Rory A. More O'Ferrall

1 H and 13C NMR spectra of α-heterocyclic ketones and assignment of keto, enol and enaminone tautomeric structures

A method for distinguishing enol and enaminone tautomers of acylmethyl heterocycles is described based on differences in 13C chemical shift between the carbonyl carbon atom of the enaminone (CO) and enolic carbon atom of the enol (C–OH). For 2-, 3- and 4-acetylmethyl, -phenacyl, and -pyridacyl-pyridines, -pyrazines, -quinolines, -quinoxalines, -phenanthrolines. -benzoxazoles and -benzothiazoles (usually in CDCl3 as solvent) measured values of δc fall in the non-overlapping ranges 179–191 ppm for the enaminones and 161–171 for the enols. Both sets of chemical shifts depend on the electronegativity of the acyl substituent and for strongly electron-withdrawing groups, such as pyruvate, δc for the enaminone does overlap the range for the enol. However the difference in chemical shifts (Δδc∼ 20) appears to be nearly independent of substituent, and in these cases the value for the enaminone can be predicted from a correlation of δc with σ* for the acyl substituent based on data of Greenhill, Loghmani-Khouzani and Maitland (J. Chem. Soc., Perkin Trans.1, 1991, 2831) for substituted quinolines. No other proton or carbon chemical shift differentiates the tautomers, but the structural assignments are corroborated by (a) comparisons with N-methyl enaminone models for the enaminone tautomers, (b) coupling constants (J34) between 3- and 4-hydrogen atoms of pyridine and quinoline rings and (c) allylic coupling between CH3 and vinyl hydrogens in enols of methyl ketones. With the exception of pyruvates, enols are observed only for 2-substituted heterocycles, usually as mixtures (in CDCl3) with the easily distinguished keto tautomers, but for (2-substituted) quinolines and quinoxalines the enaminone tautomer predominates. Structural assignments in aqueous media or other solvents where solubility precludes NMR measurements follow from correlations of UV–VIS with NMR spectra. NMR and UV–VIS measurements are considered in detail for the example of 2-phenacylpyridine.

Keto–Enol tautomerism and ionisation of 1-phenacylpyridinium ions: a model for carbanion-stabilisation of azomethine ylides

Measurement of equilibrium constants for keto–enol tautomerism (KT) and ionisation (Ka) of 1-phenacylpyridinium and 1-phenacyl(4-dimethylamino) pyridinium ions gives pKT(–log KT)= 6.10 and 5.55 and pKa= 10.90 and 13.2 respectively. The enol content and acidity of the 1-phenacylpyridinium ion is lower than that of its 2-, 3- and 4-isomers. and the possibility that this reflects impaired –M resonance by a 1-pyridinium substituent is discussed. Notional (proton) activating factors reflecting the influence of the positive charge of the 1-pyridinium substituent upon equilibrium ionisation and rates of deprotonation by lutidine and hydroxide bases are estimated from free energy correlations as 103, 17 and 5 × 103 respectively. These compare with a (methyl) activating factor of 109 derived from equilibrium ionisations of 4-chlorobenzaldehyde oxime and nitrone and a (notional) value of 106 for pyridine-N-oxide. The implications of these values for the activating effect of N-protonation of an azomethine group in models for pyridoxal-catalysed azomethine rearrangements are discussed.

Keto–Enol and imine–enamine tautomerism of 2-, 3- and 4-phenacylpyridines

Equilibrium constants for keto–enol tautomerisation and migration of hydrogen from carbon to nitrogen to form enamine or zwitterion tautomers have been measured for 2-, 3- and 4-phenacylpyridines (PyCH2COPh) in aqueous solution at 25 °C. Relative tautomeric stabilities fall in the order ketoimine > enamine > enol and (for the 3-isomer) enol > zwitterion. Values of pKT, (–log KT) where KT=[enamine (or zwitterion)]/[imine] or [enol]/[ketone], are 1.05, 5.87 and 2.42 for the enamine or zwitterion tautomerism and 2.0, 4.86 and 4.4 for keto–enol tautomerism of the 2-, 3- and 4-isomers respectively. For the enamines KT was determined kinetically by quenching the enolate anion at a pH below its pKa and monitoring its relaxation to the ketoimine spectrophotometrically: combining rate constants for this process and the reverse reaction measured by halogen trapping of the enol or enamine gave KT. Values are compared with results of earlier indirect measurements by Katritzky. For the 3-isomer, the N-methyl-3-phenacylpyridinium ion was taken as a model for the zwitferion tautomer and a ratio of enol to zwitterion concentrations of 10:1 was derived from kinetic and equilibrium ionisation measurements corrected for a methyl substituent effect. The unusually large enol content of 2-phenacylpyridine in non-polar solvents was measured spectrophotometrically and extrapolated to water. For the 4-isomer the enol content could be obtained from a correlation of pKas of phenacylpyridine enols and vinylogously related phenols. Acidic and basic pKas of all tautomers are reported including kinetically determined values for O-protonation of the enaminones. Proton activating factors for ionisation and tautomerisation have been calculated and are compared with values for the vinylogous hydroxypyridines. The influence of charge delocalisation and electrostatic interactions on the stability of enolate and carboxylate anions is discussed.

The mechanism of imine–enamine tautomerism of 2- and 4-phenacylquinolines

Isomerisations of 2- and 4-phenacylquinolines to their enaminone tautomers via 1,3 or 1,5 carbon–nitrogen hydrogen shifts occur by stepwise acid- or base-catalysed pathways similar to those for the enolisation of ketones. The reactions are observed as relaxations of the unstable to stable tautomers by stopped-flow spectrophotometry or, where the aromatic imine is the stable form, by trapping the enaminone with iodine in the reverse reaction. Evidence of mechanism comes from observations of general acid and general base catalysis, agreement between kinetically determined pKa values and independently measured values, and comparisons between rate and equilibrium constants for protonation of the enaminone tautomers and their N-methyl derivatives. The reactions show a primary isotope effect and yield normal Bronsted plots with αca. 0.5. The kinetically determined pKa values indicate N- rather than O-protonation of phenacylquinolines but for the enaminones O-protonation competes kinetically with the thermodynamically preferred C-protonation. Combination of pKa values for C-, N-, and O-protonation leads to equilibrium constants KT for enamine–imine and (protonated) keto–enol tautomerisation. The effect of 2- and 4-N-protonation (proton activating factors) upon rates and equilibria for ionisation of hydrogen from the methylene carbons is discussed and evidence of ‘imbalance’ in charge development on the carbon base in the transition state is noted. A concerted intramolecular 1,3-proton transfer is predicted but not observed.

Binding of protons and zinc ions to transition states for tautomerization of α-heterocyclic ketones: implications for enzymatic reactions

The description of catalysis in terms of binding of a catalyst to the transition state propoposed by Kurz is applied to tautomerization of the α-heterocyclic ketones phenacylpyridine, phenacylpyrazine, phenacylphenanthroline and phenylacetylpyridine catalysed by protons and zinc ions. Binding constants for protonated and zinc-coordinated transition states, KB≠ are reported and Bronsted coefficients are calculated from comparison of KB≠ with binding constants for the keto reactant and enolate anion intermediate. The formal equivalence of the binding formalism to a conventional Bronsted analysis is emphasized, and the results are compared with those from a ‘generalised’ Bronsted plot of rate constants against equilibrium constants for reactions of uncomplexed, protonated and zinc ion-coordinated ketones. This plot confirms that intrinsic reactivities of metal-coordinated and protonated substrates are similar even where differences exist between substrates. Application of a comparable Kurz–Bronsted treatment to enzymatic reactions depends in principle upon (a) dissecting binding contributions to catalysis from approximation of covalently reacting groups and (b) separating binding at the reaction site of the substrate, to which Kurzs treatment applies, from ‘remote’ binding, which, to a first approximation, is unchanged between Michaelis complex and transition state. The Bronsted relationship highlights stabilization of reactive intermediates as a thermodynamic driving force for binding catalysis at the reaction site. A formal expression which describes this stabilization, and also accommodates stabilization by remote binding of the substrate and intermediate by the enzyme, is proposed. Its relationship to the usual expression for application of the Kurz approach to enzyme catalysis, (kcat/k0)/Km = KB≠, is discussed and the usefulness of the Bronsted and Marcus relationships for interpreting KB≠ is emphasized.

Acid-catalysed aromatisation of benzene cis-1,2-dihydrodiols: a carbocation transition state poorly stablised by resonance

Acid-catalysed dehydration of 3-substituted benzene cis-1,2-dihydrodiols exhibits a Hammett plot with ρ=–8.2, consistent with reaction via a benzenonium ion-like intermediate; however, correlation of +M resonance substituents such as Me and MeO by σp rather than σ+ constants indicates a marked imbalance between resonance and inductive stabilisation of the transition state.

Keto–enol tautomerism and hydration of 9-acylfluorenes

Keto-enol tautomeric constants and ionisation constants have been measured for the keto and enol tautomers of 9-formyl-, 9-acetyl- and 9-benzoyl-fluorene in aqueous solution at 25 °C. Values of pKE(KE=[enol]/[ketone] and pKE=–log KE) are –1.22, 2.28 and 1.91, respectively, and the corresponding pKas for enolate anion formation are 6.19, 9.94 and 9.44 for the keto tautomers and 7.41, 7.66 and 7.53 for the enols. The measurements demonstrate the effectiveness of the fluorenyl group in increasing enol stability and ketone acidity. For 9-formylfluorene, for which the enol is the stable tautomer, KE is increased by a factor of more than 107 and the acidity of the keto tautomer by more than 109 relative to acetaldehyde (pKE= 6.17, pKa= 16.73). For 9-acetyl- and 9-benzoyl-fluorenes tautomeric constants were determined kinetically by combining rate constants for ketonisation measured spectrophotometrically following quenching of their enolate anions in carboxylic acid buffers with rate constants for enolisation measured by halogen trapping under the same conditions. For 9-formylfluorene rate constants for enolisation were measured by generating its unstable aldehyde tautomer from an ethanethiol hemithioacetal by reaction with iodine. Combining these rate constants with rate constants for ketonisation from trapping the aldehyde with bisulfite ion gave the tautomeric constant. In aqueous solution the aldehyde tautomer of 9-formylfluorene is appreciably hydrated and an equilibrium constant Kh=[hydrate]/[aldehyde]= 5.6 was derived from measurements of the (slower) equilibration of enol and hydrate following enolisation in acetic acid buffers. In aqueous solution therefore the enol (71%) and hydrate (24%) are the principal species. log k-pH profiles for enolisation, ketonisation and hydration reactions are reported. Intrinsic reactivities of the three enolate anions towards protonation by H3O+ and carboxylic acids are compared within an extended Bronsted plot of log k versusΔpK with measurements by Kresge for the corresponding enolate anions derived from fluorene-9-carboxylic acid and its methyl, methylthio and methylthione esters. Surprisingly, not only are the thin and thione esters less acidic than the oxygen ester, but intrinsically less reactive.

Application of tritium nuclear magnetic resonance spectroscopy to the determination of isotopic fractionation factors in methanol–methoxide solutions

Tritium n.m.r. measurements of hydroxy chemical shifts in methanolic solutions of sodium methoxide have been used to determine an isotopic fractionation factor for the inner solvation shell of the methoxide ion and contributions from inner and outer solvation shells to the methoxide ion chemical shift. The identity of protium and tritium chemical shifts and the relationship between tritium and deuterium fractionation factors φT=φ1.442 mean that measurements in MeOH and MeOD double the information available from 1H n.m.r. measurements alone. The necessary assumption previously made to derive φ from 1H measurements, that secondary solvation does not contribute to the methoxide ion chemical shift, is shown to be incorrect, but the revised value of φ(0.7) differs only slightly from earlier values, although treatment of the secondary solvation shift as a variable leads to some loss of precision in the definition of φ. At high methoxide concentrations plots of chemical shift against concentration are distinctly curved. Contributions to the curvature from breakdown of the assumption that isotopic atom fractions in the solution as a whole and in the bulk solvent are identical are evaluated.

Study of the enol–enaminone tautomerism of α-heterocyclic ketones by deuterium effects on 13C chemical shifts

Deuterium isotope effects on 13C chemical shifts have been measured for the enol and enaminone tautomers of a series of α-heterocyclic ketones. Partial deuteration of the exchangeable hydrogen bound to the oxygen atom of the enol or the nitrogen atom of the enaminone leads to deuterium induced shifts of the 13C frequencies (2DIS) which are distinctive for the two types of structures. Thus, 9-methyl-2-phenacyl-1,10-phenanthroline, 2-pyridylacyl- and 2-phenacyl-quinazolines and 2-pyridylacyl- and 2-phenacyl-quinolines, which are known from independent evidence to exist in the enaminone structures, display large and variable negative 4DIS values, –240, –93, –126, –437 and –375, respectively, at the carbon bearing the oxygen atom. By contrast, 2-pyridylacyl- and 2-phenacyl-pyrazines, which are known to exist in the enol form, show large positive 2DIS values, 527 and 479, respectively, for the oxygen bound carbon atom.

Keto–enol tautomerism of phenacylpyrazine: acid catalysis with protonation at nitrogen

Kinetic and equilibrium measurements for ionisation and enolisation of 2-phenacylpyrazine in aqueous solution at 25 °C yield a tautomeric constant pKE= 2.05 (where KE=[enol]/[ketone]) and pKas for loss of a methylene proton and for protonation at nitrogen of 11.90 and 0.40, respectively. In contrast to 2-phenacylpyridine the low basicity of the pyrazine nitrogens renders an enaminone tautomer less stable than the enol and a value of pKM= 4.4 (KM=[enaminone]/[ketone]) is estimated for this equilibrium. Evidence is presented that acid catalysis of keto-enol tautomerism occurs with protonation at the N-1 nitrogen atom rather than carbonyl group (or N-4 nitrogen) despite the proton being bound to oxygen in the enolic product. This preference reflects relative magnitudes of binding constants (1/Ka) and activating factors (PAF) for protonation at the different positions. Bronsted and Marcus equations are used to express catalytic efficiency in terms of equilibrium constants for binding the catalyst to the reactant and products of the uncatalysed reaction. The form of catalysis observed, which reflects the influence of proton binding on the activation energy of the reaction, is contrasted with that in intramolecular or enzymatic reactions, which normally derives from approximation of the reactants and is entropic in origin. The significance of optimum binding of the catalyst to the transition state in the two cases is briefly compared.