama, M. (1994). J. Chem. Soc., Chem. Commun. 2773; (j) Hay, R.W. and

330

63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80.

81. 82. 83.

84.

85.

R.S. BROWN AND A.A. NEVEROV Govan, N. (1990). J. Chem. Soc., Chem. Commun. 714; (k) Schneider, H.-J., Rammo, J. and Hettich, R. (1993). Angew. Chem. Int. Ed. Engl. 32, 1716; (l) Ragunathan, K.G. and Schneider, H.-J. (1996). Angew. Chem. Int. Ed. Engl. 35, 1219; (m) Go´mez-Tagle, P. and Yatsimirsky, A.K. (1998). J. Chem. Soc., Dalton Trans. 2957; (n) Roigk, A., Hettich, R. and Schneider, H.-J. (1998). Inorg. Chem. 37, 751 and references therein; (o) Go´mez-Tagle, P. and Yatsimirski, A.K. (2001). J. Chem. Soc., Dalton Trans. 2663; (p) Jurek, P.E., Jurek, A.M. and Martell, A.E. (2000). Inorg. Chem. 39, 1016 (a) Hegg, E.L. and Burstyn, J.N. (1998). Coord. Chem. Rev. 173, 133; (b) Jancso´, A., Mikkola, S., Lo¨nnberg, H., Hegetschweiler, K. and Gadja, T. (2003). Chem. Eur. J. 9, 5404 Feng, G., Mareque-Rivas, J.C., Martin de Rosale´s, R.T. and Williams, N.H. (2006). J. Am. Chem. Soc. 127, 13470 Stra¨ter, N., Lipscomb, W.N., Klabunde, T. and Krebs, B. (1996). Angew. Chem. Int. Ed. Engl. 35, 2024 Davies, J.F., Hostomska, Z., Hostomsky, Z., Jordan, S.R. and Mathews, D.A. (1991). Science 252, 88 Beese, L.S. and Steitz, T.A. (1991). EMBO J. 10, 25 Lahm, A., Volbeda, S. and Suck, D. (1990). J. Mol. Biol. 215, 207 Molenveld, P., Engbertsen, J.F.J. and Reinhoudt, D.N. (2000). Chem. Soc. Rev. 29, 75 Mancin, F., Scrimin, P., Tecilla, P. and Tonellato, U. (2005). Chem. Commun. 2540 Morrow, J.R. and Iranzo, O. (2004). Curr. Opin. Chem. Biol. 8, 192 Yamada, K., Takahashi, Y.-i., Yamamura, H., Araki, A., Saito, K. and Kawai, M. (2000). Chem. Commun. 1315 Feng, G., Mareque-Rivas, J.C. and Williams, N.H. Chem. Commun. 1845 Mancin, F., Rampazzo, E., Tecilla, P. and Tonellato, U. Eur. J. Chem. 281 Yang, M.-Y., Iranzo, O., Richard, J.P. and Morrow, J.R. (2005). J. Am. Chem. Soc. 127, 1064 Iranzo, O., Elmer, T., Richard, J.P. and Morrow, J.R. (2003). Inorg. Chem. 42, 7737 Iranzo, O., Richard, J.P. and Morrow, J.R. (2004). Inorg. Chem. 43, 1743 Iranzo, O., Kovalevsky, A.Y., Morrow, J.R. and Richard, J.P. (2003). J. Am. Chem. Soc. 125, 1988 Brown, D.M. and Usher, D.A. J. Chem. Soc. 6558 The methoxide reaction of 33 in pure methanol has a second-order rate constant for methoxide promoted methanolysis of (2.5670.16)  10
3 M
1s
1, while the secondorder rate constant for attack of methoxide on methyl p-nitrophenyl phosphate ((7.970.6)  10
7 M
1s
1 at 25 1C), Neverov, A.A., Brown, R.S. (2001). Inorg. Chem. 40, 3588 Tsang, J.S., Neverov, A.A. and Brown, R.S. (2003). J. Am. Chem. Soc. 125, 1559 Morrow, J.R., Buttrey, L.A. and Berback, K.A. (1992). Inorg. Chem. 31, 16 A universal expression for both strong and weak binding scenarios is given below: kobs ¼ kcat ð1 þ K d  ½33 þ ½La3þ   K d X Þ=ð½33  ð2K d ÞÞ where Kd refers to the dissociation constant for La3+:33 La3++33; [33] and [La3+] are total concentrations and X is given as: X ¼ fð1 þ 2K d  ½1 þ 2  ½Zn2þ   K d þ K 2d  ½12
2  K 2d  ½Zn2þ ½1 þ ½Zn2þ 2  K 2d g0:5 . The equation was derived from the equations for equilibrium binding and for conservation of mass by using the commercially available MAPLE software, Maple V Release 5, Waterloo Maple Inc., Waterloo, Ontario, Canada At first glance the process described in Equation (6) is bimolecular in [La3+:33], but the kinetics strictly adhere to a first-order process for the loss of starting material. However, the subsequent observation that the products of the reaction actually catalyze the decomposition of starting material allow us to treat the kinetics at each ss pH according to 3+ a simple one-site binding model: kobs ¼ kmax :33]init/(Kd+[La3+:33]init) cat [La Liu, C.T., Neverov, A.A. and Brown, R.S. (2007). Inorg. Chem. 46, 1778

METAL-CATALYZED ALCOHOLYSIS

331

86. Breslow, R., Berger, D. and Huang, D.-L. (1990). J. Am. Chem. Soc. 112, 3686 report that 0.5 mM Zn2+ at pH 7 in water at 37 oC produces a pseudo-first-order rate constant of 1.71  10
2 h
1 (4.75  10
6 s
1) from which a second-order rate constant of 9.5  10
3 dm3 mol
1 s
1 can be calculated at that pH. 87. A diffusion limited proton abstraction would occur at kdif ¼ 1010 M
1 s
1, so at ss pH 7.1 where [
OEt] ¼ 10
12 the maximum pseudo-first-order rate constant would be
2
1 kmax s cat ¼ 10 88. Bonfa´, L., Gatos, M., Mancin, F., Tecilla, P. and Tonellato, U. (2003). Inorg. Chem. 42, 3943 89. Since the active species for the ethanolysis of 2 is 9: Z n2+(
OEt), the true-second-order rate constant for the reaction would be twice the gradient of the plot, or 1.62 dm3 mol
1 s
1 90. Rawlings, J., Cleland, W.W. and Hengge, A.C. (2003). Inorg. Biochem. 93, 61 91. Ligand 35 was originally prepared by Weisman, G.R., Vachon, D.J., Johnson, V. B., Gronbeck, D.A. (1987). J. C. S. Chem. Commun. 886, along with numerous others containing the [12]aneN3 and [9]aneN3 binding group 92. Kim, J. and Lim, H. (1999). Bull. Korean. Chem. Soc. 20, 491 93. Haidar, R.H., Ipek, M., Dasgupta, B., Yousef, M. and Zompa, L. (1997). J. Inorg. Chem. 36, 3125 94. Dasgupta, B., Haidar, R., Hsieh, W.-Y. and Zompa, L. (2000). J. Inorg. Chim. Acta 306, 78 95. Neverov, A.A., Lu, Z.L., Maxwell, C.I., Mohamed, M.F., White, C.J., Tsang, J.S.W. and Brown, R.S. (2006). J. Am. Chem. Soc. 128, 16398 96. Lu, Z.-L and Brown, R.S. To be published 97. Vichard, C. and Kaden, T.A. (2002). Inorg. Chim. Acta 337, 173 98. The second-order rate constant for hydrolysis of BNPP at 35 1C is reported to be 2.4  10
4 s
1 38b 99. Diebler, H., Secco, F. and Vetorini, M. (1989). J. Phys. Chem. 93, 1691 100. Liu, C.T., Bunn, S., Neverov, A.A. and Brown, R.S. To be published