摘要:

J. Chem. Soc., Perkin Trans. 2, 1998 1 Nitrous and nitric acid ionisation equilibria and nitrous acidcatalysed nitration of 4-fluorophenol, in aqueous trifluoroacetic acid Ben D. Beake and Roy B. Moodie * Department of Chemistry, The University of Exeter, Stocker Road, Exeter, UK EX4 4QD The acidity dependencies of the equilibria between nitrous acid and nitrosonium ion, and between nitrate ion and nitric acid in aqueous trifluoroacetic acid (TFA) have been investigated. The decomposition of nitrous acid in air-saturated aqueous TFA has been studied and is considerably faster than in dilute HCl.Nitrous acid and 4-fluorophenol in TFA react to give quantitatively 4-fluoro-2-nitrophenol. The kinetics are shown to be consistent with the mechanism previously proposed for similar substrates in dilute aqueous acid. Introduction In our previous report on the nitrous acid-catalysed nitration and oxidation of phenols and related compounds in dilute hydrochloric acid, we focused on the crucial rôle played by the underlying inorganic nitrogen chemistry.1–5 Our study was aided by the availability of reliable literature values for several of the individual rate constants for the various equilibria and oxidation processes of nitrogen species which together comprise the decomposition of nitrous acid in dilute aqueous acid.4 An analysis of this decomposition using a numerical integration package enabled determination of the less well-established rate constants.With all the individual rate constants of the inorganic processes known, numerical methods led to successful mechanistic understanding and a quantitative fit to the observed data for the nitration and/or oxidation reactions of 4-phenoxyphenol,1 4-methoxyphenol,2 hydroquinone3 and ascorbic acid.5 In this paper, we have extended our previous work to investigate the reaction of 4-fluorophenol, (FP), with nitrous acid. This substrate shows no observable reaction in dilute aqueous acid, but reacts at a convenient rate in ca. 60% aqueous tri- fluoroacetic acid (TFA). There is no oxidation or nitrosation, but quantitative nitration to form 4-fluoro-2-nitrophenol (FNP). The medium has a high acidity with respect to the HR scale,6 low viscosity,7 and has none of the solubility limitations of aqueous systems.8 Despite these factors, to date there have been only scattered reports in the scientific literature on the potential viability of TFA as a nitrating medium.Widely varying yields of nitroproducts 9–11 have been reported for the nitrating system sodium nitrite–sodium nitrate–trifluoroacetic acid. In the first section of this paper we report details of an investigation into the inorganic nitrogen equilibria and decomposition chemistry which occur when sodium nitrite is added to aqueous TFA in the absence of other organic species. This includes an investigation of the acid–base equilibria between nitric acid and nitrate ion, and between nitrous acid and nitrosonium ion, concurrent decomposition of nitrous acid being avoided as far as possible by using low concentrations of nitrous acid for this part of the study.In view of the possible medium effects operating on the UV spectrum, as have been reported for sulfuric acid solutions,12,13 the absorption data for the equilibrium studies were analysed using characteristic vector analysis (CVA).14–16 This technique is able to provide information on protonation equilibria where spectral changes due to protonation are accompanied by medium effects on the spectra of the acid and base species.In the second section, we describe the results of the study of the reaction of nitrous acid with 4-fluorophenol. This substrate is known to react with a mixture of nitrous and nitric acids in aqueous sulfuric acid to give 4-fluoro-2-nitrophenol in quantitative yield.17 A similar conversion occurs with nitrous acid in TFA with or without added nitrate, and the mechanism proposed is that occurring for similar substrates in media of lower acidity.1–5 Results Inorganic equilibria and decomposition reactions The equilibrium between nitrous acid and nitrosonium ion is observable between 84.5 and 98.1% TFA (Fig. 1) (Here and below, x% TFA means a mixture of water and TFA containing x% by weight of TFA). Above 98.1% TFA, the absorbance at 280 nm was found to decrease (Fig. 2). The equilibrium was found to be unaffected by changes in the initial concentration of nitrous acid (though this was kept very low to avoid as far as possible concomitant decomposition) or by the presence/ absence of nitrate ion/nitric acid.Addition of sulfuric acid to 100% TFA caused the absorbance to decrease still further (Fig. 2). Characteristic vector analysis (CVA) was hindered by limitations in accessible wavelengths (above 265 nm) where solvent absorbance did not interfere with that of the nitrosonium ion whose broad peak is centred at about 250–260 nm.CVA showed Fig. 1 Absorbance-wavelength scans of HNO2/NO1 (4 × 1023 M) in TFA solutions of varying acidity. a = 99.6%; b = 95.3%; c = 91.4%; d = 87.8%; e = 84.5%. Repeat scans (pecked lines) taken 2 min later show decomposition2 J. Chem. Soc., Perkin Trans. 2, 1998 that below 98.1% TFA medium effects on the system (shown to be present by the absence of an isosbestic point in Fig. 1) were small (the first characteristic vector, associated with the effect of protonation, accounted for over 99% of the variation), thus allowing absorbance values to be used directly to study the equilibrium using eqn.(1), where e(HNO2) and e(NO1) are [NO1] [HNO2] = e 2 e(HNO2) e(NO1) 2 e (1) the extinction coefficients of molecular nitrous acid and nitrosonium ion respectively. These are taken as the e values obtained in 84.5 and 98.1% TFA respectively. The plot of log ([NO1]/ [HNO2]) vs. 2HR fitted eqn. (2).For comparison with results for aqueous sulfuric acid, see Table 1. log ([NO1]/[HNO2]) = 1.12 (± 0.15)(2HR) 2 11.4 (± 1.5) (2) The aerobic decomposition of nitrous acid, as previously,4 had an approximately first-order kinetic form, and the observed first-order rate constant exhibited a steep dependence (approximately second order) on the initial concentration of nitrous acid (Fig. 3). Dependence on the acidity of the medium is illustrated in Fig. 4. Solutions where NIII was present mainly as nitrosonium ion were relatively stable whereas those at lower acidity were found to decompose to nitric acid (Fig. 1). Below 68% TFA, the rate of the decomposition decreased markedly with the drop in acidity of the medium. The nitrate ion–nitric acid equilibrium was observable between 42 and 99% TFA. Characteristic vector analysis revealed that the scalar multiples (s) associated with the first characteristic vector, from which [HNO3]/[NO3 2] was estimated, exhibited a sigmoidal dependence on acidity (Fig. 5) and accounted for 85% of the variance. The plot of log [HNO3]/ [NO3 2] vs. 2H0 fitted eqn. (3) (see Table 1). log [HNO3]/[NO3 2] = 0.89 (± 0.06) (2H0) 21.35 (± 0.08) (3) Fig. 2 Variation in absorbance at 280 nm due to HNO2/NO1 (1 × 1024 M) in TFA solutions of varying acidity. The addition of small quantities of sulfuric acid to a UV cell containing 3 cm3 of TFA, to increase the acidity over that in 100% of TFA, is shown in the inset. Table 1 Comparison with acidity dependence at 25 8C of the equilibria in sulfuric acid Medium TFA H2SO4 log ([NO1]/[HNO2]) slope vs. 2HR 1.12 a 0.75 c log ([HNO3]/[NO3 2]) slope vs. 2H0 0.89 b 0.61 d a This study. Half-ionised in 93.5% TFA. b This study. Half-ionised in 82.2% TFA. c Half-ionised in 56.3% H2SO4 (ref. 13). d Half-ionised in ca. 44% H2SO4 (refs. 25 and 26). Nitration of 4-fluorophenol Most of our studies relate to 61.8% TFA because this gave rise to convenient rates of reaction.It is clear from the work described above that at this acidity the predominant NIII and NV species, as in 1 M HCl,1–5 are HNO2 and NO3 2 respectively. The reaction exhibited good first-order kinetics when nitrous acid was present in excess. The value of kobs increased with nitrous acid, proportionately at the lower concentrations but much less steeply at the higher concentrations (Fig. 6). The value of kobs was insensitive to the wavelength used for its determination, indicating the absence of an intermediate in signifi- cant concentration.Conversion to FNP was quantitative. When FP was initially in excess the reaction could be fitted only imperfectly to a first-order form, and the best-fitting firstorder rate constant changed little with the concentration of the reagent (FP) present in excess. The reaction did not conform to Fig. 3 Dependence of the rate of nitrous acid decomposition on its initial concentration in 61.8% TFA Fig. 4 Variation of kobs with acidity for the decomposition of nitrous acid in aqueous TFA Fig. 5 Acidity dependence of the scalar multiples (s) associated with the first characteristic vector obtained by CVA of UV data for solutions of sodium nitrate in TFAJ. Chem. Soc., Perkin Trans. 2, 1998 3 the simple rate equation, eqn. (4). Thus division of the observed rate = k2[HNO2][FP] (4) first-order rate constant by the concentration of the reagent (HNO2 or FP) present in excess, each over a range of concentrations, gave a wide range of different values of k2.Added sodium nitrate was found markedly to enhance the rate. When FP and HNO2 were initially present in equal amount with a large excess of sodium nitrate the reaction exhibited good firstorder kinetics. At high concentrations of added nitrate ion, kobs approached a limiting value. A brief investigation of the acidity dependence revealed that kobs increased with acidity, and more steeply the higher the nitrous acid concentration. (Fig. 7). The reaction of added nitrate with FP in the absence of Fig. 6 Variation in kobs for the nitration of FP (1 × 1024 M) with nitrous acid concentration in 61.8% TFA Fig. 7 Acidity dependence of the nitrous acid catalysed nitration of FP (1 × 1024 M) in TFA. Nitrous acid concentration (1 × 1022 M, j; 1 × 1023 M, d; 5 × 1024 M, m). Fig. 8 Reaction of added nitrate (1 × 1022 M) with FP (1 × 1024 M) in the absence of added nitrous acid in 61.8% TFA solution added nitrous acid was very slow and a marked and variable induction period was observed (Fig. 8). The product spectrum was essentially the same as that with nitrous acid present. Discussion The analysis of the HNO2/NO1 equilibrium was limited by the solvent absorbance masking a considerable portion of the peak. CVA of the absorbance data suggested that medium effects on this equilibrium were small below ca. 98% TFA and therefore absorbance values could be used directly to follow the equilibrium.The observed decrease in absorbance above 98.1% TFA was surprising (Fig. 2). There are several possible explanations for this. A decrease in the protonating power of the medium over 98.1% TFA was considered. A minimum in H0 has been reported at ca. 98% TFA.6 However the HNO2/NO1 equilibrium is thought to proceed via protonation–dehydration and hence would be expected to more closely follow the HR acidity function (see below) where no such minimum has been observed below 100% TFA.6 This is supported by the observation of the effect of adding small amounts of sulfuric acid to 100% TFA.This would surely increase the acidity although the absorbance at 280 nm continues to decrease (Fig. 2). An effect caused by a change in the ionic atmosphere of the medium may also be discounted since the addition of a small amount of sulfuric acid to 100% TFA has the opposite effect to addition of small amounts of water. The loss of NIII through decomposition and the subsequent relaxation of the related equilibria (see later) may be considered.It has been shown11 that gaseous NO is evolved on addition of sodium nitrite to 100% TFA, although the loss of NIII was small. Studies with differing initial concentrations of nitrous acid or added nitric acid showed little change in absorbance and hence changes in the related inorganic equilibria do not appear to affect that between nitrous acid and nitrosonium ion. It seems most probable that the observed decrease in absorbance at 280 nm above 98.1% TFA, and the further decrease on addition of small quantities of sulfuric acid, is a medium effect shifting the absorbance of the nitrosonium ion to lower wavelength and/or lowering its intensity.A similar effect has been observed 13 in sulfuric acid solutions above 70%. Log ([NO1]/[HNO2]) in 84.5–98.1% TFA increases similarly to, but slightly more steeply than, 2HR (Table 1). This is reasonable since it is formally similar to the protonation– dehydration equilibrium on which the HR acidity function was determined using a range of triarylcarbinols.The acidity dependence of the equilibrium formation of nitric acid from nitrate ion in 42.2–99.1% TFA relates more closely to H0. Table 1 shows the comparison between the acidity dependencies of these equilibria in TFA with those previously determined in sulfuric acid. The studies of the HNO2/NO1 and NO3 2/HNO3 equilibria in sulfuric acid reveal that they are much less steeply dependent on the acidity of the medium than in aqueous trifluoroacetic acid.It seems that the peculiarity of the individual medium is an important factor in determining the positions of these ionisation equilibria, and that acidity scales in themselves are not a reliable predictive guide. The observed first-order rate constant, kobs for decomposition of nitrous acid (0.004 M) in 61.8% TFA is approximately 20 times greater than in 1 M HCl, even though in both media essentially all of the NIII is present as nitrous acid.The decomposition also exhibits a different dependence on the initial concentration of nitrous acid (Fig. 3). [In our study of the nitrous acid decomposition in dilute acid,4 the order in initial nitrous acid was found to be constant (1.65)]. There are several possibilities for the enhanced kobs in TFA. It is probable 18 that the solubility of oxygen in 61.8% TFA is higher than in dilute HCl.However, since the oxygen is present in deficit this would be expected, on the basis of the reaction in4 J. Chem. Soc., Perkin Trans. 2, 1998 dilute aqueous HCl,4 to have a minimal effect on kobs. In airsaturated 61.8% TFA solution the extent of decomposition is consistent with an oxygen concentration of ca. 3 × 1024 M. This figure has been used in the calculations. Since the reactions (a) and (b) in Scheme 1 involve water they 2NO2 k1 k21 HNO2 1 NO3 2 1 H1 (a) 2HNO2 k2 k22 NO 1 NO2 (b) 2NO 1 O2 k3 2NO2 (c) FC6H4OH 1 HNO2 k4 k24 FC6H4O? 1 NO (d) FC6H4O? 1 NO2 k5 FC6H3(NO2)OH (e) Scheme 1 will be affected by changes in the activity of water, which in this medium is 0.85.19 This departure from unity is insufficient in itself to account for the observed marked change in rate although it may be a contributory factor.Of greater signifi- cance is the fact that the k21 step is acidity dependent, and the value of k21[H1], going from 1 M HCl to 61.8% TFA, will have increased to an uncertain degree.There is reason to suspect that reaction (c) 5,20 may have an acid-catalysed component. In addition to all this, medium effects on the rate constants can be anticipated. The set chosen (see footnotes to Table 2) represent those that can adequately explain both the inorganic decomposition and the organic process in 61.8% TFA, although their exact values should not be considered to be precisely determined. The choice of rate constants in the fitting procedure is discussed further below.Nitration of 4-fluorophenol We discuss the nitration results in terms of the scheme put forward previously,1–3 to explain nitration of 4-substituted phenols with nitrous acid in dilute hydrochloric acid, which is related to that proposed for nitrous acid-catalysed nitration of phenols and substituted phenols in nitric or sulfuric acids.8,21,22 In Scheme 1, water has been omitted [steps (a), (b) and (d )], to make the significance of the rate constants clear.It is a min- Table 2 Comparison of fitted with observed behaviour a [HNO2]/ 1024 M 200 100 50 10 10 10 1 1 1 0.1 50 30 [NaNO3]/ 1024 M 0 0 0 0 120 20 1000 100 0 0 0 0 [FP]/1024 M 1 1 1 1 1 1 1 1 10 100 0 0 mean kobs/ 1023 s21 17 12 9.3 2.2 54 8.2 7.6 3.0 0.60 1.0 40 9.1 kcalc/1023 s21 15 13 8.3 1.7 68 26 9.8 9.0 0.50 1.1 40 12 a Air-saturated solution containing TFA (61.8%). The last two rows refer to the rate of decomposition of nitrous acid.The values for the individual rate constants used in the fitting procedure were as follows (values for the inorganic reactions previously used for dilute aqueous HCl are in parentheses): k1 = 1 × 108 M21 s21 (1 × 108 M21 s21); k21[H1] = 10 M21 s21 (k21 = 5 × 1023 M22 s21); k2 = 1 × 102 M21 s21 (15 M21 s21); k22 = 2 × 108 M21 s21 (2 × 108 M21 s21); k3 = 4 × 107 M22 s21 (2.1 × 106 M22 s21); k4 = 1 × 102 M21 s21; k24 = k5 = 1 × 108 M21 s21, with the concentration of dissolved oxygen taken to be 3 × 1024 M. imal scheme, omitting also the low concentration intermediates almost certainly involved.These three steps are probably rate determined (described for the reactions in the forward direction) by water attack on N2O4, homolysis of N2O3, and homolysis of a nitroso-substituted cyclohexadienone intermediate, respectively. In what follows, we have attempted to explain the kinetics of nitration and the inorganic decomposition reaction with one set of rate constants.In contrast to the similar reactions in dilute aqueous acid 1–5 none of the rate constants has been independently determined, and together they provide too many fitting parameters for any one to be considered precisely determined. The set chosen is in the footnotes to Table 2, where they are compared with the values used for 1 M HCl, given in parentheses. No attempt has been made to refine these fitting parameters to better than one significant figure.The important feature is that, as is clear from Table 2, a set can be chosen which adequately explain both the inorganic decomposition and the organic process under a wide range of conditions in 61.8% TFA. The contrast between conditions where nitrous acid is in excess and those where FP is in excess is well reproduced, as is the effect of substantial addition of nitrate. The effect of small concentrations of nitrate is not so well fitted (rows 6 and 8), but further refinement was not considered worthwhile at the present time.The values chosen for the rate constants of steps (a)–(c) were as similar as possible to those for dilute aqueous HCl. With k1 and k22 unchanged, it was necessary to make substantial increases in k21[H1], and k3, both reasonable in view of the increased acidity, and in k2, partially explained by the reduced activity of water aiding the pre-equilibrium dehydration to N2O3. The surprising feature of the organic rate constants necessary to achieve a fit was that the rate constants k24 and k5, which are for combination of the 4-fluorophenoxy radical with NO and NO2 respectively, had to be made approximately equal.(The treatment was insensitive to the actual value provided it was large.) This was necessary to accommodate the reactions both in the absence of added nitrate (when the numerical integration showed [NO]/[NO2] always to be large so that k5 was rate-determining in the organic sequence) and in the presence of added nitrate (where k4 can become rate-determining.) We had previously deduced a value of ca. 500 for the ratio k5/k24 for the uncatalysed similar reaction of 4-(RO)-phenol (R = Me, H).3 One possible explanation is that the reactive species under these more acidic conditions is the protonated 4-fluorophenoxy radical, (4-fluorophenol radical cation) and that it reacts on encounter with both NO and NO2. It is noteworthy in this connection that the reaction is acid catalysed (Fig. 7). The induction period in the absence of nitrous acid The presence of an induction period has been observed in several nitrating systems where nitrous acid is absent initially.23,24 Clemens, Ridd and Sandall noted 23 an induction period in the nitric acid nitration of p-nitrophenol in 90% TFA which was absent when a small amount of nitrous acid was initially present. The authors concluded that nitrous acid has an autocatalytic effect and suggested that it may be formed through side-reactions involving the substrate, although these were not detailed.We note the similarity but can cast no further light on the origin of this effect. Experimental Materials Trifluoroacetic acid (TFA, > 99% purity, Lancaster Chemicals) was further purified by fractional distillation from sulfuric acid, and the fraction boiling between 70–72 8C was retained. All TFA solutions are quoted as % w/w. Sodium nitrate, sodium nitrite and sulfuric acid (98%) were AR reagents.Trifluoroacetic anhydride (TFAA) was obtained from Lancaster (highpurity reagent) and used without further purification. SodiumJ. Chem. Soc., Perkin Trans. 2, 1998 5 nitrite and sodium nitrate were AR reagents. 4-Fluorophenol (FP, 99%, ex Aldrich) and 4-fluoro-2-nitrophenol (FNP, 98%, ex Lancaster) were used without further purification. Procedure for inorganic equilibria and decomposition study The equilibrium between nitrous acid and nitrosonium ion was studied by UV spectroscopic methods.Aqueous sodium nitrite (2 × 102524 × 1023 M) was added via syringe to a 1 cm UV cell containing an aqueous solution of TFA thermostatted to 25.0 8C. After mixing the cell was transferred to the thermostatted (25.0 8C) cell compartment of a Perkin-Elmer Lambda 5 spectrometer. Scan spectra were recorded immediately over 400–265 nm at 240 nm min21, baseline corrected against the same cell before addition of the sodium nitrite (solvent absorbance is negligible above 265 nm).Single wavelength work was carried out at 280 nm. Characteristic vector analysis of the absorption data was done by computer. The absorbance A at r wavelengths is obtained for n different acid concentrations and a resultant nrow, r-column data matrix is set up. Statistical manipulation of this matrix produces a series of characteristic vectors v, each with associated scalar multiples, s. The value of the absorbance at a given acidity is given by eqn.(5). The scalar multiples (si) Ar = A� r 1 s1v1,r 1 s2v2,r 1. . . (5) represent the amount of each characteristic vector (vi, j) which has to be added to the mean absorbance (A� r) to reproduce the original curve (Ai). It has been found that the absorbance data can be described by only two characteristic vectors, and it is chemically reasonable to associate the first of these (account for most of the variance) with the effect of protonation and the second with a medium effect.The necessity of adding via syringe aqueous solutions of sodium nitrite (for accuracy, and to minimise loss of gaseous NO11) precluded reaching 100% TFA directly. 100% TFA was reached by allowing a known amount of trifluoroacetic anhydride [TFAA] to be hydrolysed by the water added with the sodium nitrite. The rate of TFAA hydrolysis (and thus the time taken to complete hydrolysis) was separately determined. With 2.75 × 1025 M TFAA at 25.0 8C, monitoring the reaction spectrophotometrically at 280 nm, the reaction conformed approximately to the kinetic form, eqn.(6). The deduced value 2d[TFAA]/dt = k2[TFAA][H2O] (6) of k2 varied from 53 to 65 dm3 mol21 s21 over the range 1–5 × 1024 M H2O. The decomposition of nitrous acid in air-saturated TFA solutions was monitored at 368 nm using full 4 cm UV cells to minimise ingress of oxygen by phase transfer from the air above the solution. The effects of varying the acidity and initial concentration of nitrous acid on its decomposition were investigated.The procedure for investigation of the nitric acid/nitrate ion equilibrium was analogous to that followed in the study of the nitrosonium ion/nitrous acid equilibrium, with sodium nitrate being used in the place of sodium nitrite. Extinction data were calculated for 6 wavelengths between 270 and 320 nm for 11 acidities between 42.2 and 99.1% TFA. CVA was used to analyse the data. Procedure for nitration study In a typical kinetic run, sodium nitrite was added last to a UV cell at 25.0 8C containing FP (1 × 1024 M) and aqueous TFA to generate nitrous acid in situ in aqueous TFA of known composition.The kinetics of nitration were folloby changes in absorbance at 360 nm (the position of the absorbance maximum of FNP). Conclusions The acidity dependencies of the equilibrium between nitrous acid and nitrosonium ion and the equilibrium between nitrate ion and nitric acid have been determined at 25.0 8C in aqueous trifluoroacetic acid (TFA).The ionisation ratios for both these equilibria increase with acidity (as measured by the relevant acidity function) more rapidly than in aqueous sulfuric acid. The decomposition of nitrous acid in air-saturated ca. 60% TFA is considerably faster than that previously observed in dilute HCl. Trifluoroacetic acid is a good solvent for nitration of 4-fluorophenol by nitrous acid. Quantitative conversion to the 2-nitroproduct was observed whether or not NV was added initially. A consequence of the changes to the underlying inorganic equilibria was that the effect of added NV was much greater in this system than in aqueous HCl.The results are consistent with the mechanism proposed. Acknowledgements We thank the EPRSC for financial support. References 1 B. D. Beake, J. Constantine and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2, 1992, 1653. 2 B. D. Beake, J. Constantine and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2, 1994, 335. 3 B. D. Beake, R. B. Moodie and J. P. B. Sandall, J. Chem. Soc., Perkin Trans. 2, 1994, 957. 4 B. D. Beake and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2, 1995, 1045. 5 B. D. Beake, R. B. Moodie and D. Smith, J. Chem. Soc., Perkin Trans. 2, 1995, 1251. 6 T. W. Toone, U. A. Spitzer and R. Stewart, Can. J. Chem., 1976, 54, 440. 7 M. G. Harris and J. B. Milne, Can. J. Chem., 1971, 49, 1888. 8 G. D. Tobin, Ph.D. Thesis, University of Exeter, 1978. 9 U. A. Spitzer and R. Stewart, J. Org. Chem., 1974, 39, 3936. 10 S. Uemura, A. Toshimitsu and M. Okano, J. Chem. Soc., Perkin Trans. 1, 1978, 1076. 11 B. Milligan, J. Org. Chem., 1983, 48, 1495. 12 N. S. Bayliss and D. W. Watts, Aust. J. Chem., 1963, 16, 933. 13 S. N. Richards, Ph.D. Thesis, University of Exeter, 1986. 14 T. W. Anderson, Introduction to Multivariate Statistical Analysis, Wiley, New York, 1958. 15 J. L. Simonds, J. Opt. Soc. Am., 1963, 58, 968. 16 R. L. Reeves, J. Am. Chem. Soc., 1966, 88, 2440. 17 U. Al-Obaidi, Ph.D. Thesis, University of Exeter, 1986. 18 D. V. Bryk, R. G. Makitra, Ya. N. Pirig and Yu. V. Stefanyk, Zh. Prikl. Khim., 1988, 61, 91. 19 N. G. Zarakhani and N. P. Vorob’era, Russ. J. Phys. Chem. (Engl. Trans.), 1972, 46, 1392. 20 M. Pires, M. J. Rossi and D. S. Ross, Int. J. Chem. Kinet., 1994, 26, 1207. 21 U. Al-Obaidi and R. B. Moodie, J. Chem. Soc., Perkin Trans. 2, 1985, 467. 22 M. Ali and J. H. Ridd, J. Chem. Soc., Perkin Trans. 2, 1986, 327. 23 A. H. Clemens, J. H. Ridd and J. P. B. Sandall, J. Chem. Soc., Perkin Trans. 2, 1984, 1667. 24 M. J. Thompson and P. J. Zeegers, Tetrahedron Lett., 1988, 29, 2471. 25 N. C. Deno, H. J. Peterson and E. J. Sacher, J. Phys. Chem., 1961, 65, 199. 26 N. C. Marziano, A. Tomasin and P. G. Traverso, J. Chem. Soc., Perkin Trans. 2, 1981, 1070 and references therein. Paper 7/07021G Received 29th September 1997 Accepted 20th October 1997

ISSN:1472-779X    

DOI:10.1039/a707021g

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1999, 5.13 5 A computational study of the reactivity of diethenylnaphthalenes towards anionic polymerization Ahu Ak©¥n,a Safiye Sag¢§ Erdem,a Turgut Nugay,a Viktorya Aviyente a and Haluk Res¢�at b a Chemistry Department, Bog¢§ azici University, 80815, Bebek, I¡�stanbul, Turkiye b Faculty of Arts and Sciences, Koc University, I¡�stinye, I¡�stanbul, Turkiye Received 3rd June 1998, Accepted 11th November 1998 Diethenyl, di(1-methylethenyl), and di(1-phenylethenyl) naphthalenes are known to be difunctional initiators used in the synthesis of thermoplastic elastomers.Semiempirical (AM1, PM3) and ab initio calculations (HF/6-31G, HF/6-31G*) have been carried out to determine the reactivity of these compounds towards anionic polymerization. For this purpose, geometrical parameters, electrostatic potentials, and frontier orbitals have been analyzed. Reaction paths starting from the diethenylnaphthalenes and reaching the proposed products have been studied, and transition structures along the paths have been located.The minimum energy conformers were determined through a conformational search around single bonds for a series of diethenylnaphthalenes. We have attempted to predict how the location of the vinyl groups aVects the reactivity of diethenylnaphthalenes. Our results have revealed that the most suitable difunctional initiators for anionic polymerization are the compounds where the substituents lie away from the naphthalene bridge. We have also found that in some cases the substituents are conjugated with each other and di(1-phenylethenyl)naphthalenes are more reactive than diethenylnaphthalenes which in turn are more reactive than di(1-methylethenyl)naphthalenes towards anionic polymerization.Introduction Vinyl-diene triblock copolymers are among the most extensively investigated thermoplastic elastomers because they were the earliest to be produced and commercialized.Thermoplastic elastomers are block copolymers that exhibit rubberlike elasticity 1 without requiring chemical crosslinking. The term ¡°rubberlike elasticity¡± implies that a material can be extended to several times its original length, yet return rapidly to nearly its initial dimensions upon removal of the deforming force. Block copolymers that behave as thermoplastic elastomers are described as either ABA or (AB)n polymers, according to the number and type of repeating units per macromolecule.Homogeneous anionic polymerization methods 2.4 provide close control of the molecular weight, molecular weight distribution and composition of each block in the ABA copolymers. Certain vinyl, diene and cyclic monomers can be polymerized by an anionic mechanism with no termination step.3,4 Since the growing chain remains active in such ¡°living polymers¡±, diVerent monomers may be added stepwise to build each block in sequence. Anionic polymerization proceeds by the addition of monomers to active centers bearing a whole or a partial negative charge.The active center is regenerated in each step. The chain propagation is illustrated in eqn. (1). Difunctional anionic initiators are used because triblock copolymers can be synthesized in only two or one monomer addition steps.5 Difunctional initiators are also important when the second monomer is incapable of reinitiating the polymerization of the first monomer.6 The synthesis of the difunctional anionic initiators has generally followed two methods: (i) the generation of ion.radical species which couple to yield the dicarbanionic initiator,7,8 (ii) the reaction of butyllithium with a diethenyl compound.9 For vinyl and diene monomers alkyl lithiums are particularly favored initiators.The reaction between 1,5-diethenylnaphthalene and sec-butyllithium has produced a new difunctional organolithium initiator [eqn. (2)] which is soluble in non-polar solvents and eVective in the synthesis of the styrene.isoprene.styrene triblock copolymer.10 The purpose of this research is to assist experimentalists by giving a means of predicting reactivity trends of diethenylnaphthalenes towards sec-butyllithium.We have attempted to understand how the location of the vinyl groups and the nature of the substituent on the vinyl groups will aVect the reactivity of the diethenylnaphthalenes towards anionic polymerization. To this end, we have modeled several diethenylnaphthalenes, di(1-methylethenyl)naphthalenes and di(1-phenylethenyl)naphthalenes.Our approach was to determine the most stable conformer of each species and to analyze its properties of interest such as the transition structures for the addition of sec-butyllithium, the frontier orbitals and the electrostatic potentials. The nomenclature used in the discussion is shown in Fig. 1, the terms exy, mxy, pxy have been used to denote diethenylnaphthalenes, di(1-methylethenyl)naphthalenes and di(1-phenylethenyl) naphthalenes respectively where x and y show the position of the substituent on the naphthalene ring.The numbering system used throughout this article is shown in Fig. 2. Methodology All possible conformers for diethenylnaphthalenes, di(1-methylethenyl) naphthalenes and di(1-phenylethenyl)naphthalenes have been investigated using the MM2 force field provided with-6 J. Chem. Soc., Perkin Trans. 2, 1999, 5–13 in the Spartan package.11 The following parameters have been investigated: for diethenylnaphthalenes, free rotation around the naphthalene–vinyl bond; for di(1-methylethenyl)naphthalenes free rotation around the naphthalene–vinyl and the vinyl–methyl bonds, for di(1-phenylethenyl)naphthalenes free rotation around the naphthalene–vinyl and the phenyl–vinyl Fig. 1 Nomenclature for diethenylnaphthalenes. R = H, diethenylnaphthalenes: e15, e16, e17, e26, e27; R = CH3, di(1-methylethenyl)- naphthalenes: m15, m16, m17, m26, m27; R = Ph, di(1-phenylethenyl)- naphthalenes: p15, p16, p17, p26, p27.Fig. 2 Numbering system used for diethenylnaphthalenes. bonds. The stationary points having the minimum energy with the MM2 force field—three for each molecule—have been chosen as potential candidates for further study employing semiempirical methods. The choice of the semiempirical method to be used is based upon a comparison between the experimental and calculated values for the torsion angle between the vinyl group and naphthalene in 1-vinylnaphthalene.The high resolution NMR spectra of 1-vinylnaphthalene have been used to deduce a value of 36.7–45.98 for the angle between the vinyl group and the ring planes.12 The value calculated by AM1 is 40.18 whereas PM3 has set this value to 08. Based on these findings, AM1 has been chosen for further investigations. For each compound, the conformers located as minima on the potential energy surface have been further optimized with HF/6-31G using the GAUSSIAN94 program.13 The energetics have also been reported with single point HF/6-31G* (HF/6-31G*//HF/6-31G) calculations (Table 1).Selected dihedral angles are gathered in Table 2, bond lengths are given in Table 3. The activation barriers for the addition of sec-butyllithium to diethenylnaphthalenes have been evaluated by considering a stepwise addition of the lithium salt (Figs. 3 and 4). The transition structures are four centered: the C]] C bond and the Li–C bonds break, whereas new C–Li and C–C bonds form (Fig. 5). Due to the size of the compounds, this reaction path has been modeled for 2,6-diethenylnaphthalenes only. The parameters for Li are available in PM3 (tm) but not in AM1, thus the PM3 method has been used to evaluate the relative energies for the addition reaction. Many diVerent reactivity measures have been introduced to quantify the chemical activity of various groups or sites of molecules.14,15 To name a few, free valencies, frontier orbitals and the molecular electrostatic potentials can be used as the indicators of molecular reactivity.The magnitude of the coe Ycients of the lowest unoccupied molecular orbitals (LUMO’s) is an indication of the sites which are more susceptible to attack by ahile. Any electric charge distribution creates an electrostatic potential. For example, the nucleus and the electrons of the atoms give rise to a potential field around the receptor molecule.The sign of the potential V(r) in any particular region depends on whether the potential of the nuclei or of the electrons is the dominating factor. The molecular electrostatic potential of the receptor creates favorable binding Table 1 Relative energies (kcal mol21) with respect to the most stable compoundsa Compound e15 e16 e17 e26 b e27 m15 m16 m17 m26c m27 p15 p16 p17 p26d p27 DHf(rel) AM1 3.35 1.59 1.65 0.00 0.02 3.72 1.80 1.77 0.00 0.02 3.86 1.92 2.06 0.00 0.13 Er(rel) HF/6-31G 5.54 2.62 2.64 0.00 0.64 4.86 2.26 2.39 0.00 0.02 Et(rel) HF/6-31G*//HF/ 6-31G 11.72 2.59 2.62 0.00 7.06 4.47 2.07 2.20 0.00 0.04 a Energies are given for the most stable conformer of each compound.b DHf = 73.74 kcal mol21. Et (HF/6-31G = 2536.933525 H, HF/6-31G* = 2537.121.332 H). c DHf = 61.47 kcal mol21. Et (HF/6-31G = 2614.97209 H, HF/6-31G* = 2615.19194 H). d DHf = 129.37 kcal mol21.J. Chem. Soc., Perkin Trans. 2, 1999, 5–13 7 Table 2 Selected dihedral angles (8) for diethenylnaphthalenes, di(1-methylethenyl)naphthalenes (HF/6-31G) and di(1-phenylethenyl)naphthalenes (AM1) Dihedral angle C12–C11–C1–C2 C12–C11–C2–C3 C14–C13–C5–C6 C14–C13–C6–C5 C14–C13–C7–C6 C12–C11–C1–C2 C12–C11–C2–C3 C14–C13–C5–C6 C14–C13–C6–C5 C14–C13–C7–C6 H27–C21–C11–C12 H30–C24–C13–C14 C12–C11–C1–C2 C12–C11–C2–C3 C14–C13–C5–C6 C14–C13–C6–C5 C14–C13–C7–C6 C20–C15–C11–C1 C20–C15–C11–C2 C22–C21–C13–C5 C22–C21–C13–C6 C22–C21–C13–C7 e15 45.0 — 245.0 —— m15 78.4 — 78.4 —— 4.5 4.6 p15 112.9 — 2112.6 —— 238.9 — 38.9 —— e16 43.6 —— 2177.6 — m16 78.5 —— 148.8 — 4.4 7.8 p16 271.1 —— 39.4 — 238.6 —— 39.8 — e17 44.1 ——— 2.4 m17 279.0 ——— 134.7 24.4 4.4 p17 114.0 ——— 42.9 239.4 ——— 2138.8 e26 — 0.72 — 2179.3 — m26 — 232.0 — 245.5 — 7.8 4.2 p26 — 240.1 — 239.6 —— 240.9 — 240.3 — e27 — 0.5 —— 2153.5 m27 — 232.4 —— 232.5 7.2 7.8 p27 — 40.2 —— 40.1 — 40.6 —— 40.7 sites for the ligands or restricts the approach of the ligands to certain regions of the molecule.An approaching nucleophile would move favorably towards regions of positive V(r). To investigate the reactivity of the studied diethenylnaphthalenes, we analyzed the frontier orbitals resulting from AM1 optimization. We have also created the electrostatic potential at the molecular surfaces. To create the electrostatic potentials, first Mulliken charges for the optimized geometries were determined. Then, using Mulliken charges, the electrostatic potential was determined at the molecular surfaces using the Grasp program. 16 Since the nucleophilic attachment site for the sec-butyl anion is the end vinylic carbon, only the surface points which are within 2.5 Å from the end vinylic carbon with a potential larger than 20 kT (about 0.5 eV at room temperature) were assumed to form the reactive region.Changing the radius and the threshold potential did not aVect the relative trend of the reactivity of the compounds.Results and discussion A. Energetics and geometries of diethenylnaphthalenes The energetics of the compounds considered are seen to depend upon the position of the substituents on the rings. The relative energies (Table 1) show that substitutions at 2,6- or 2,7-positions give rise to the more stable compounds while substitutions at 1,5-positions near the bridge produce isomers which are less stable. As expected 1,6- and 1,7-substituted naphthalenes lie in between and are almost isoenergetic.We have also optimized 1,8-substituted naphthalenes which turned out to be highly energetic in comparison to the rest of the compounds. Obviously, steric hindrance is the major factor aVecting the relative stability of these compounds. When positions away from the bridgehead are occupied the compound is more stable. It is interesting to notice that the ranking of the compounds is not sensitive to the method used, both AM1 and HF/6-31G produce the same trend regarding their stabilities.For all the compounds studied, 2,6-substituted isomers are found to be more stable than the others whereas 1,5-substituted isomers are found to be less stable than the others. The internal rotation in styrene and substituted styrenes has been investigated by a large variety of techniques including microwave,17a infrared,17b Raman,17c ultraviolet,17d fluorescence, 17e photoelectron17f and NMR spectroscopy,17g molecular rotatory polarization,17h calorimetric,17i molecular mechanics calculations,17j semiempirical 17k and ab initio molecular orbital calculations.17l The consensus from these investigations is that the internal rotation in styrene is governed largely by a twofold barrier with the planar form being the most stable.For styrene, the magnitude of the observed quadrupolar splittings with high resolution deuterium NMR spectra were used to calculate the average value of the dihedral angle between the vinyl and ring planes as 16.58.18 Klemm et al.concluded on the basis of ultraviolet spectroscopy that 1-vinylnaphthalene exists in a non-planar anti conformation, 1-anti whereas 2- vinylnaphthalene exists in a coplanar conformation 2-syn.12 Similar conclusions were reached on the basis of 1H NMR chemical shifts.19 On the basis of proton–proton nuclear Overhauser eVects, it was concluded that the minimum energy conformation for 1-vinylnaphthalene was 1-anti with the deviation of the torsion angle a from planarity being equal to 388.The same torsional angle a determined by high-field NMR spectroscopy was shown to be 41.8 ± 4.18 for 1-vinylnaphthalene and as 18.3 ± 3.18 for 2-vinylnaphthalene.18 In our calculations, analyses of the dihedral angles between the vinyl group and the naphthalene ring planes have demonstrated that the substituents orient themselves in such a way as to minimize their interaction with the ring and with each other (Table 2). We have compared the optimum geometries8 J.Chem. Soc., Perkin Trans. 2, 1999, 5–13 Fig. 3 Reaction path for the addition of sec-butyllithium to a, 2,6-diethenylnaphthalene, b, 2,6-di(1-methylethenyl)naphthalene, c, 2,6-di(1- phenylethenyl)naphthalene. with HF/6-31G for diethenylnaphthalenes with the findings in the literature for ethenylnaphthalenes. The vinyl group at position 1 for compounds e15, e16 and e17 is anti to the bridge and tilted by approximately 458 from planarity in agreement with the experimental findings for 1-vinylnaphthalene.18 The compounds with vinyl substituents at positions 2, 6 and 7 can be compared to 2-vinylnaphthalene.Similar to experimental results for 2-vinylnaphthalene, HF/6-31G calculations also give rise to coplanar vinyl groups with the naphthalene ring in e16, e17, e26 and e27. However, the vinyl group adopts an anti orientation with respect to the bridge in these compounds except for position 7 in e27. We have found out that the energies of diVerent conformers for a given diethenylnaphthalene compound are within 2 kcal mol21 of each other. The substituents might aVect each other’s orientation with respect to the naphthalene ring.The value of the dihedral angle between the vinyl group and the naphthalene ring increases by substituting methyl and phenyl groups on the vinyl group in compounds m15, m16, m17, p15, p16 and p17. For compounds m26, m27, p26 and p27 the dihedral angle C12–C11–C2–C3 is greater than zero, the coplanar situation of the vinyl group with respect to naphthalene is destroyed, due to the presence of bulky substituents.For di(1-methylethenyl)naphthalenes one of the H’s of the methyl group is eclipsed with the vinylic double bond, this may be due to long range stabilizing interactions between the p electrons and the hydrogen.J. Chem. Soc., Perkin Trans. 2, 1999, 5–13 9 It is expected that variations in bond distances will provide evidence for the conjugation eVect (Table 3).The bond lengths in the naphthalene ring adjacent to the substituents are longer than their homologues: for example in e15 the C1–C2, C1–C9, C5–C6, C5–C10, C9–C10 bonds are longer by 0.01–0.02 Å than C3–C4, C2–C3. The same is true for all the compounds. This elongation of the bonds vicinal to the substituents may be interpreted as a reflection of the electron withdrawing character of the vinyl groups. On the other hand, the C]] C bond in the vinyl group is shorter by about 0.03 Å than the double bonds in naphthalene, indicating that delocalization of p-electrons over the naphthalene ring does not extend to the external vinylic fragment with the same eYciency.The C–C bond between the Table 3 Bond lengths (Å) of diethenylnaphthalenes, di(1-methylethenyl) naphthalenes (HF/6-31G) and di(1-phenylethenyl)naphthalenes (AM1) Bond length C1–C2 C3–C2 C4–C3 C10–C4 C10–C5 C6–C5 C7–C6 C8–C7 C8–C9 C9–C1 C10–C9 C11a C12–C11 C13b C14–C13 C1–C2 C3–C2 C4–C3 C10–C4 C10–C5 C6–C5 C7–C6 C8–C7 C8–C9 C9–C1 C10–C9 C11a C12–C11 C13b C14–C13 C21–C11 C24–C13 C1–C2 C3–C2 C4–C3 C10–C4 C10–C5 C6–C5 C7–C6 C8–C7 C8–C9 C9–C1 C10–C9 C11a C12–C11 C13b C14–C13 C15–C11 C21–C13 e15 1.369 1.410 1.361 1.421 1.434 1.369 1.410 1.361 1.421 1.434 1.417 1.482 1.326 1.482 1.326 m15 1.366 1.412 1.360 1.422 1.433 1.366 1.412 1.360 1.421 1.433 1.417 1.500 1.326 1.500 1.326 1.512 1.512 p15 1.380 1.413 1.372 1.422 1.430 1.380 1.413 1.372 1.422 1.430 1.421 1.472 1.341 1.472 1.341 1.466 1.466 e16 1.369 1.413 1.360 1.420 1.417 1.368 1.422 1.361 1.423 1.432 1.414 1.482 1.326 1.477 1.327 m16 1.368 1.414 1.360 1.420 1.419 1.368 1.423 1.360 1.421 1.431 1.412 1.499 1.326 1.491 1.330 1.512 1.512 p16 1.380 1.414 1.372 1.422 1.421 1.380 1.423 1.371 1.423 1.429 1.419 1.472 1.341 1.467 1.343 1.466 1.466 e17 1.369 1.412 1.360 1.418 1.418 1.360 1.420 1.370 1.420 1.434 1.414 1.482 1.326 1.478 1.327 m17 1.367 1.414 1.360 1.418 1.419 1.360 1.422 1.369 1.421 1.433 1.413 1.500 1.326 1.492 1.329 1.512 1.512 p17 1.381 1.413 1.372 1.421 1.422 1.371 1.423 1.380 1.421 1.430 1.419 1.472 1.341 1.467 1.342 1.466 1.466 e26 1.370 1.424 1.360 1.421 1.418 1.370 1.424 1.358 1.421 1.416 1.409 1.477 1.327 1.474 1.327 m26 1.369 1.425 1.360 1.420 1.417 1.369 1.424 1.361 1.419 1.418 1.409 1.491 1.330 1.492 1.329 1.512 1.512 p26 1.381 1.424 1.371 1.422 1.419 1.380 1.423 1.372 1.422 1.420 1.418 1.467 1.343 1.467 1.343 1.466 1.467 e27 1.369 1.423 1.360 1.418 1.418 1.360 1.423 1.369 1.417 1.419 1.411 1.478 1.327 1.478 1.327 m27 1.369 1.425 1.360 1.418 1.418 1.360 1.425 1.369 1.419 1.419 1.408 1.492 1.330 1.492 1.330 1.512 1.512 p27 1.380 1.424 1.372 1.421 1.421 1.372 1.424 1.380 1.421 1.421 1.418 1.467 1.343 1.467 1.343 1.467 1.467 a C1 in 1,5-, 1,6- and 1,7-diethenylnaphthalenes; C2 in 2,6- and 2,7- diethenylnaphthalenes.b C5 in 1,5-; C6 in 1,6- and 2,6-; C7 in 1,7- and 2,7-diethenylnaphthalenes.Fig. 4 Energetics for the addition reaction of sec-butyllithium to a, 2,6-diethenylnaphthalene, b, 2,6-di(1-methylethenyl)naphthalene, c, 2,6-di(1-phenylethenyl)naphthalene. Fig. 5 Geometry for the transition states for the addition of secbutyllithium to disubstituted naphthalenes. Distances are given in plain for R = H, in underlined for R = CH3 and in bold for R = Ph.10 J. Chem. Soc., Perkin Trans. 2, 1999, 5–13 vinyl group and the ring is longer for methyl and phenyl substituted compounds.This result can be justified by noting that methyl or phenyl substitution increases the steric interactions of the substituent with the naphthalene ring and the vinyl group tends to be away from the ring. However, comparison of positions 1 and 2 in terms of the bond length between the vinyl group and the ring indicates that this bond is shorter in position 2 because at this position, the substituent suVers less from steric repulsions. Also, due to the quasi planarity of these compounds migration of electrons towards the vinyl group can cause shortening of the bond.For substituents at positions 2, 6 and 7, diVerences in the length between double and single bonds diminish in comparison to positions 1 and 5. p electrons are more delocalized in compounds which are 2,6,7-substituted rather than 1 and 5 substituted ones, confirming the extra stability of the former compounds over the others. The trends observed for the bond lengths in the ring with HF/6-31G for diethenylnaphthalenes and di(1-methylethenyl)naphthalenes is also observed with di(1-phenylethenyl)naphthalenes. For di(1- phenylethenyl)naphthalenes, the C–C bond distance between the substituent and the ring is only 0.05 Å longer than the single bond in the ring, the same value is 0.06 Å for the ethenyl substituents and 0.08 Å for the methylethenyl substituents.These findings are due to the electron withdrawing character of the phenyl group which pulls the electrons from the naphthalene ring more than hydrogen and methyl species.B. Modeling the addition of sec-butyllithium to diethenylnaphthalenes We have considered the addition reaction of sec-butyllithium to the most stable 2,6-diethenylnaphthalenes in Fig. 3 where ts1 and ts2 represent the first and second transition states, p1 and p2 stand for the first and second products after the addition of one and two moles of sec-butyllithium respectively.The addition of sec-butyllithium to the first vinylic double bonds is an exothermic reaction. For all the substituents, the first step is more exothermic than the second one. The first activation barrier is considered as the rate determining step since the Table 4 The coeYcients for the LUMO in diethenylnaphthalenes e15 s px py pz e16 s px py pz e17 s px py pz e26 s px py pz e27 s px py pz C11 0.00051 0.09952 0.01370 0.05323 0.01440 0.10579 20.05147 0.01493 20.00049 0.03841 20.11072 0.03180 0.00128 20.04280 0.00320 0.13550 0.00164 20.02047 20.10399 20.03123 C12 20.01915 0.20076 0.00739 20.08027 20.00692 20.19086 0.11083 20.03366 20.01908 20.03633 0.22102 20.04400 20.00847 0.09848 0.00044 20.24946 20.00607 0.03917 0.18730 0.07325 C13 0.00053 20.09929 20.01342 20.05282 0.00064 0.09410 20.05121 20.04489 20.00161 0.10425 20.03943 0.05742 20.00128 20.04280 0.00320 20.13550 0.00164 0.02047 0.10399 20.03123 C14 20.01923 0.20025 20.00803 0.07943 20.01779 20.16029 0.09337 0.12049 0.00811 20.20693 0.07014 20.09191 0.00847 0.09848 0.00044 0.24946 20.00607 20.03917 20.18730 0.07325 second barrier can be overcome by the energy given oV during the first step.The energetics displayed in Fig. 4 suggest that the substituents lower the activation barrier. The methyl group with its electron donor ability and the phenyl group with its electron withdrawing character polarize the double bond towards the addition reaction. The geometry of the transition states is such that the double bond character of the vinylic double bond (1.33 Å) has disappeared (1.41 Å).The C–Li bond elongates from 2.01 Å in sec-butyllithium to 2.18 Å. The other sides of the quadrilateral are of 2.20 Å and 2.09 Å respectively. As seen from Fig. 5 the nature of the substituent does not aVect the geometrical parameters of the transition structures. C. Frontier orbitals The LUMO’s of the disubstituted naphthalenes have been considered since the attacking anion will be accommodated in the atomic orbitals with higher coeYcients.Consideration of Tables 4–6 shows that coeYcients for C12 and C14 are larger than the ones for C11 and C13. As expected the anion will attack the end carbons (C12 and C14) rather than the central vinylic carbons (C11 and C13). Comparison of the coeYcients at diVerent positions indicates that the LUMO’s have larger coeYcients at positions 2, 6 and 7 which are away from the bridgehead.Thus, naphthalenes with substituents at positions 2, 6 and 7 are preferentially attacked by the alkyl anion. For the symmetrical 2,6- and 2,7- diethenylnaphthalenes, both positions, 2 and 6 or 2 and 7 are equally susceptible towards the nucleophilic attack. The ratio of the coeYcients has been considered for the comparison of the three substituents (R = H, CH3, Ph): when the vinyl groups are away from the bridgehead positions, the phenyl group favors the attack of the anion somewhat more than methyl and hydrogen.On the other hand when the vinyl groups are close to the bridgehead the substitution eVect is not observed. D. Electrostatic potential The reactivity of molecules can be calculated from the electrostatic potential V(r) in several ways. The molecular surface potentials of the studied naphthalenes are reported in Figs. 6 and 7. The electrostatic potential energy of an anion with a charge q = 21 would be U(r) = qV(r) = 2V(r). The most straightforward analysis would be to calculate the patch area on the surface defining the reaction region.If one assumes random collisions between the reacting molecules, the reaction rate would simply be proportional to the reaction patch area. In this, it is implicitly assumed that the contribution to the reactivity is the same in each part of the reaction patch regardless of the value of the electrostatic potential. Results of this type of analysis are reported in Table 7 under PA, the patch area.As described in the Methodology section, the reaction region was defined as the surface points which have an energy lower than 20 kT and are within a distance of 2.5 Å from the vinylic end carbon atoms. A slightly more complicated analysis would weight the contributions of diVerent regions of the reaction zone with an appropriate Boltzmann factor NKa = AÚ s e2buda (3) where A is a proportionality constant and da is the infinitesimal area element of the integrated surface.This type of analysis has been used before in determining the association constant of reactions. Because of the exponential dependence, points with large potentials would dominate the integral NKa. Therefore, the observed trend in the maximum value of the electrostatic potential V(r) and in the weighted sum of the above integral are the same. It should however be kept in mind that the electrostatic potentials were generated when the anionic ligand was not in the vicinity.Therefore, since they willJ. Chem. Soc., Perkin Trans. 2, 1999, 5–13 11 be aVected less, the reaction patch area sizes might be a better indicator of the polymer’s stability. Results reported in Table 7 show that positions away from the bridgehead are more susceptible to being attacked by the anion and the 1,5-substituted compounds are the least reactive ones toward the anionic Table 5 The coeYcients for the LUMO in di(1-methylethenyl)- naphthalenes m15 s px py pz m16 s px py pz m17 s px py pz m26 s px py pz m27 s px py pz C11 20.00009 0.01027 0.06443 0.00486 0.00059 0.04344 0.02100 0.04370 0.00031 20.02516 0.03027 20.05110 0.00104 0.06801 0.01208 20.10235 0.00122 20.08224 0.02851 20.04854 C12 0.02737 20.02139 20.10891 20.06474 0.02649 20.04400 20.07586 20.09252 0.02698 0.01398 20.10130 0.07899 0.01829 20.14488 20.03473 0.16726 0.01540 0.13280 20.06501 0.09583 C13 0.00009 0.02547 0.01680 0.05774 0.00076 0.10622 0.04370 20.00217 20.00006 0.01316 20.06724 20.08255 20.00104 0.06801 0.01208 0.10235 0.00122 0.08220 20.02851 20.04853 C14 20.02739 20.02139 20.10891 20.06474 0.01719 20.17718 20.10024 20.00395 0.01266 20.05787 0.12188 0.14856 20.01829 20.14488 20.03473 20.16726 0.01540 20.13274 0.06501 0.09580 Table 6 The coeYcients for the LUMO in di(1-phenylethenyl)- naphthalenes p15 s px py pz p16 s px py pz p17 s px py pz p26 s px py pz p27 s px py pz C11 20.00423 0.00158 20.04225 20.06510 0.00358 20.01967 0.02887 0.04681 0.00330 0.04221 20.03419 20.04986 0.00136 0.08671 20.10385 20.03163 0.00198 0.03644 20.10905 0.00407 C12 20.02602 20.06261 20.08230 0.10078 0.02732 0.08319 20.04404 20.04939 0.02564 20.12196 0.05837 0.04857 0.01371 20.18943 0.16327 0.04909 0.01422 20.04197 0.19349 20.01260 C13 20.00425 20.00164 0.04176 0.06447 0.00110 0.04134 20.01141 0.12956 20.00213 20.03361 0.05680 20.11125 0.00212 0.07733 0.11346 0.02990 0.00198 20.08357 0.03740 0.06989 C14 20.02609 0.06269 20.08151 20.09940 0.01134 20.04445 0.04473 20.23738 20.01744 0.05203 20.12566 0.17901 0.01745 20.16484 20.18176 20.04626 0.01421 0.13634 20.04999 20.13554 polymerization. This can be justified by steric restrictions around the bridgehead limiting the access to the binding site.Among other substitutions, 1,7-substitution is not favored either, which might be due to the fact that in this configuration the substitutions are placed very close to each other. Compounds with the 1,6-, 2,6- and 2,7-substitutions are more Fig. 6 Electrostatic potential at the molecular surfaces of 2,6-substituted diethenylnaphthalenes. The orientation of the molecule is such that one of the vinyl groups is facing the reader on the left upper corner, and the other is looking into the page on the right. a, e26, b, m26, and c, p26.12 J. Chem. Soc., Perkin Trans. 2, 1999, 5–13 susceptible to a nucleophilic attack. Judging from the patch areas the 2,7-substituted compounds would polymerize slightly better than the equivalent 1,6- and 2,6-substituted ones.The observed trend in the maximum value of the electrostatic potential V(r) reflects the eVect of the substituents on the reactivity of the compounds of interest. The values for V(r) in the reaction region for 1,5-, 1,7- and 2,6-substituted com- Fig. 7 Reactive patch areas reported in Table 7 for the 2,6-substituted diethenylnaphthalenes. Small white spheres around the vinyl groups show the defined reactive molecular surface regions.a, e26, b, m26, and c, p26. pounds are seen to mimic the electron withdrawing abilities of the substituents, phenyl substituted compounds have the greatest V(r) while methyl substituted compounds have the lowest V(r). This trend is not observed for 1,6- and 2,7-substituted compounds. We have explained this diVerence in behavior between these two classes of compounds by drawing resonance structures for representatives of each class.Fig. 8 shows that for 1,5-, 1,7- and 2,6-substituted compounds the two substituents are in resonance with each other through a continuous delocalization path. This will enhance the electron withdrawing abilities of the substituents. On the other hand, it is diYcult to predict a trend for compounds with 1,6- and 2,7-substituents where the two substituents are not conjugated due to a discontinuous delocalization path. Conclusions The following conclusions can be drawn from the results presented. 1. Both AM1 and HF/6-31G can be used to suYciently discuss the geometries and energies of the compounds of interest. 2. Substitution away from the bridgehead of naphthalene has generated stable compounds due to the delocalization which Fig. 8 Delocalization in diethenylnaphthalenes. a, 2,6-diethenylnaphthalenes, b, 1,6-diethenylnaphthalenes. Table 7 Electrostatic potentials in the reaction region a Compound PA log10 (NKa) Vmax(r) Compound PA log10 (NKa) Vmax(r) Compound PA log10 (NKa) Vmax(r) e15 1.13 23.26 44.9 m15 0.68 24.54 41.7 p15 0.91 21.43 49.5 e16 1.33 22.11 48.7 m16 0.89 1.92 58.2 p16 1.23 1.56 57.3 e17 1.25 22.39 48.1 m17 0.81 23.43 45.6 p17 1.02 20.65 52.2 e26 1.34 1.16 56.4 m26 1.00 0.00 53.2 p26 1.23 3.62 62.1 e27 1.34 2.82 60.2 m27 1.07 1.25 56.1 p27 1.59 20.05 53.0 a The reaction region is described in the text.PA is the reaction patch area, log10 (NKa) is the base 10 logarithm of the association constant calculated using eqn.(3). Both the PA and NKa are normalized with their respective values for the m26 compound. Vmax(r) is the largest V(r) among the surface points of the reaction region in the units of kBT where kB is the Boltzmann’s constant and T is the temperature.J. Chem. Soc., Perkin Trans. 2, 1999, 5–13 13 was possible because of the quasi planarity of these compounds. 3. Frontier orbitals and electrostatic potentials have shown that positions 2, 6 and 7 are more reactive than the other positions. 4. Methyl and phenyl substituted vinyl compounds have been shown to decrease the activation barriers in comparison to the non-substituted vinyl groups. These findings are confirmed by the presence of coeYcients for end carbons in the LUMO’s. 5. Electrostatic potential analysis results further support the conclusion that the 1,6-, 2,6- and 2,7-substituted compounds are more susceptible to a nucleophilic attack, and therefore, should have better polymerization characteristics. 6. In some cases—1,5-, 1,7- and 2,6-substituted compounds— the substituents are in conjugation with each other. It is then possible to state that the phenyl group with its electron withdrawing character enhances polymerization more than hydrogen and methyl groups. Acknowledgements The authors acknowledge support from the Bog¡aziçi University Research Funds. References 1 L. R. G. Treloar, Rubber Chem. Technol., 1974, 47, 625. 2 L. J. Fetters, J. Res. Nat. Bur. Stand., Sect.A, 1966, 70, 421. 3 L. J. Fetters, J. Polym. Sci., Part C, 1969, 26, 1. 4 M. Morton and L. J. Fetters, Macromol. Rev., 1967, 2, 71. 5 D. C. Ailport and W. H. James, Block Copolymers, Halsted Press, New York, USA, 1973. 6 A. Noshay and J. E. McGrath, Block Copolymers, Academic Press, London, UK, 1977. 7 M. Morton and L. J. Fetters, Macromolecules, 1969, 2, 453. 8 S. F. Reed, J. Polym. Sci., Part A: Polym. Chem., 1972, 10, 1187. 9 L. H. Tung, G. Y. S. Lo and D. E. Beyer, Macromolecules, 1978, 11, 616. 10 (a) T. Nugay and S. Küçükyavuz, Polym. Int., 1992, 29, 195; (b) T. Nugay and G. Riess, Elf Atochem S. A., Fr. Pat. 95400334.9, 1995. 11 SPARTAN Version 4.0 Wavefunction, Inc., 18401 Von Karman Ave., #370 Irvine, CA 92715, 1996. 12 L. H. Klemm, H. ZiVer, J. W. Sprague and W. Hodes, J. Org. Chem., 1955, 20, 190. 13 GAUSSIAN94 Revision C.3, M. J. Frish, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Cheeseman, T. Keith, G. A. Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayakkara, M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. V. Wong, J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defrees, J. Baker, J. J. P. Stewart, M. Head-Gordon, C. Gonzales and J. A. Pople, Gaussian Inc., Pittsburgh, PA, 1995. 14 W. B. Smith, Theoretical Organic Chemistry, VCH Publishers, Inc., 1996. 15 J. S. Murray, T. Brinck and P. Politzer, Chem. Phys., 1996, 204, 289. 16 A. Nicholls, K. A. Sharp and B. Honig, Proteins, 1991, 11, 281. 17 (a) W. M. Ralowski, P. J. Mjöberg and S. O. Ljunggren, J. Mol. Struct., 1976, 30, 1; (b) W. G. Fately, G. L. Carlson and F. E. Dickson, Appl. Spectrosc., 1968, 22, 650; (c) J. M. Hollas, H. Musa, T. Ridley, P. H. Turner, K. H. Weisenberger and V. Fawcett, J. Mol. Spectrosc., 1982, 94, 437; (d ) J. M. Hollas, E. Khalilipour and S. N. Thakur, J. Mol. Spectrosc., 1978, 73, 240; (e) J. M. Hollas and T. Ridley, Chem. Phys. Lett., 1980, 75, 94; ( f ) J. P. Maier and D. W. Turner, J. Chem. Soc., Faraday Trans. 2, 1973, 69, 196; (g) R. Laatikainen and E. Kolehmainen, J. Magn. Reson., 1985, 65, 89; (h) P. Castan, A. Lopez and R. Martino, Tetrahedron, 1979, 35, 1093; (i) K. S. Pitzer, L. Guttman and F. E. Westrum, Jr., J. Am. Chem. Soc., 1946, 68, 2209; ( j) N. L. Allinger and J. T. Sprague, J. Am. Chem. Soc., 1973, 95, 3893; (k) I. Baraldi, F. Momicchioli and G. Ponterini, J. Mol. Struct., 1984, 110, 187; (l) T. Schaefer and G. H. Penner, Chem. Phys. Lett., 1985, 114, 526, and references therein. 18 K. L. Facchine, S. W. Staley, P. C. M. van Zijl, P. K. Mishra and A. A. Bothner-By, J. Am. Chem. Soc., 1988, 110, 4900. 19 S. W. Staley, C. K. Dustman and G. E. Linkowski, J. Am. Chem. Soc., 1981, 103, 1069. Paper 8/04188A

ISSN:1472-779X    

DOI:10.1039/a804188a

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1998 7 AlCl3-mediated migration of the benzamido group of N-phenoxybenzamide derivatives to the phenyl group Etsuko Miyazawa, Takeshi Sakamoto and Yasuo Kikugawa * Faculty of Pharmaceutical Sciences, Josai University, 1-1 Keyakidai, Sakado, Saitama 350-02, Japan AlCl3-mediated decomposition of N-phenoxybenzamide derivatives in dichloromethane mainly leads to regioselective intramolecular migration of the benzamido group from the oxygen to the ortho position of the phenyl group via electron-deficient nitrogen intermediates.Previously we reported AlCl3-mediated regioselective migration of the methoxy group of N-methoxy-N-phenylamides from the nitrogen to the ortho position of the phenyl group.1 This rearrangement reaction suggests generation of an N-acyl-Narylnitrenium ion 2 and subsequent nucleophilic migration of the methoxy group to the ortho position of the phenyl group via a tight ion pair intermediate. In an extension of this work, we have investigated the reaction of N-phenoxybenzamide derivatives (1) with AlCl3.Treatment of N-phenoxybenzamide (1a) with AlCl3 (5 equiv.) in CH2Cl2 for 1 h at room temperature gave N-(2-hydroxyphenyl)- benzamide (2a) (85%). Several ring substituted phenoxyamines were synthesized by the literature and patent methods 3 and reacted with benzoyl chloride in pyridine with cooling to give 1. Compounds 1 thus obtained were reacted with AlCl3 in this way, and the results are presented in Table 1. We have investigated the reaction of 1 with various Lewis acids [AlCl3, BF3?OEt2, AlMe3 and La(OTf)3] in aprotic solvents (CH2Cl2 and benzene) and found that AlCl3-mediated decomposition of 1 leads (i) to the electrophilic intramolecular migration of the benzamido group from the oxygen to the ortho position of the phenyl ring to give 2 as major products, (ii) to the ipso position of the phenolic moiety followed by loss of the oxygen from the molecule to give 3, and (iii) in the case of 1i to the fluorine substituted carbon through ‘back-donation’ 4 by fluorine atom, followed by migration of the fluorine atom and subsequent elimination of hydrogen fluoride in the molecule to give 2l, as a minor product (Table 1).Proposed reaction mechanisms are shown in Schemes 1 and 2. There was no migration of the benzamido group to the para position. The structure of 2l was confirmed by spectral analyses and chemical transformation.Thus, it was hydrogenated over 10% palladium on carbon in ethyl acetate in the presence of triethylamine (2.8 equiv.) for 3 h to give 2g in 73% yield. Five equiv. of AlCl3 to 1 were needed to obtain a high yield of 2 and 3, and starting compound (1e) was recovered by use of 2 equiv. of AlCl3 in 1 h (76%). Use of BF3?OEt2 instead of AlCl3 required prolonged reaction time (2–3 days at room temperature) and in this case the carbonyl oxygen rather than the benzamido nitrogen migrated to the phenyl group which upon hydrolysis gave 2-hydroxy-5-methylphenyl benzoate (5a) in yields of 62% in CH2Cl2 and 74% in benzene.It seems that the rearrangement proceeds in a concerted mechanism via a sixmembered cyclic transition state. In the case of La(OTf)3 starting compound (1e) was recovered quantitatively. Reactions of 1 with AlMe3 gave N-methylbenzamide (6) and the rearrangement products; production of 6 indicates that N-benzoylnitrenium ion was generated and trapped by AlMe3.The results are presented in Table 2. N-(4-Methylphenoxy)acetamide (7) reacted with AlCl3 at room temperature for 1 h to give N-(2-hydroxy-5-methylphenyl) acetamide (8) (26%) and starting compound (7, 33%). It seems that the benzamido group migrates more easily than the acetamido group. Structural elucidation of the products was performed by the measurements of IR, NMR and mass spectra, and elemental analyses. The key feature of the assignments is that in NMR spectra N-(2-substituted phenoxy)amide derivatives exhibit signals at unusually low field for the H-6 proton owing to steric effects 5 and intramolecular hydrogen bonding between the amide N]H proton and an ortho substituent.6 In the case of N-(2-substituted phenoxy)amide derivatives, the H-6 proton appeared down field and reasonable correspondence of the chemical shift of H-6 proton to these structures was observed, PhCONHO X Y HO X Y PhCONH PhCONH X Y Z OH HO PhCOO MeCONHO X HO MeCONH Me Me a X = H, Y = H b X = 4-F, Y = H c X = 6-F, Y = H d X = 3-Cl, Y = H e X = 4-Cl, Y = H f X = 6-Cl, Y = H g X = 5-Cl, Y = H h X = 3-Me, Y = H i X = 5-Me, Y = H j X = 3-Cl, Y = 5-Cl k X = 3-F, Y = 6-Br l X = 4-Br, Y = 5-Cl m X = 3-Br, Y = 6-Br 2 a X = Y = Z = H b X = 2-Cl, Y = 4-Cl, Z = H c X = 4-Me, Y = 5-Cl, Z = H d X = 2-Cl, Y = 4-Cl, Z = 6-Cl e X = 2-Br, Y = 4-Cl, Z = 5-Br a X = H, Y = H b X = 3-F, Y = H c X = 2-Cl, Y = H d X = 3-Cl, Y = H e X = 4-Cl, Y = H f X = 2-Me, Y = H g X = 4-Me, Y = H h X = 2-Cl, Y = 4-Cl i X = 2-F, Y = 5-Br j X = 2-Br, Y = 5-Br 5a X = Me b X = H 3 1 4a 2-OH b 4-OH 6 8 PhCONHMe 78 J.Chem. Soc., Perkin Trans. 2, 1998 except the case of entries 9 and 10 in Table 3. Thus, introduction of a methyl group at the 3 or 4 position causes pronounced up field shifts of the ortho proton resonances, the reason for which is not fully understood at present. It was reported that this ortho effect cannot be extended to ortho substituted phenolic esters.5b Indeed, the H-6 proton of 5a showed an usual chemical shift (d 6.96–7.03).Therefore, to confirm the structures of 2h and 2i, these were hydrolyzed 7 by heating in aqueous NaOH which transformed them into the corresponding aniline derivatives. To investigate the mechanistic aspects of the rearrangement, we have undertaken to study this reaction using benzene as solvent. Treatment of 1a with AlCl3 (5 equiv.) in benzene for 1 h at room temperature gave 2a (36%) and benzanilide (3a) (58%).Evidently, coordination of AlCl3 with the phenoxy oxygen induces heterolytic cleavage of the N]O bond to give an Nacylnitrenium ion and a phenoxide anion. The positive charge thus produced was trapped intramolecularly by a phenoxide anion and canonical forms involving the benzene ring to give 2a or trapped intermolecularly by benzene to give 3a. It is very difficult to prove the presence of an acylnitrenium ion; however, a concerted mechanism is less favorable because thermodynamically unstable four-membered cyclic transition state must be considered in the case of production of ortho migration products.Scheme 1 Proposed mechanism (main route) O AlCl3 O AlCl3 Ph N+ O Al PhCONH N H O Ph Al – Cl3AlO + 1a AlCl3 Cl2AlO PhCON Al AlCl3 – HCl – + – HO PhCONH – HCl H2O 2a Table 1 Reaction of N-phenoxybenzamides (1) with AlCl3 (5 equiv.) in CH2Cl2 at room temperature Entry 1 2 3 4 5 6 7 8 9 10 Starting compound 1a 1b 1c 1d 1e 1f 1g 1h 1i 1j t/h 1 3 a 0.5 1 a 0.8 0.5 1 1.5 5.5 4.5 Product (yield, %) 2a (85) 2b (48), 2c (22) 2d (49), 2e (53), 2f (20) 2g (65), 2h (58) 2i (69), 2j (44), 2k (19), 21 (46) 2m (22), 3b (19) 3b (20) 3c (9) 3d (25) 3e (13) a Reflux.Table 2 Reaction of N-phenoxybenzamides (1) with AlMe3 (5 equiv.) in CH2Cl2 at room temperature Entry 1 2 3 4 Starting compound 1c 1e 1f 1g t/h 2 a 4 4 0.5 a Product (yield, %) 2d (4), 6 (74), 1c (7) 2g (46), 6 (14), 1e (9) 2h (52), 6 (13), 1f (4) 2i (71), 6 (4), 1g (4) a Ice cooling.In contrast, Shudo and co-workers have previously reported the reaction of the same compound 1a with a strong protonic acid.8 Treatment of 1a with a mixture of trifluoroacetic acid (TFA) and trifluoromethanesulfonic acid (TFSA) in benzene gave 2- and 4-hydroxybiphenyls (4a and 4b) (6% and 3%, respectively) and 2-hydroxyphenyl benzoate (5b) (43%). Protonation of the amide carbonyl of 1a and subsequent heterolytic cleavage of the N]O bond brought about formation of a phenoxenium ion, which was trapped intermolecularly by benzene to give 4a and 4b or trapped intramolecularly by the carbonyl oxygen to give 5b.It is interesting to note that the mode of cleavage of the N]O bond was completely dependent on the kind of acid used and that an N-acylnitrenium ion or a phenoxenium ion is produced from the same compound by use of a Lewis acid or a strong protonic acid, respectively (Scheme 3).In conclusion, it is assumed that successful use of AlCl3 for the generation of acylnitrenium ions and subsequent electrophilic migration to the phenyl ring is due mainly to the fact that the acylnitrenium–AlCl3 complex will be sufficiently stable and Scheme 2 Proposed mechanism (minor route) O Cl AlCl3 Ph N+ O Al Cl Cl3AlO N O Ph Al Cl Cl2AlO N O Ph Al N Cl Cl PhCONH Cl Cl O Ph Al – HF O AlCl3 Ph N+ O Al H2O H2O – AlCl2OH N O Ph 3b 2l O AlCl3 Cl – – Al 1e F 2AlCl3 – HCl Br – Br F + PhCON Al – HCl – Br + Cl2AlO F – H Cl 1i 2AlCl3 Cl2AlO PhCON Br Cl Al HO PhCON Br Cl H Cl H Cl –J.Chem. Soc., Perkin Trans. 2, 1998 9 Table 3 1H NMR chemical shift data for H-6 in benzanilides H O N Ph X Y Z H dH Entry 1 2 3 4 5 6 7 8 9 10 11 12 Compound 3a 3b 3e 2a 2b 2d 2e 2g 2h 2i 2j 2l X H Cl Br OH OH OH OH OH OH OH OH OH Y H 4-Cl 4-Cl H 4-F 3-Cl 4-Cl 5-Cl 3-Me 5-Me 3-Cl 4-Br Z H H 5-Br H H H H H H H 5-Cl 5-Cl H-6 7.53–7.61 8.07 8.96 7.88 7.70 7.42–7.66 7.82 7.91–8.13 7.07 7.37 7.87 8.24 NH 7.76–7.86 9.01 8.40 9.24 9.66 8.36 9.47 9.47 8.59 8.94 9.67 9.87 OH — — — 8.99 9.47 7.81–8.00 9.13 9.31 8.10 8.59 —a 9.31 a Signal not identified in the spectrum.long-lived enough to react with a phenyl ring, since arylnitrenium –AlCl3 complexes 9 are reported to be more stable than the nitrenium ions themselves. Experimental Melting points are uncorrected and were taken on a Yanagimoto hot-stage melting point apparatus. 1H NMR spectra were measured on a JEOL JNM-PMX60SI spectrometer with tetramethylsilane (Me4Si) as an internal reference and CDCl3 as the solvent.J values are given in Hz. Infrared (IR) spectra were recorded on a JASCO IR810 spectrometer. Low and high resolution mass spectra (MS) were obtained with a JEOL JMS-DX300 spectrometer with a direct inlet system at 70 eV. Elemental analyses were performed in the microanalytical laboratory of this University. Compounds 1 were synthesized by benzoylation of phenoxyamines with benzoyl chloride in pyridine with ice cooling later at room temperature for periods ranging from several hours to overnight in 70–90% yields.Spectral data for new compounds are listed in Table 4. 1a, mp 138–140 8C (benzene), lit.,10 mp 137–138 8C (EtOH); 1d, mp 109.5–110 8C (benzene), lit.,10 mp 102–103 8C (CH2Cl2–hexane); 1e, mp 121–122 8C (benzene), lit.,10 mp 117–118 8C (CH2Cl2–hexane); 1g, mp 131– 133 8C (benzene), lit.,10 mp 137–138 8C (benzene); 2a, mp 171– 173 8C (EtOH), lit.,11 mp 169–170 8C; 2e, mp 233–235 8C Scheme 3 O H2O O HN + PhH O 1a H Ph TFSA AlCl3 O AlCl3 4a, b + 5b 2a (36%) + 3a (58%) H2O + and Ph N O – PhH PhH TFA Al + (AcOEt), lit.,12 mp 231–232 8C; 2f, mp 161–162 8C (AcOEt), lit.,13 mp 155 8C; 2g, mp 239–240 8C (AcOEt), lit.,14 mp 229– 230 8C; 2i, mp 198–201 8C (AcOEt), lit.,15 mp 191–192 8C; 2j, mp 215–218 8C (AcOEt), lit.,14 mp 215–216 8C; 3b, mp 118– 119 8C (benzene), lit.,16 mp 115 8C; 3c, mp 123–125 8C (AcOEt), lit.,17 mp 122 8C; 3d, mp 175–178 8C (benzene), lit.,18 mp 175 8C; 6, mp 81–82 8C (hexane), lit.,19 mp 80–81 8C; 8, mp 159–160 8C (benzene), lit.,20 mp 159–160 8C.Reaction of N-phenoxybenzamide derivatives (1) with AlCl3 in CH2Cl2. Typical procedure To 1a (200 mg, 0.94 mmol) in CH2Cl2 (8 cm3) was added AlCl3 (625 mg, 4.69 mmol) with cooling. After stirring the reaction mixture for 1 h at room temperature, 10% HCl (10 cm3) was added with cooling. The aqueous layer was extracted with CH2Cl2 (30 cm3 × 2), and the combined organic layer was washed with brine (30 cm3), dried over Na2SO4 and concentrated.The residue was chromatographed on a column of silica gel with benzene–ethyl acetate (10 : 1) as an eluent to give 2a (170 mg, 85%), mp 167–169 8C, which was recrystallized from EtOH, mp 171–173 8C (lit.,11 mp 169–170 8C). Physical and spectral data of all new compounds are listed in Tables 4 and 5. Reaction of N-phenoxybenzamide derivatives (1) with AlMe3 in CH2Cl2.Typical procedure To 1f (100 mg, 0.4 mmol) in CH2Cl2 (4 cm3) was added a solution of AlMe3 in hexane (1.01 M, 2.18 cm3, 2.18 mmol) with cooling. After stirring the reaction mixture for 4 h at room temperature, 10% HCl (10 cm3) was added with cooling. The aqueous layer was extracted with ethyl acetate (30 cm3 × 2), and the combined organic layer was washed with brine (30 cm3), dried over Na2SO4 and concentrated. The crude products were chromatographed on a flash column of silica gel.First elution with diethyl ether–light petroleum (1 : 2) afforded 1f (4 mg, 4%) and 2h (52 mg, 52%), mp 168–169 8C (AcOEt) (Found: C, 73.77; H, 5.80; N, 6.03. C14H13NO2 requires C, 73.99; H, 5.77; N, 6.16%); m/z 227 M1; nmax(KBr)/cm21 3320, 3260 (OH, NH), 1640 (CON); dH(CDCl3) 6.82 (1H, t, Ph), 6.93 (1H, d, Ph), 7.07 (1H, d, Ph), 7.47–7.65 (3H, m, Ph), 8.10 (1H, br s, NH), 8.59 (1H, s, OH). Further elution with the same solvent mixture (5 : 1) afforded 6 (8 mg, 13%), mp 80–81 8C (benzene) (lit.,19 mp 80–81 8C).10 J.Chem. Soc., Perkin Trans. 2, 1998 Table 4 Spectral data for new compounds Compound 1b 1c 1f 1h 1i 1j 2b 2c 2d 2k 2l 2m 3e 7 nmax/cm21 3150, 1660 3150, 1650 3130, 1650 3150, 1670 3120, 1660 3180, 1660 3420, 3050 1640 3400, 3250, 1650 3430, 3200, 1665 3330, 3060, 1640 3425, 3200, 1660 3290, 3080, 1640 3290, 1660 3120, 1670 dH 6.10–7.90 (9H, m, Ph) 9.83 (1H, br s, NH) 6.60–7.97 (9H, m, Ph) 9.12 (1H, br s, NH) 2.23 (3H, s, CH3) 6.50–7.93 (9H, m, Ph) 9.23 (1H, br s, NH) 6.73–7.93 (8H, m, Ph) 9.93 (1H, br s, NH) 6.03–8.07 (8H, m, Ph) 9.17 (1H, br s, NH) 6.73–8.00 (8H, m, Ph) 9.10 (1H, br s, NH) 6.65 (1H, ddd, J 9.5, 8.0, 2.8, Ph) 6.73 (1H, dd, J 9.5, 2.8, Ph) 7.48–7.66 (3H, m, Ph) 7.70 (1H, dd, J 8.0, 6.1, Ph) 7.96–8.09 (2H, m, Ph) 9.47 (1H, br s, NH) 9.66 (1H, br s, OH) 6.72 (1H, t, J 9.8, Ph) a 6.87 (1H, d, J 8.4, Ph) 7.05–7.16 (1H, m, Ph) 7.48–7.69 (3H, m, Ph) 7.87–7.99 (2H, m, Ph) 8.20 (1H, br s, NH) 9.76 (1H, br s, OH) 6.91 (1H, t, J 8.1, Ph) 7.08–7.22 (2H, m, Ph) 7.42–7.66 (3H, m, Ph) 7.81–8.00 (3H, m, Ph 1 NH) 8.36 (1H, br s, NH) 6.95 (1H, t, J 9.3, Ph) 7.14 (1H, dd, J 8.8, 4.8, Ph) 7.50–7.72 (3H, m, Ph) 7.94–8.07 (2H, m, Ph) 8.47 (1H, br s, NH) 9.89 (1H, br s, OH) 7.30 (1H, s, Ph) a 7.47–7.70 (3H, m, Ph) 7.99–8.07 (2H, m, Ph) 8.24 (1H, s, Ph) 9.31 (1H, br s, NH) 9.87 (1H, br s, OH) 7.08 (1H, d, J 8.6, Ph) 7.37 (1H, d, J 8.6, Ph) 7.51–7.76 (3H, m, Ph) 7.97–8.07 (2H, m, Ph) 8.45 (1H, br s, NH) 10.32 (1H, br s, OH) 7.50–7.67 (3H, m, Ph) 7.68 (1H, s, Ph) 7.89–7.95 (2H, m, Ph) 8.40 (1H, br s, NH) 8.96 (1H, s, Ph) 1.98 (3H, s, CH3) 2.25 (3H, s, CH3) 6.53–7.25 (4H, m, Ph) m/z (M1) 231 (M1, 15%) 247 (M1, 27%) 249 (M1 1 2, 10%) 227 (M1, 39%) 281 (M1, 11%) 283 (M1 1 2, 7%) 285 (M1 1 4, 1%) 309 (M1, 7%) 311 (M1 1 2, 7%) 369 (M1, 6%) 371 (M1 1 2, 11%) 373 (M1 1 4, 6%) 231 (M1, 16%) 231 (M1, 15%) 247 (M1, 17%) 249 (M1 1 2, 6%) 309 (M1, 8%) 311 (M1 1 2, 8%) 325 (M1, 5%) 327 (M1 1 2, 7%) 329 (M1 1 4, 2%) 369 (M1, 3%) 371 (M1 1 2, 6%) 373 (M1 1 4, 3%) 387 (M1, 3%) 389 (M1 1 2, 7%) 391 (M1 1 4, 5%) 393 (M1 1 6, 1%) 165 (M1, 46%) a [2H6]Acetone. Reaction of N-(4-methylphenoxy)benzamide (1g) with BF3?OEt2 in CH2Cl2 To 1g (200 mg, 0.88 mmol) in CH2Cl2 (8 cm3) was added BF3?OEt2 (0.54 cm3, 4.40 mmol) with cooling.After stirring the reaction mixture for 3 days at room temperature, 10% HCl (10 cm3) was added with cooling.The aqueous layer was extracted with ethyl acetate (50 cm3 × 2), and the combined organic layer was washed with brine (50 cm3), dried over Na2SO4 and concentrated. The residue was chromatographed on a column of silica gel with benzene–ethyl acetate (20 : 1) as an eluent to give 5a (124 mg, 62%), mp 169–171 8C (MeOH–H2O) (Found: C, 73.55; H, 5.26. C14H12O3 requires C, 73.67; H, 5.30%); m/z 228J. Chem. Soc., Perkin Trans. 2, 1998 11 Table 5 Physical constants and microanalytical data of new compounds Found (%) Required (%) Compound 1b 1c 1f 1h 1i 1j 2b 2c 2d 2k 2l 2m 3e 7 Mp/8C 92–93 141–143 139–142 132–133 123–123.5 145–146 220–222 145–146 141–143 125–126 248–251 177.5–178 178–180 144–146 Solvent AcOEt–hexane benzene benzene benzene benzene benzene AcOEt AcOEt AcOEt AcOEt AcOEt benzene AcOEt benzene Molecular formula C13H10FNO2 C13H10ClNO2 C14H13NO2 C13H9Cl2NO2 C13H9BrFNO2 C13H9Br2NO2 C13H10FNO C13H10FNO C13H10ClNO2 C13H9BrFNO C13H9NO2BrCl C13H9Br2NO C13H8Br2ClNO C9H11NO2 C 67.41 63.02 73.92 55.15 50.28 42.09 67.51 67.42 62.96 50.24 47.77 42.02 40.02 65.36 H 4.48 4.23 5.84 3.33 2.99 2.53 4.45 4.49 4.22 3.03 2.86 2.49 2.28 6.71 N 5.96 5.56 6.11 4.91 4.32 3.58 5.89 5.93 5.72 4.45 4.28 3.72 3.64 8.37 C 67.53 63.04 73.99 55.35 50.35 42.08 67.53 67.53 63.04 50.35 47.81 42.08 40.31 65.44 H 4.36 4.07 5.77 3.22 2.93 2.45 4.36 4.36 4.07 2.93 2.78 2.45 2.07 6.71 N 6.06 5.66 6.16 4.96 4.52 3.78 6.06 6.06 5.66 4.52 4.29 3.78 3.62 8.48 (M1); nmax(KBr)/cm21 3400 (OH), 1725 (COO); dH(CDCl3) 2.31 (3H, s, CH3), 5.29 (1H, s, OH), 6.96–7.03 (3H, m, Ph), 7.48–7.59 (2H, m, Ph), 7.61–7.72 (1H, m, Ph), 8.19–8.26 (2H, m, Ph).Reaction of N-phenoxybenzamide (1a) with AlCl3 in benzene To 1a (200 mg, 0.94 mmol) in benzene (8 cm3) was added AlCl3 (625 mg, 4.69 mmol) with cooling. After stirring the reaction mixture for 1 h at room temperature, 10% HCl (10 cm3) was added with cooling. The aqueous layer was extracted with ethyl acetate (30 cm3 × 2), and the combined organic layer was washed with brine (30 cm3), dried over Na2SO4 and concentrated. The crude products were chromatographed on a column of silica gel.First elution with benzene–ethyl acetate (20 : 1) afforded 3a (101 mg, 58%), mp 162–164 8C (lit.,21 mp 164– 166 8C). Further elution with the same solvent mixture afforded 2a (71 mg, 36%), mp 168–169 8C (lit.,11 mp 169–170 8C). References 1 Y. Kikugawa and M.Shimada, J. Chem. Soc., Chem. Commun., 1989, 1450. 2 R. A. Abramovitch and R. Jeryaraman, in Azides and Nitrenes, ed. E. F. V. Scriven, Academic Press, New York, 1984, pp. 297–357. 3 E. Miyazawa, T. Sakamoto and Y. Kikugawa, Org. Prep. Proced. Int., 1997, 29, 594, and references cited therein. 4 G. A. Olah, B. P. Singh and G. Liang, J. Org. Chem., 1984, 49, 2922; G. T. Tisue, M. Grassmann and W. Lwowski, Tetrahedron, 1968, 24, 999; L. H. Horner and H. Steppan, Liebigs Ann.Chem., 1957, 606, 24. 5 (a) A. Ribera and M. Rico, Tetrahedron Lett., 1968, 535; (b) M. Zanger, W. W. Simons and A. R. Gennaro, J. Org. Chem., 1968, 33, 3673. 6 R. F. C. Brown, L. Radom, S. Sternhell and I. D. Rae, Can. J. Chem., 1968, 46, 2577; J. R. Bartels-Keith and R. F. W. Cieciuch, Can. J. Chem., 1968, 46, 2593; B. D. Andrews, I. D. Rae and B. E. Reichert, Tetrahedron Lett., 1969, 1859; J. M. Appleton, B. D. Andrews, I. D. Rae and B. E. Reichert, Aust. J. Chem., 1970, 23, 1667; G. W. Gribble and F. P. Bousquet, Tetrahedron, 1971, 27, 3785. 7 K. Paluch, Rocz. Chem., 1965, 39, 744 (Chem. Abstr., 1965, 63, 8250). 8 Y. Endo, K. Shudo and T. Okamoto, J. Am. Chem. Soc., 1982, 104, 6393. 9 H. Takeuchi, M. Maeda, M. Mitani and K. Koyama, J. Chem. Soc., Perkin Trans. 1, 1987, 57. 10 Y. Endo, K. Shudo and T. Okamoto, Synthesis, 1980, 461. 11 A. Ito, Yakugaku Zassi, 1962, 82, 875. 12 J. T. Edward, J. Chem. Soc., Chem. Commun., 1956, 222. 13 G. R. Clemo and A. F. Daglish, J. Chem. Soc., Chem. Commun.. 1950, 1481. 14 R. Adams and J. M. Stewart, J. Am. Chem. Soc., 1952, 74, 5876. 15 L. C. Raiford and J. Couture, J. Am. Chem. Soc., 46, 2305. 16 F. D. Chattaway, K. J. P. Orton and W. H. Hurtley, Chem. Ber., 1899, 32, 3635. 17 J. P. Wibaut, Recl. Trav. Chim. Pays-Bas, 32, 244 (Chem. Abstr., 1914, 8, 1268). 18 M. P. Grammaticakis, Bull. Soc. Chim. Fr., 1949, 761. 19 A. W. Titherley, J. Chem. Soc., 1901, 79, 391. 20 E. Nöelting and O. Kohn, Ber., 1884, 17, 351. 21 C. G. Derick and J. H. Bornmann, J. Am. Chem. Soc., 1913, 35, 1269. Paper 7/06488H Received 16th June 1997 Accepted 8th September 1997

ISSN:1472-779X    

DOI:10.1039/a706488h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1998 13 Elimination mechanisms in the anilinolysis of sulfamoyl chlorides in chloroform and acetonitrile William J. Spillane,* Francis A. McHugh and Padraig O. Burke Chemistry Department, National University of Ireland, Galway, Ireland The kinetics of the reaction of various sulfamoyl chlorides, R1R2NSO2Cl (R1 5 Ph, Me, c-C6H11, But, R2 5 H and R1 5 R2 5 Me and R1 5 R2 5 PhCH2) with anilines in chloroform and acetonitrile have been studied. Reaction is first order in both chloride and aniline and pseudo first order and second order rate constants have been determined for various chlorides and para- and meta-substituted anilines.Hammett Ò values vary from 24.76 to 22.63 for varying XC6H4NH2 depending on the halide, X. Substituents in XC6H4NHSO2Cl have only small effects on the rates. There is a ca. 106-fold rate difference between PhNHSO2Cl and (PhCH2)2NSO2Cl reacting with p-anisidine in chloroform at 25 8C. This finding, coupled with the observation of hydrogen/deuterium isotope effects, supports the operation of an elimination mechanism involving an N-sulfonylamine, [PhN]] SO2], as a transient species.In chloroform an E2-type mechanism is suggested, while in acetonitrile the activation parameters (DH, DS) may indicate a more E1cB-like E2 mechanism. For quite a few years there has been widespread interest in the kinetics and mechanism of reactions of sulfonyl halides in hydrolysis, alcoholysis (including phenolysis) and aminolysis processes (Scheme 1) leading to sulfonic acid, sulfonate ester and sulfonamide products, respectively.This area has been authoritatively reviewed.1 A number of mechanisms have been supported depending on substrate, reagents and solvents and these range from nucleophilic substitutions (SN) through addition–eliminations (SAN) to eliminations (E1, E2 and E1cB variants). Studies on the mechanism(s) of reaction of the related sulfamoyl halides R = RNH or R1R2N in 1 [reaction (1)] have not RSO2X 1 Y2 æÆ RSO2Y 1 X2 1 (1) been extensive despite their importance as intermediates in the pharmaceutical and agrochemical industries 2 and their utility in the synthesis of sulfamides.3 The hydrolysis of N,N-disubstituted sulfamoyl halides R1R2NSO2Cl has been controversial, with both SN1 and SN2 processes being favoured by different groups.4a–f The kinetics of halide exchange using a 36Cl label have been reported for N,Ndimethylsulfamoyl chloride.4g We could only find one piece of kinetic information on the reaction of a sulfamoyl halide with an amine, namely the rate of reaction of N,N-dimethylsulfamoyl chloride with piperidine in aqueous dioxane giving N,N-dimethyl-N9-piperidylsulfamide.4a This paucity of information on sulfamoyl group transfers to nitrogen centres prompted us to probe the kinetics and mechanism of such transfers using a series of anilines and a number of sulfamoyl chlorides. Results and discussion Preliminary studies using N-phenylsulfamoyl chloride and aniline at varying concentrations in chloroform showed that the UV kinetics were first order in chloride and in aniline.Second order rate constants were obtained from plots of kobs, the pseudo first-order rate constant, determined using excess amine from plots of kobs vs. [aniline]. The observed second order kinetics are compatible with the following mechanistic types: SN2, SAN, E2 and E1cB. However, the pronounced acidity of the N]H group in monosubstituted sulfamates compared to that of the C]H group in phenylmethylsulfonate esters, such as PhCH2SO2OC6H4CH3-p,5 which are well known to react via elimination mechanisms involving sulfenes,6 suggests that sulfamoyl halides will be very likely to involve elimination pathways in their decomposition.King has shown that the analogous phenylmethylsulfonyl halides, PhCH2SO2X react with aniline and other amines via an elimination path involving a sulfene.7 Williams 8 has determined the acidities of the aryl methylsulfamates, MeNHSO2OAr in 50% EtOH]H2O as being in the range 10.53 (Ar = C6H5) to 8.70 (Ar = m-NO2C6H4).We feel that the pKa of PhNHSO2Cl may be lower, ca. 7–8. This is based on comparison of this system with the sulfonamides shown. The pKa values of the latter have been determined in water.9 Comparison of the sp value for chlorine (0.227) with the s values for a series of C6H4X groups indicated that the electronic demands of Cl would probably be best matched by p-nitrophenyl, which has sp = 0.23. The pKa of PhNHSO2- C6H4NO2-p was found to be 7.50 in water and it is reasonable to consider that the pKa of phenylsulfamoyl chloride should be ca. 7–8 in H2O and ca. 9 in 50% EtOH.10 PhNHSO2Cl PhN _ SO2Cl 1 H1 PhNHSO2C6H4X PhN _ SO2C6H4X 1 H1 Evidence for an elimination mechanism The data in Table 1 support the operation of an elimination mechanism in the aminolysis of monosubstituted sulfamoyl chlorides, which are seen to react ca. 106 times more rapidly than disubstituted sulfamoyl chlorides. The latter probably react via bimolecular nucleophilic attack by the amine at the sulfur of the sulfamoyl centre. This test has also been used to show that monosulfamate esters, RNHSO2OR9 react by a Table 1 Second order rate constants for the reaction of some monoand disubstituted sulfamoyl chlorides R1R2NSO2Cl with p-anisidine in chloroform at 25 8C R1 Ph Bu Ph PhCH2 R2 HH Me PhCH2 k2/dm3 mol21 s21 10.3 13.9 1.3 × 1025 5.8 × 102614 J.Chem. Soc., Perkin Trans. 2, 1998 different mechanism to disulfamate esters R2NSO2OR9.8,11 Further support for an elimination mechanism possibly involving an N-sulfonylamine [PhN]] SO2], came from the reaction of N-phenyl sulfamoyl chloride (1, R = PhNH, X = Cl) with panisidine- ND2 (>90% D) (10 fold excess) in deuteriated chloroform. The product formed, N-phenyl-N9-p-anisylsulfamide, was fully deuteriated on both nitrogens.Had substitution taken place by amine attack at sulfur, a product containing only one deuterium would be expected, i.e. PhNHSO2NDC6H4OMe-p. Fig. 1 shows the 1H NMR spectra for the products of reaction when (a) non-deuteriated and (b) deuteriated materials were used. The possibility that the deuteriated anisidine might exchange with the starting sulfamoyl chloride or the product sulfamide seems unlikely since it is a weak base. This is supported by the fact that in our attempts to fully deuteriate panisidine, equilibration with DCl in D2O for two days at room temperature resulted in only ca. 40% incorporation of deuterium. We synthesised the fully deuteriated material by a phase transfer method instead (see Experimental section). Further support for an elimination mechanism comes from a ‘reverse’ isotope experiment in which we probed the hydrogen deuterium kinetic isotope effect by synthesising PhNDSO2Cl. Table 2 records the observed kH and kD values and the ratio kH/kD.The observation of these effects effectively rules out an (E1cB)rev mechanism and would point towards either an E2 or an (E1cB)irrev mechanism (Scheme 1). An (E1cB)rev mechanism involves a non-rate determining deprotonation equilibrium for which a kinetic isotope effect would not be expected. Substituent eVects Hammett Ò values. The effects of varying substituents both in aniline and in sulfamoyl chloride have been probed.The results for various anilines are displayed in Tables 3–5, Table 6 (runs 1–4) and Table 7 and for sulfamoyl chloride in Table 8 and Table 6 (runs 2, 5 and 6). For the seven anilines in Table 3, a Hammett r of 23.57 (correlation coefficient, r = 0.975; stand. error = 0.364) at 25 8C in chloroform was obtained, in excellent agreement with the r obtained from the more limited plot for the four anilines (runs 1–4) in Table 6 which gave a r of 23.54 (r = 0.997; stand. error = 0.181). Changing the sulfamoyl chloride from N-phenyl- or N-methyl- to p-methylphenyl- and p-chlorophenyl- gave r values of 24.76 (r = 0.99; stand.Scheme 1 ArNH2 + PhNHSO2Cl ArNH2 H Ph N SO2 .... .... d+ d– ....Cl d– k1 k2 ArNH3•Cl– + [PhN=SO2] PhNHSO2NHAr ArNH2 fast E2 + ArNH2 + PhNHSO2Cl ArNH3 + PhNSO2Cl + – k1 ArNH3Cl– + [PhN=SO2] PhNHSO2NHAr ArNH2 fast E1cB + –Cl– k2 Table 2 Kinetic isotope effects for the reaction of N-phenylsulfamoyl chloride with anilines (XC6H4NH2) in chloroform at 25 8C X p-OMe p-Me H kH a/dm3 mol21 s21 10.3 3.1 1.0 kD b/dm3 mol21 s21 3.9 3.5 0.78 0.86 0.19 kH/kD 2.6 2.9 4.0 3.6 5.3 a From Table 3.b Using PhNDSO2Cl (97% D by 1H NMR spectroscopy) and CDCl3. error = 0.171) for the six anilines in the Table reacting with N-phenylsulfamoyl chloride. Large negative r values, usually in the range of 22.0 to 22.9 for the aminolysis of sulfonyl halides are common1 and more negative values have been reported. The reaction of phenylmethanesulfonyl halides with anilines (involving an elimination pathway with a sulfene intermediate 7 and a substrate that may be viewed as the carbon analogue of phenylsulfamoyl chloride) gives r values of ca. 23.50 for variation of the aniline base with a number of ring-substituted phenylmethanesulfonyl Table 3 Second order rate constants for the reaction of N-phenylsulfamoyl chloride with anilines (XC6H4NH2) in chloroform at 25 8C X p-OMe p-OEt p-Me Hm -OMe p-Cl p-Br l/nma 305 305 294 287 287 314 298 k2/dm3 mol21 s21 10.3 7.8 3.1 1.0 0.39 0.31 0.08 log k2 1.014 0.892 0.495 0.002 20.412 20.507 21.1 sb 20.2688 20.24 20.17 0.0 0.115 0.227 0.232 pKa c 5.34 5.20 5.08 4.60 4.23 3.98 3.86 a Analytical wavelength used.b D. H. McDaniel and H. C. Brown, J. Org. Chem., 1958, 23, 458. 0.23 for p-nitrophenyl (see text) from C. Hansch and A. J. Lee, Substituent Constants for Correlation Analysis in Chemistry and Biology, Wiley-Interscience, New York, 1979, p. 128. c Refs B66, H16 (for p-OEt) and F14 (for p-CN in Table 7), D. D. Perrin, Dissociation Constants of Organic Bases in Aqueous Solution, Butterworths, London, 1965. Table 4 Pseudo-first order rate constants for the reaction of p-methylphenylsulfamoyl chloride with anilines (XC6H4NH2) in chloroform at 25 8Ca X p-OMe p-OEt p-Me Hm -OMe p-Cl kobs/s21 0.88 0.65 0.17 0.67 0.012 0.003 log kobs 20.56 20.187 20.769 21.187 21.921 22.511 s 20.268 20.24 20.17 0 0.115 0.227 a [Aniline] 6 × 1024 mol dm23, [sulfamoyl chloride] 3.75 × 1025 mol dm23.Table 5 Pseudo first order rate constants for the reaction of pchlorophenylsulfamoyl chloride with anilines (XC6H4NH2) in chloroform at 25 8Ca X p-OMe p-OEt p-Me Hm -OMe p-Cl kobs 0.71 0.52 0.41 0.12 0.07 0.04 log kobs 20.150 20.282 20.384 20.907 21.186 21.45 s 20.268 20.24 20.17 0.0 0.115 0.227 a Concentrations of reactants as in Table 4 footnote. Table 6 Pseudo-first order rate constants for the reaction of sulfamoyl chlorides (RNHSO2Cl) with anilines (XC6H4NH2) in chloroform at 20 8Ca Run 123456 R Me c-C6H11 Ph X Hp -OMe p-Cl p-CN p-OMe p-OMe l/nm 287 311 299 270 305 305 kobs/1023 s21 0.34 3.2 0.04 0.0017 b 18.0 34.6 a [Aniline] 12 × 1024 mol dm23, [sulfamoyl chloride] 0.75 × 1024 mol dm23.b Estimated (see Experimental section).J. Chem. Soc., Perkin Trans. 2, 1998 15 chlorides.12 The r values of 23.57 (phenylsulfamoyl chloride) and 23.54 (methylsulfamoyl chloride) are strikingly close to this value for the related phenylmethanesulfonyl chloride system.Changing from chloroform to acetonitrile produces a lower r value of 23.09 and this would be expected on the basis of the change to a more polar solvent: the relative permittivities (e) of chloroform and acetonitrile are 4.81 and 35.9, respectively. 13 For phenylmethanesulfonyl chloride the r values were measured in MeOH–CH3CN mixtures varying from 100% MeOH (e = 32.7 13) to 50% MeOH (v/v). These r values determined here indicate that the anilines play a major role by proton abstraction in an E2 transition state or in the first step (k1) of an (E1cB)irrev mechanism (Scheme 1).The much lower and higher r values measured for p-chlorophenyland p-methylphenylsulfamoyl chlorides of 22.63 (Table 5) and 24.76 (Table 4), respectively can be understood in terms of relative acid strength. The p-chloro compound should be more acidic than the parent sulfamoyl chloride and a diminished role can be envisaged for the anilines since the N–H is already weakened.However, in the p-methyl compound the hydrogen of the N–H bond should be more tightly held and consequently the bases will have a greater role. In Table 8 the effects of a limited study of the variation of rates with varying substitutent in XC6H4NHSO2Cl are presented. In chloroform the reaction is virtually insensitive to changing substituent and interestingly a similar result was obtained in the hydrolysis in aqueous dioxane at 29 8C of the sulfamoyl azides (RNHSO2N3, R = Ph, p-ClC6H4 and p- MeOC6H4), which gave rates of 8.75, 6.68 and 8.68 (×102 min21) respectively.14 In acetonitrile in our system there is a notable increase in rate with a change from an electrondonating to an electron-withdrawing substituent (Table 8).Also, it is noticeable that the rates are considerably faster in acetonitrile: kCH3CN/kCHCl3 for p-Me, H and p-Cl being respectively 19.5, 20.1 and 41.4.The more polar solvent facilitates reaction and in the case of p-chlorophenylsulfamoyl chloride a 40-fold acceleration is found. The extra rate enhancement for the chloro compound may well be due to the fact that in either an E2 or (E1cB)irrev mechanism the chloro substituent will stabilise an intermediate N-sulfonylamine, whereas a p-methyl group or the unsubstituted parent compound would not provide such stabilisation; hence their 20-fold acceleratory action may be due simply to the change to the more polar solvent.In Table 6 an approximately tenfold acceleration is noted from runs 2, 5 and 6 for a change in sulfamoyl chloride in the Table 7 Second order rate constants for the reaction of Nphenylsulfamoyl chloride with anilines (XC6H4NH2) in acetonitrile at 25 8C X p-Me Hm -OMe p-Cl p-CN p-NO2 k2/dm3 mol21 s21 89.1 20.3 11.2 2.68 0.28 0.08 log k2 1.95 1.31 1.05 0.428 20.55 21.12 sa 20.17 0 0.115 0.227 0.66 0.778 pKa a 5.08 4.60 4.23 3.98 1.74 1.00 a For s and pKa values see Table 3 footnotes b and c.Table 8 Second order rate constants for the reaction of parasubstituted phenylsulfamoyl chlorides (XC6H4NHSO2Cl) with aniline at 25 8C in chloroform and with aniline in acetonitrile a Chloroform Acetonitrile X p-OMe p-Me Hp -Cl k2/dm3 mol21 s21 0.98 0.88 1.01 0.84 X p-Me Hp -Cl k2/dm3 mol21 s21 17.1 20.3 34.8 reaction with p-anisidine in chloroform at 20 8C. These data gave a very poor Taft plot when log kobs was plotted vs.s*. The rate constants do, however, increase with increasing s*. Brønsted ‚ values. Brønsted b values have been determined for the data in Tables 3 and 7 using pKa values measured in water (see Table 3, footnote c). In CHCl3 at 25 8C the bnuc for aniline attack is 1.27 (n = 0.982, stand. error = 0.11) and in CH3CN at 25 8C the value is 0.685 (n = 0.980, stand. error = 0.076). These Brønsted coefficients could be substantially in error since the rate data in the Tables should of course be plotted using pKa values measured in the same solvents in which the rates were measured.At least, however, it may be taken that the bnuc values for the reaction are reasonably positive. In agreement with this, bnuc values of ca. 1.2 have been reported for the anilinolysis of phenylmethanesulfonyl chloride in 100% MeOH, 80% and 50% MeOH–CH3CN (v/v) 12 and of ca. 0.6–1.1 for the elimination reaction of 2-arylethylquinuclidinium ions, ArCH2CH2 1X in 60% DMSO–H2O (v/v).15 Using pKa values for anilinium ions, determined in dimethyl sulfoxide (DMSO),16 an approximate straight line is obtained when they are plotted against those determined in water. The values for p-Cl and p-NO2 are available in DMSO,16 and the other four pKa values in DMSO required for the data in Table 7 were obtained by interpolation (for m-OMe and p-CN) and by a slight extrapolation (for p-Me and H).A plot of log k2 values in Table 7 vs. these pKa values gave a slope of 0.62 (r = 0.976, stand.error = 0.069). Since the relative permittivity of DMSO (e = 46.5 13) is closer to that of acetonitrile (e = 35.9) than H2O (e = 78.3 13) such a plot may be more meaningful and seems to point to a reasonable degree of proton transfer in the slow step of the elimination. Activation data Table 9 contains DH‡ and DS‡ values for the reaction of phenylsulfamoyl chloride with anilines in chloroform and acetonitrile. The lower enthalpies in acetonitrile are immediately clear and the ca. 20-fold times faster reaction in this solvent has been referred to above. Acetonitrile, being more polar than chloroform, can provide greater stabilisation of the transition state. However, the lower enthalpies may indicate a shift to a more E1cB type mechanism, where proton removal is an important process occurring in the transition state. The enthalpies associated with proton dissociation of various anilinium ions in water lie in the range of ca. 16–30 kJ mol21.17 Some data for different temperatures are plotted in Fig. 2. The entropy data for acetonitrile compared to that for chloroform may also indicate a move to a mechanism with a greater charge build-up in the transition state and thus a somewhat greater change in entropy on going from ground state reactants to activation complex. The evidence in support of E2 or (E1cB)irrev mechanisms has been given above. A priori the involvement of an NTable 9 Activation parameters for the reaction of N-phenylsulfamoyl chloride with anilines in chloroform and acetonitrile Solvent CHCl3 a CH3CNd Substituent p-OMe p-OEt p-Me Hb m-OMeb p-Cl c Hm -OMe p-Cl DH‡/kJ mol21 24.3 ± 4 23.2 ± 2 23.4 ± 4 30.2 ± 2 53.6 ± 5 50.5 ± 7 17.7 ± 9 26.5 ± 3 37.3 ± 1 DS‡/J K21 mol21 2141.7 ± 12 2151.9 ± 5 2154.5 ± 12 2143.5 ± 6 273.2 ± 2 285.1 ± 23 2160.4 ± 1 2136.2 ± 7 2112.0 ± 1 a Temperature range 298–313 K (four temperatures). b Temperature range 298–313 K (three temperatures).c Temperature range 298–317.5 K (three temperatures).d Temperature range 298–314 K (four temperatures).16 J. Chem. Soc., Perkin Trans. 2, 1998 sulfonylamine is very likely. The general conditions for generation of an N-sulfonylamine 18 correspond to the reaction conditions used in this work. The likely mechanistic possibilities are shown in Scheme 1. The reaction may involve a central E2 type mechanism where departure of chloride and removal of the amino hydrogen in the sulfamoyl chlorides may both be concerted, or an (E1cB)irrev mechanism where a negative charge is first deposited on nitrogen prior to carbon–chlorine bond cleavage. In acetonitrile an E1cB-like E2 mechanism is proposed and in chloroform the mechanism may be shifted to a more E2 type.Experimental Solvents and starting materials Chloroform was refluxed over PCl5 for 12 h and fractionally distilled a number of times until a satisfactory low water content, which was determined by a Dean–Stark apparatus, was obtained.The solvent used in kinetic runs was distilled weekly and stored in an opaque container. Pyridine was left to stand for 24 h over KOH and then refluxed over fresh KOH, following which it was fractionally distilled. The first and last 10% of the distillate were discarded. Benzene stock solution was stored over CaCl2 and was refluxed and fractionally distilled over CaH2. It was stored over sodium wire. Acetonitrile was refluxed and distilled over CaH2 and stored in a dark container.Light petroleum was fractionally distilled and stored over sodium wire at least 72 h before use. Anilines. All liquid amines were refluxed over KOH and then fractionally distilled through an efficient Vigreux column, being sure to discard the first and last 15% of the distillate. All solid amines were sublimed or distilled under reduced pressure using a Kugel–Rohr distillation unit. All amines were stored under a nitrogen atmosphere at ca. 5 8C. Sulfamoyl chlorides. These were prepared by the method of Kloek and Leschinsky,19 except for N-methylsulfamoyl chloride, which was prepared by the method of Weiss and Schulze.20 Deuteriated materials. p-Anisidine-ND2 was prepared as follows: p-anisidine (0.5 g, 4.1 mmol) was dissolved in CDCl3 (10 ml) together with tetramethylammonium chloride (0.022 g, 0.2 mmol). To this was added 10 ml of D2O and sodium deuteroxide (0.008 g, 0.2 mmol); the reaction mixture was stirred for 24 h following which time the organic and aqueous layers were separated and the aqueous layer was replaced with a fresh batch of 10 ml of D2O, and the procedure was repeated.The layers were then separated and the chloroform was removed under vacuum. The deuteriated amine was purified by bulb to bulb distillation. The percentage of deuterium was determined by 1H NMR spectroscopy to be 94%. PhNDSO2Cl was prepared in >97% yield as follows: phenylsulfamoyl chloride (4.0 g, 0.021 mol) was added slowly to a solution of 15 ml of D2O containing NaOD (0.021 mol).This solution was heated to 40 8C for 2 h to ensure the complete hydrolysis of the chloride to the sulfamic acid. The D2O was removed under vacuum and the remaining sulfamate salt was recrystallised from 4: 1 CH3CN–D2O. The percentage of deuterium incorporated into the sodium phenylsulfamate salt was estimated by 1H NMR spectroscopy to be 98%. This was then reacted in the usual way in [2H6]benzene with phosphorus pentachloride to form the N-deuteriated phenylsulfamoyl chloride.Preparation of N-phenyl-N9-p-anisylsulfamide Non-deuteriated material. N-Phenylsulfamoyl chloride (0.25 g, 0.0013 mol) was dissolved in dry chloroform (50 ml) and p-anisidine (0.32 g, 0.0026 mol) was added. The solution was stirred for 2 h at room temperature, by which time complete reaction had occurred. The p-anisidine hydrochloride which forms precipitated from solution and was filtered off; removal of the solvent under vacuum gave crude product sulfamide (0.341 g, 94%).This was flash chromatographed to yield the pure product (87%). dC 55.190, 114.186, 117.919, 121.904, 122.490, 128.929, 130.667, 138.433, 155.831 ppm. 1H NMR spectrum in [2H6]DMSO, see Fig. 1(a). Deuteriated material. This reaction was conducted as in the previous experiment, except that a 10 : 1 molar ratio of p-anisidine-ND2 to phenylsulfamoyl chloride was used. The product sulfamide was redissolved in [2H6]DMSO (in CDCl3 the 1H NMR signals for the amino hydrogens fall under the aromatic peaks) and its 1H NMR spectrum was recorded, see Fig. 1(b). Two deuterium atoms, one on each of the nitrogens, are indicated and thus the reaction product is PhNDSO2NDC6H4OMe- p. Kinetic measurements The reactions between various aromatic amines and the sulfamoyl chlorides were carried out at 25 ± 0.2 and 20 ± 0.2 8C unless otherwise stated in the text and were monitored by UV spectroscopy using a Shimadzu 260 instrument. Reactions were generally followed by monitoring the decrease in absorbance of the amine peak at a suitable wavelength (see Tables 3 and 6).The reactions were started by the addition of 1 ml of a sulfamoyl chloride stock solution to an equal volume of the appropriate amine solution equilibrating in a quartz cuvette in the sample holder of the instrument. The initial sulfamoyl chloride concentration was 3.75 × 1025 mol dm23 (runs at 25 8C) and 0.75 × 1024 mol dm23 (runs at 20 8C).Amine concentrations varied from 4.5 × 1024 mol dm23 to 24 × 1024 mol dm23. Plots of |At 2 A•| as a function of time were found to be linear for over 90% reaction (>3 half lives). From 15 to 20 points were used in each plot. Second order rate constants were determined from plots of kobs as a function of amine concentration, the best line being determined by least-squares analysis. Some typical plots are shown in Fig. 2.Rate constants (Tables 1–8) were accurate to within ±5% based on replicate runs. The reaction of p-cyanoaniline in CHCl3 with MeNHSO2Cl was extremely slow and the spectral changes were very small (thus the method of initial rates was not used). The pseudo first-order Fig. 1 1H NMR spectra for (a) PhNHSO2NHC6H4OCH3-p and (b) PhNDSO2NDC6H4OCH3-pJ. Chem. Soc., Perkin Trans. 2, 1998 17 rate constant for this reaction, kobs (Table 6), was estimated as follows: (i) MeNHSO2Cl (1.0 × 1024 mol dm23) was reacted with p-anisidine (4.0 × 1024 mol dm23) in chloroform at 20 8C, a half-life of 4.6 min being observed; (ii) the sulfamoyl chloride (0.025 mol dm23) was reacted with p-cyanoaniline (0.1 mol dm23) in chloroform at 20 8C, a half-life of 35 min being observed; the reaction was performed in 100 ml of solution, with appropriate dilution of sample in chloroform prior to absorbance readings.The above results (i) and (ii) indicate that p-cyanoaniline is 1902 (35/4.6 × 250) times less reactive than panisidine towards N-methylsulfamoyl chloride.Thus an estimated rate constant for the reaction of N-methylsulfamoyl chloride (0.75 × 1024 mol dm23) and p-cyanoaniline (12.0 × 1024 mol dm23) in chloroform at 20 8C is 1.003 × 1024 min21 (0.190 76/1902 min21) or 0.001 67 × 1023 s21. Product runs Since the reaction of sulfamoyl chlorides with amines is a well documented synthetic route to unsymmetrical sulfamides, extensive product studies were not carried out.TLC analysis, to ensure that clean reactions were occurring, was performed, however. Bearing in mind the careful drying of the solvents and amines which was carried out the formation of sulfamic acids by reaction of H2O with the sulfamoyl chlorides would be negligible. Product runs with N-methylsulfamoyl chloride (5 mmol) and with aromatic amine (10 mmol) in chloroform (2 ml) were carrried out by refluxing for 40 min. The usual work-up for sulfamides gave from the appropriate amine the following: Nmethyl- N9-phenylsulfamide, 71%, mp 80–81 8C, N-methyl- Fig. 2 Plots of pseudo first order rate constants as a function of amine concentration at four different temperatures for the reaction of phenylsulfamoyl chloride (3.75 × 1025 mol dm23) with p-ethoxyaniline (4.5– 12 × 1024 mol dm23) in chloroform N9-p-anisylsulfamide, 96%, mp 86–87 8C and N-methyl-N9-pchlorophenylsulfamide, 97%, mp 112–113 8C. The yields are yields of crude product.A product run with phenylsulfamoyl chloride (0.25 g, 1.3 × 1023 mol) and p-anisidine (0.32 g, 2.6 × 1023 mol) in CHCl3 (50 ml) was carried out at room temperature over 2 h and 94% crude sulfamide was formed. After purification an 87% yield of pure sulfamide giving C, H and N microanalyses within ±5% and good 13C and 1H NMR spectra (see above) was obtained. Sulfamides were purified by recrystallisation using EtOH–H2O (50/50 v/v) or flash chromatography. Characterisation was by IR spectroscopy, mp or 13C/1H NMR spectroscopy.References 1 I. M. Gordon, H. Maskill and M.-F. Ruasse, Chem. Soc. Rev., 1989, 18, 123. 2 G. A. Benson and W. J. Spillane, Chem. Rev., 1989, 80, 151; G. A. Benson and W. J. Spillane, Sulphamic Acid and Derivatives, in The Chemistry of Sulphonic Acids, Esters and their Derivatives, eds. S. Patai and Z. Rappoport, Wiley, 1991, ch. 22, pp. 947–1036. 3 S. D. McDermott and W. J. Spillane, Org. Prep. Proced. Int., 1984, 16, 49. 4 (a) H. K. Hall, J. Am. Chem. Soc., 1956, 78, 1450; (b) H. K. Hall and C. H. Lueck, J. Org. Chem., 1963, 28, 2818; (c) O. Rogne, J. Chem. Soc. B, 1969, 663; (d ) E. C. F. Ko and R. E. Robertson, J. Am. Chem. Soc., 1972, 94, 573; (e) E. C. F. Ko and R. E. Robertson, Can. J. Chem., 1972, 50, 946; ( f ) B. C. Lee and I. Lee, Taehan Hwakakhee Chi, 1980, 24, 342 (Chem. Abstr., 1981, 94, 102 400); (g) I. Lee and S. C. Kim, J. Korean Chem. Soc., 1973, 17, 406. 5 A. Williams and K. T. Douglas, Chem. Rev., 1975, 75, 627. 6 S. Thea and A. Williams, J. Chem. Soc., Perkin Trans. 2, 1981, 72 and references therein. 7 (a) J. F. King and S. Skoneiczny, Tetrahedron Lett., 1987, 28, 5001; (b) J. F. King, Acc. Chem. Res., 1975, 8, 10. 8 K. T. Douglas and A. Williams, J. Chem. Soc., Perkin Trans. 2, 1974, 1727. 9 A. V. Willi, Helv. Chim. Acta, 1956, 39, 46. 10 J. A. Mastrukova, Yu. N. Sheiner, I. K. Kuznetsova, E. M. Peresleni, T. B. Sakharova and M. I. Kabachnik, Tetrahedron, 1963, 19, 357. 11 W. J. Spillane, G. Hogan and P. McGrath, J. Phys. Org. Chem., 1995, 8, 610. 12 I. Lee, H. K. Kang and H. W. Lee, J. Am. Chem. Soc., 1987, 109, 7472. 13 C. Reichardt, Solvents and Solvent Effects in Organic Chemistry, VCH, Weinheim, 2nd edn., 1988, Table A-1, pp. 408–410. 14 W. L. Matier, W. T. Comer and D. Deitchman, J. Med. Chem., 1972, 15, 538. 15 J. R. Gandler and W. P. Jencks, J. Am. Chem. Soc., 1982, 104, 1937. 16 A. G. Cook and G. W. Mason, J. Inorg. Nucl. Chem., 1966, 28, 2579. 17 P. D. Bolton and F. M. Hall, J. Chem. Soc. B, 1969, 259; P. D. Bolton and F. M. Hall, J. Chem. Soc. A, 1969, 1212. 18 G. M. Atkins and E. M. Burgess, J. Am. Chem. Soc., 1972, 94, 6135. 19 J. A. Kloek and K. L. Leschinsky, J. Org. Chem., 1976, 41, 4028. 20 G. Weiss and G. Schulze, Liebigs Ann. Chem., 1969, 729, 40. Paper 7/06410A Received 2nd September 1997 Accepted 9th October 1997

ISSN:1472-779X    

DOI:10.1039/a706410a

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1999, 15–21 15 Chiral recognition of an anionic tetrahelicene by native cyclodextrins. Enantioselectivity dominated by location of a hydrophilic group of the guest in a cyclodextrin cavity Koji Kano,*a Hisanobu Kamo,a Shigeru Negi,a Takashi Kitae,a Ryoko Takaoka,a Masahiko Yamaguchi,b Hitoshi Okubo c and Masahiro Hiramac a Department of Molecular Science and Technology, Faculty of Engineering, Doshisha University, Kyotanabe, Kyoto 610-0321, Japan b Pharmaceutical Institute, Tohoku University, Aoba, Sendai 980-0845, Japan c Department of Chemistry, Graduate School of Science, Tohoku University, Aoba, Sendai 980-0845, Japan Received (in Cambridge) 3rd August 1998, Accepted 5th November 1998 Chiral recognition of 1,12-dimethylbenzo[c]phenanthrene-5,8-dicarboxylic acid (1) by native b- (b-CD) and g-cyclodextrins (g-CD) has been studied by means of 1H NMR spectroscopy.The binding constant (K) for complexation of (M)-1 with b-CD (18700 ± 1700 dm3 mol21) is much larger than that for (P)-1 (2200 ± 100 dm3 mol21), DDG being 5.2 kJ mol21.g-CD forms less stable inclusion complexes (K = 3100 ± 100 and 690 ± 20 dm3 mol21 for (M)-1 and (P)-1, respectively). The 2D ROESY spectra indicate that both CO2 2 groups of 1 are placed near the rim of the secondary OH group side of b-CD though the (P)-1 molecule penetrates into the host cavity somewhat more deeply than the (M)-1 molecule. The deeper penetration of the (P)-1 molecule seems to be an enthalpically unfavourable but entropically favourable process because such a complexation needs dehydration from the CO2 2 group(s) of 1.The enantioselectivity of b-CD toward 1 is dominated by the diVerence in the enthalpy changes due to the diVerence in the extent of penetration between the enantiomers of 1. The 2D NMR spectra clearly indicate that at least one CO2 2 group of 1 is located inside of the g-CD cavity resulting in extensive dehydration from the guest molecule.Such an endothermic process reduces the K value for the 1-g-CD complex. The diVerence in the structures of the complexes between the guest enantiomers might be ascribed to the chiral helix-structure of the CD taken upon complexation in water. Since cyclodextrins (CD) are cyclic oligosaccharides composed of chiral glucopyranose units, a racemic guest yields diastereomeric isomers upon complexation with a host CD. If there is a diVerence in binding constants (K) between guest enantiomers, it is said that CD recognizes the chirality of the guest molecule.Lots of studies have been carried out with chiral recognition by CDs.1 However, most examples show a poor ability of CDs to discriminate between the enantiomers of guests in aqueous solutions, diVerences in DG values (DDG) for complexation between guest enantiomers being mostly less than 1 kJ mol21.2 Although molecular mechanics–molecular dynamics (MM–- MD) calculations suggest the possibility of chiral recognition of a-amino acids such as tryptophan by native a-CD,3 no data about the diVerences in binding constants (K) between the enantiomers have been reported.Improvements have been carried out by modifying native CDs. For example, protonated aminocyclodextrins are known as hosts which can recognize the central chirality of 1-phenylpropanoic acid,4 mandelic acid and its derivatives 5 and N-acetylated a-amino acids 6 in their anionic forms.In these cases, however, the DDG values for chiral recognition are still small. Metal complexes of cyclodextrins having ligand moieties show extremely large K values for complexation with native a-amino acids.7 These complexes also show small DDGs. As a general conclusion, it might be said that the CDs are poor hosts for recognition of the central chirality. Meanwhile, recent studies reveal that heptakis(2,3,6- tri-O-methyl)-b-CD (TMe-b-CD) shows an excellent ability to discriminate between the enantiomers of 1,19-bi-2-naphthol, 1,19-binaphthyl-2,29-diyl hydrogen phosphate (BNP) and 1,19- binaphthyl-2,29-dicarboxylic acid (BNC).8 In the case of BNC in a neutral form, the DDG value for chiral recognition is 4.5 kJ mol21.Such a finding is epoch-making in CD chemistry because, in general, it has been assumed that CDs are trivial hosts in chiral recognition in spite of their wide use as chiral selectors in HPLC, GLC and capillary electrophoresis (CE).9 In order to generalize the characteristic nature of CDs in chiral recognition, it is necessary to gather further examples which exhibit the high ability of CDs to discriminate between enantiomers of guests.In a previous communication,10 we reported distinct chiral recognition by b-CD of the helicity of a chiral tetrahelicene, 1,12-dimethylbenzo[c]phenanthrene-5,8-dicarboxylic acid (1). This is the first example of the recognition of helicity by CD, though it has been reported briefly that the conformation of achiral benzo[c]phenanthrene is fixed to take (P)-helicity upon complexation with g-CD.11 The DDG value for chiral recognition of 1 by b-CD is 5.2 kJ mol21 which might be the largest in chiral recognition by CDs reported so far.The present paper is a continuation of the previous communication and reports further detailed results and the mechanism for chiral recognition. H2 H3 H4 COO– H6 H7 –OOC H9 H10 H11 O O OH OH OH A B C D 1 H-1 H-2 H-3 H-4 H-5 H-6 H-6 n CD16 J.Chem. Soc., Perkin Trans. 2, 1999, 15–21 Results 1H NMR spectra Fig. 1 shows the 1H NMR spectra of (±)-1 in D2O containing various CDs at pD 7.0. Under the conditions, tetrahelicene 1 exists as a dianion. The spectrum is not aVected by a-CD, indicating no interaction between 1, having a large molecular size, and a-CD with a relatively small cavity size. Upon addition of b-CD, the signal of each proton of 1, except for H2(H11) and CH3, shifts downfield and is split.Splitting of each signal is ascribed to a larger complexation-induced shift in chemical shift (CIS) of (M)-1 as compared with that of (P)-1. Splitting of the signals of the H2(H11) and CH3 protons, which are located at a hydrophobic part of 1, is too small to be detected. Addition of g-CD causes splitting of all signals due to 1. The signals of (M)-1 shift downfield more than those of (P)-1. The signals of the H2(H11) and CH3 protons of 1 are also split by complexation with g-CD.The eVects of alkylated b-CDs such as heptakis(2,6-di-O-methyl)-b-CD (DMe-b-CD) and heptakis- (2,3,6-tri-O-methyl)-b-CD (TMe-b-CD) are not significant. In particular, TMe-b-CD does not interact with 1 at all. The cavity of TMe-b-CD is so hydrophobic that the hydrophilic dianion of 1 might not penetrate into such a host cavity. Fig. 2 shows the CISs of b- and g-CDs observed when (M)-1 and (P)-1 were added. All proton signals of the hosts shift upfield upon complexation because of the ring current eVect of 1.In the case of b-CD, the most remarkable upfield shift was observed with H-3 of b-CD, suggesting that 1 penetrates into the host cavity from the secondary OH group side of b-CD. The signals of H-1, H-2 and H-4, which are located at the outside of the host cavity, also shift upfield. A part of the 1 molecule might be situated outside of the cavity because the size of 1 is too large to be incorporated fully into the b-CD cavity.A characteristic of the g-CD system is that CISs of H-5 and H-6 are larger than that of H-3. The 1 molecule seems to be incorporated into the g-CD cavity more deeply than the case of b-CD. A continuous variation method was applied to determine the stoichiometry of the inclusion complexes of 1 and CDs. The 1 : 1 complex of each enantiomer of 1 and b- or g-CD was Fig. 1 1H NMR spectra of (±)-1 (2 × 1023 mol dm23) in D2O at pD 7.0 and 25 8C in the absence and the presence of various CDs (1 × 1022 mol dm23).suggested from the Job’s plot for the changes in the chemical shifts of H4(H9) and H6(H7) (the data are not shown herein). Binding constants and thermodynamic parameters The binding constants (K) were determined from the 1H NMR titration curves which were analyzed by a non-linear leastsquares method.12 The results are listed in Table 1. The K value for the (M)-1-b-CD complex is 18700 dm3 mol21 which is an anomalously large K for complexation of a dianionic guest with native b-CD.Meanwhile, the K value for the (P)-1-b-CD complex is relatively small (2200 dm3 mol21), but still large for a guest with two negative charges. The diVerence in free energy changes for complexation between the (M)- and (P)-enantiomers (DDG), a measure of enantioselectivity, is 5.2 kJ mol21 which might be the largest enantioselectivity in chiral recognition by CDs so far reported. From the viewpoint of size-fitting, g-CD seems to bind 1 more strongly than b-CD.However, the K value of the g-CD complex is much smaller than that of the b- CD complex against the same enantiomer of 1. Still, g-CD also shows a considerably high enantioselectivity (DDG = 3.7 kJ mol21). Table 2 summarizes the temperature dependence of K and Fig. 2 Changes in the 1H NMR chemical shifts of a) b-CD and b) g-CD (5 × 1024 mol dm23) in D2O at pD 7.0 and 25 8C upon addition of (M)- and (P)-enantiomers of 1 (1 × 1023 mol dm23), (j) (M)-1, (h) (P)-1.J.Chem. Soc., Perkin Trans. 2, 1999, 15–21 17 the thermodynamic parameters for complexation determined from the van’t HoV plots. Regardless of which enantiomer is used, the complexation of 1 with b-CD is enthalpically favourable and entropically unfavourable. As compared with the case of (M)-1, the complexation of (P)-1 with b-CD is entropically more favourable but enthalpically less favourable. The smaller K values for the 1-g-CD system are ascribed to the larger DH values compared with the 1-b-CD system. It is noteworthy that the complex formation of g-CD shows a relatively large DS.In particular, a positive but small DS was obtained for the (P)-1-g- CD system. pKa of 1 Measurement of pKa of a guest carboxylic acid in the presence of CD is one of the methods which can be used to locate the position of the CO2 2 group of the guest. If the CO2 2 group is located inside of a CD cavity, the pKa of the guest should be raised.The apparent pKa values of (M)-1 and (P)-1 in water in the presence of b- and g-CDs are shown in Table 1. Although the precipitation of 1 in the CO2H form at lower pH prevented the determination of the precise pKa of 1 in water, it was roughly estimated to be around 3. In the presence of b-CD, the apparent pKa values of the (M)- and (P)-enantiomers of 1 are 3.1 and 3.4, respectively. Successive acid–base equilibria of 1 due to two CO2H groups of 1 could not be measured from the titration curves.The pKa value of (M)-1 indicates that the CO2 2 groups of the guest are located outside of the b-CD cavity while slightly deeper penetration is suggested with (P)-1. Complexation with g-CD causes a remarkable increase in the pKa values of both enantiomers of 1. DpKa is around 1.0, Table 1 Binding constants K and enantioselectivities DDG for complexation of 1 with b- and g-CDs in D2O at pD 7.0 and 25 8C and pKa values of 1 in the presence of CDs at 25 8Ca Host b-CD b-CD g-CD g-CD Guest (M)-1 (P)-1 (M)-1 (P)-1 K/dm3 mol21 18700 ± 1700 2200 ± 100 3100 ± 100 690 ± 20 DDG/kJ mol21 5.2 3.7 pKa 3.1 3.4 4.1 4.1 a The K values were determined from the NMR titration curves measured for 1 × 1023 mol dm23 of 1 in D2O.The pKa values of 1 (1 × 1025 mol dm23) in water containing CDs (1 × 1022 mol dm23) were determined from the changes in the optical densities of 1 at 298.4 nm as a function of pH. In this case, the phosphate buVer (3.3 × 1022 mol dm23) was mostly used to adjust pH.Table 2 Binding constants K for complexation of the enantiomers of 1 with b- and g-CDs as a function of temperature and thermodynamic parameters for complexation System (M)-1-b-CD (P)-1-b-CD (M)-1-g-CD (P)-1-g-CD T/K 288.15 298.15 308.15 318.15 288.15 298.15 308.15 318.15 288.15 298.15 308.15 318.15 288.15 298.15 308.15 318.15 K/dm3 mol21 37100 18700 9750 4920 3720 2200 1400 940 4590 3244 2037 1431 818 678 543 438 DH/kJ mol21 251.1 ± 0.8 235.1 ± 1.1 230.2 ± 1.4 216.0 ± 0.3 DS/J mol21 K21 290.1 ± 3.6 253.2 ± 2.7 234.4 ± 4.7 0.45 ± 0.76 indicating that at least one CO2 2 group of 1 is located on the inside of the g-CD cavity.Structures of complexes—ROESY spectra Figs. 3 and 4 are the 2D ROESY spectra of the (M)-1- and (P)- 1-b-CD systems in D2O. The ROESY spectrum of the (M)-1-b- CD system (Fig. 3) shows correlations between the H-5 protons of the host and the CH3 and H3 protons of the guest. Such correlations clearly indicate that a hydrophobic part of the (M)-1 molecule is located inside of the CD cavity.Since a Corey–Pauling–Koltun (CPK) molecular model does not allow simultaneous penetration of both rings A and D of 1, ring A seems to penetrate into the CD cavity resulting in the protrusion of ring D. The ROESY spectrum of the (P)-1-b-CD system is more complicated. The cross peaks between the H-6 protons of the host and the CH3, H2 and H3 protons of the guest indicate a deeper penetration of ring A of (P)-1 into the b-CD cavity.The cross peaks between the H-3 protons of the host and the CH3, H2(H11) and H4(H8) protons of the guest were measured. If these correlating protons of the guest belong to ring D, it can be concluded that ring D of (P)-1 is located at the rim of the secondary OH group side and the mobility of this enantiomer in the complex is restricted more seriously than that of (M)-1. The results of the 2D NMR spectroscopy are in good agreement with those of the pKa measurements.The slightly larger pKa value of (P)-1 in the presence of b-CD can be interpreted in terms of the somewhat deeper penetration of ring A of (P)-1 compared with the case of the (M)-1-b-CD system. The ROESY spectra of the (M)-1- and (P)-1-g-CD systems are shown in Figs. 5 and 6, respectively. It is noteworthy that the H6(H7) protons of both enantiomers of 1 interact with the H-5 protons of g-CD. Such a result clearly indicates the deeper penetration of 1 to place the CO2 2 group(s) in the hydrophobic g-CD cavity.The correlations between the CH3, H2(H11) and Fig. 3 ROESY spectrum of the b-CD-(M)-1 system in D2O at pD 7.0 and 25 8C. The spectrum was measured for the solution of a mixture of b-CD (5 × 1023 mol dm23) and (M)-1 (3 × 1023 mol dm23) in N2- saturated D2O. The mixing time for the ROESY measurement was 250 ms.18 J. Chem. Soc., Perkin Trans. 2, 1999, 15–21 H6(H7) protons of the guest and the H-6 protons of the host are observed in the (P)-1-g-CD system while such correlations are negligible in the (M)-1-g-CD system.Stronger correlation with the H-6 protons in the (P)-1 complex might be explained by the deeper penetration of this guest. Fig. 4 ROESY spectrum of the b-CD-(P)-1 system. The conditions of the measurement are the same as those of the (M)-1 system. Fig. 5 ROESY spectrum of the g-CD-(M)-1 system in D2O at pD 7.0 and 25 8C. The spectrum was measured for the solution of a mixture of g-CD (1.5 × 1022 mol dm23) and (M)-1 (3 × 1023 mol dm23) in N2- saturated D2O.The mixing time for the ROESY measurement was 250 ms. Discussion The present study reveals the excellent ability of native CDs such as b- and g-CDs to discriminate between (M)- and (P)- enantiomers of tetrahelicene dicarboxylate 1. The high ability of TMe-b-CD to recognize the axial chirality of the binaphthyl derivatives has already been reported.8 The chiral recognition of 1,7-dioxaspiro[5.5]undecane by TMe-b-CD,13 the conformational enantiomerism of bilirubin induced by native CDs14 and cationic CD,15 g-CD-induced conformational enantiomerism of pamoic acid 16 and the formation of an optically active pyrene-dimer in the g-CD cavity 17 are also regarded as recognition of the chiral twisted-structures of the guest molecules.These results strongly suggest that the CDs are good hosts for recognizing chiral twisted-structures of some guests, though they are poor hosts for guests with central chirality.In the case of the chiral recognition of the binaphthyl derivatives such as BNP and BNC by TMe-b-CD, the complexation of a preferred guest [(S)-BNP or (R)-BNC] is a process accompanied by a negative enthalpy change (DH) and a positive entropy change (DS) while the complexation of another enantiomer [(R)-BNP or (S)-BNC] is an entropically unfavourable process.8c The positive entropy change in complexation of (S)- BNP in an anionic form has been ascribed to dehydration from the guest as well as the host upon complexation.The BNP monoanion penetrates into the CD cavity from the secondary OCH3 group side of TMe-b-CD to place the hydrophilic anion group (cyclic phosphate) on the inside of the hydrophobic CD cavity. Such a process is accompanied by dehydration from the phosphate anion group of BNP yielding the entropic gain. On the other hand, the phosphate anion group of (R)-BNP is located outside of the CD cavity. In this case, van der Waals interaction is the main driving force for complexation showing negative DH and negative DS.Similar entropic gain accompanied by extensive dehydration has been found in the complexation of p-methylbenzoate anion by protonated heptakis(6-amino-6-deoxy)-b-CD as well as native b-CD.18 The thermodynamic parameters are very important for discussing mechanisms for chiral recognition by CDs as well as for formation of inclusion complexes.19 Fig. 6 ROESY spectrum of the g-CD-(P)-1 system. The conditions of the measurement are the same as those of the (M)-1 system.J. Chem. Soc., Perkin Trans. 2, 1999, 15–21 19 Fig. 7 Structures of the b-CD-(M)- and (P)-1 complexes proposed from the ROESY spectra. The thermodynamic behaviour in the complexation of 1 by b-CD apparently diVers from that reported for the BNP and p-methylbenzoate anions. In spite of two CO2 2 groups of 1, the complex formation of both enantiomers of 1 shows the negative and large entropy changes.Such a result strongly suggests that the CO2 2 groups of 1 bound to b-CD are located outside of the CD cavity. The results of the 1H NMR spectroscopy as well as those of the pKa measurements indicate the structures of the 1-b-CD complexes in which two CO2 2 groups of 1 are placed outside of the CD cavity (Fig. 7). The enantioselectivity of b-CD toward 1 is controlled by the diVerence in the enthalpy changes. The complexation of (P)-1, the undesirable guest, is entropically more favourable but enthalpically less favourable than the complexation of (M)-1. Such a diVerence in the thermodynamic parameters might be interpreted in terms of the diVerence in the location of the CO2 2 groups of 1 in the inclusion complexes.As the ROESY spectra suggest, the (P)-1 molecule penetrates into the CD cavity somewhat more deeply than (M)-1 resulting in a partial penetration of the CO2 2 group of 1 into the b-CD cavity (Fig. 7). The partial penetration of the CO2 2 group needs dehydration from the (P)-1 molecule to some extent. Such a process is enthalpically unfavourable but entropically favourable. In the case of (M)-1, two CO2 2 groups of 1 are distinctively located outside of the b-CD cavity. Therefore, the complexation of (M)-1 does not need extensive dehydration from the guest. The relatively small K values for the g-CD complexes can also be explained in the same manner. The 2D NMR spectra clearly indicate the deep penetration of the molecule of 1 into the g-CD cavity.At least one CO2 2 group of 1 is located inside of the g-CD cavity. Such an inclusion requires the release of lots of water molecules from dianionic 1 yielding the enthalpic loss20 J. Chem. Soc., Perkin Trans. 2, 1999, 15–21 Fig. 8 Structures of the g-CD-(M)- and (P)-1 complexes proposed from the ROESY spectra. and the entropic gain. Indeed, both DH and DS values for the g-CD complexes are larger than those for the b-CD complexes, supporting more extensive dehydration from the guest.In particular, a small but positive DS value was obtained in the case of (P)-1. Such a positive DS value can be explained by the deep penetration of (P)-1 which causes extensive dehydration from the CO2 2 group(s) of the guest. The ROESY spectra show that (P)-1 penetrates into the g-CD cavity more deeply than (M)-1. Thus, the results of the 2D NMR spectroscopy are in good agreement with those of the thermodynamic parameters.As a consequence, it might be concluded that the enantioselectivity of native b- or g-CD toward 1 is dominated by the diVerence in the location of the hydrophilic CO2 2 group of 1 between the guest enantiomers. Penetration of the anionic group of a guest into a hydrophobic CD cavity has been interpreted in terms of an ion–dipole interaction between host and guest.20 The next problem is the reason why the location of a hydrophilic group of an enantiomer diVers from that of another enantiomer. This question might be answered by assuming a chiral twisted-structure of CD complexed with a guest in water.It is known that noncyclic dextrin takes a right-handed helix structure.21 One turn of the helix is composed of six or seven glucopyranose units. The cyclic structure of native CD is stabilized by intramolecular hydrogen bonding between the secondary OH groups of adjacent glucopyranoses.22 In water, however, the hydration of the secondary OH groups of CD weakens intramolecular hydrogen bonding.Therefore, it might be reasonable to assume that the structure of native CD as well as permethylated CD tends to take a right-handed helix structure when it complexes with a guest. Such a chiral helical structure of a host should be suitable for recognizing axial chirality or helicity of a guest.J. Chem. Soc., Perkin Trans. 2, 1999, 15–21 21 At present, we cannot explain the reason(s) for the abnormally large K values of the native CD complexes of dianionic 1 and the extremely weak interaction between 1 and TMe-b-CD.† Experimental a-, b- and TMe-b-CDs (Nacalai) were purchased and washed with THF using a Soxhlet extractor.g-CD (Nacalai) was used as received. TMe-a-CD was prepared and purified in our laboratory using an ordinary method.23 The purity of TMe-aand -b-CDs were found by 1H NMR to be satisfactory. Racemic tetrahelicene dicarboxylic acid 1 was prepared and its enantiomers were resolved by a diastereomer-method using quinine.24 The 1H NMR spectra were taken on a JEOL JNM-A400 (400 MHz) spectrometer in D2O (CEA, 99.8%) using 3-trimethylsilyl[ 2,2,3,3-2H4]propionate (TSP, Aldrich) as an external standard.ROESY spectra were recorded with a spectral width of 3358 Hz. The 908 pulse was 33.3 ms, mixing time was 250 ms, delay time was 2.0 s and 512 × 512 data points were recorded. The pD values were adjusted by using Na2CO3 and DCl.The MM–MD calculations involving the eVects of solvent were carried out as previously reported.25 Acknowledgements This work was supported by a Grant-in-Aid for Scientific Research on Priority Areas, ‘New Polymers and Their Nano- Organised Systems’ (No. 277/09232258), a Grant-in-Aid for Scientific Research B (10440211) and a subsidy to RCAST of Doshisha University from the Ministry of Education, Science, Sports and Culture, Japan. References 1 For reviews, see (a) K.Kano, in Bioorganic Chemistry Frontiers, ed. H. Dugas and F. P. Schmidtchen, Springer-Verlag, Berlin, Heidelberg, 1993, vol. 3, ch. 1; (b) K. Kano, J. Phys. Org. Chem., 1997, 10, 286. 2 (a) A. Cooper and D. D. MacNicol, J. Chem. Soc., Perkin Trans. 2, 1978, 760; (b) R. Fornasier, P. Scrimin and U. Tonellato, Tetrahedron Lett., 1983, 24, 5541; (c) I. Tabushi, Y. Kuroda and T. Mizutani, J. Am. Chem. Soc., 1986, 108, 4514; (d) Y. Ihara, E. Nakanishi, M. Nango and J. Koga, Bull.Chem. Soc. Jpn., 1986, 59, 1901; (e) S. E. Brown, J. H. Coates, S. F. Lincoln, D. R. Coghlan and C. J. Easton, J. Chem. Soc., Faraday Trans., 1991, 87, 2699; ( f ) S. Li and W. C. Purdy, Anal. Chem., 1992, 64, 1405; ( g) B. F. Feibush, C. L. Woolley and V. Mani, Anal. Chem., 1993, 65, 1130. 3 K. B. Lipkowitz, S. Raghothama and J.-A. Yang, J. Am. Chem. Soc., 1992, 114, 1554. 4 S. E. Brown, J. H. Coates, P. A. Duckworth, S. F. Lincoln, C. J. Easton and B. L. May, J. Chem.Soc., Faraday Trans., 1993, 89, 1035. † The possibility of hydrogen-bond formation between the CO2 2 groups of 1 and the OH groups of b-CD could be studied to explain the large K values. The cavity of TMe-b-CD might be too hydrophobic to include anionic 1. 5 T. Kitae, H. Takashima and K. Kano, J. Inclusion Phenom. Mol. Recognit. Chem., in the press. 6 T. Kitae, T. Nakayama and K. Kano, J. Chem. Soc., Perkin Trans. 2, 1998, 207. 7 (a) R. Corradini, A. Dossena, G. Impellizzeri, G.Maccarrone, R. Marchelli, E. Rizzarelli, G. Sartor and G. Vecchio, J. Am. Chem. Soc., 1994, 116, 10267; (b) S. E. Brown, C. A. Haskard, C. J. Easton and S. F. Lincoln, J. Chem. Soc., Faraday Trans., 1995, 91, 1013. 8 (a) K. Kano, K. Yoshiyasu and S. Hashimoto, J. Chem. Soc., Chem. Commun., 1989, 1278; (b) K. Kano, Y. Tamiya, C. Otsuki, T. Shimomura, T. Ohno, O. Hayashida and Y. Murakami, Supramol. Chem., 1993, 2, 137; (c) K. Kano, Y. Kato and M. Kodera, J. Chem. Soc., Perkin Trans. 2, 1996, 1211. 9 (a) V. Schurig and H.-P. Nowotny, Angew. Chem., Int. Ed. Engl., 1990, 29, 939; (b) S. Li and W. C. Purdy, Chem. Rev., 1992, 92, 1457; (c) S. F. Y. Li, J. Chromatogr. Libr., 1992, 52, 201; (d ) R. Kuhn and S. HoVstetter-Kuhn, Chromatographia, 1992, 34, 505; (e) S. Fanali, J. Chromatogr. A, 1996, 735, 77. 10 K. Kano, S. Negi, H. Kamo, T. Kitae, M. Yamaguchi, H. Okubo and M. Hirama, Chem. Lett., 1998, 151. 11 G. Le Bas, C. de Rango, N. Rysanek and G. Tsoucaris, J.Inclusion Phenom. Mol. Recognit. Chem., 1984, 2, 861. 12 H.-J. Schneider, R. Kramer, S. Simova and U. Schneider, J. Am. Chem. Soc., 1988, 110, 6442. 13 K. Yannakopoulou, D. Mentzafos, I. M. Mavridis and K. Dandika, Angew. Chem., Int. Ed. Engl., 1996, 35, 2480. 14 (a) D. A. Lightner, J. K. Gawrinski and K. Gawronska, J. Am. Chem. Soc., 1985, 107, 2456; (b) K. Kano, K. Yoshiyasu and S. Hashimoto, J. Chem. Soc., Chem. Commun., 1988, 801; (c) K. Kano, K. Yoshiyasu, H. Yasuoka, S. Hata and S. Hashimoto, J. Chem. Soc., Perkin Trans. 2, 1992, 1265. 15 K. Kano, S. Arimoto and T. Ishimura, J. Chem. Soc., Perkin Trans. 2, 1995, 1661. 16 K. Kano, M. Tatsumi and S. Hashimoto, J. Org. Chem., 1991, 56, 6579. 17 (a) K. Kano, K. Matsumoto, S. Hashimoto, M. Sisido and Y. Imanishi, J. Am. Chem. Soc., 1985, 107, 6117; (b) K. Kano, H. Matsumoto, Y. Yoshiyasu and S. Hashimoto, J. Am. Chem. Soc., 1988, 110, 204. 18 K. Kano, T. Kitae, H. Takashima and Y. Shimofuri, Chem. Lett., 1997, 899. 19 For example: (a) Y. Inoue, T. Hakushi, Y. Liu, L.-H. Tong, B.-J. Shen and D.-S. Jin, J. Am. Chem. Soc., 1993, 115, 475; (b) Y. Liu, B.-H. Han, B. Li, Y.-M. Zhang, P. Zhao, Y.-T. Chen, T. Wada and Y. Inoue, J. Org. Chem., 1998, 63, 1444. 20 K. Kano, N. Tanaka, H. Minamizono and Y. Kawakita, Chem. Lett., 1996, 925. 21 R. E. Rundle and F. C. Edwards, J. Am. Chem. Soc., 1943, 65, 2200. 22 (a) W. Saenger, Angew. Chem., Int. Ed. Engl., 1980, 19, 344; (b) K. K. Chacko and W. Saenger, J. Am. Chem. Soc., 1981, 103, 1708; (c) T. Steiner, S. A. Mason and W. Saenger, J. Am. Chem. Soc., 1990, 112, 6184; (d ) K. Harata, Chem. Lett., 1984, 641. 23 S. Hakomori, J. Biochem. (Tokyo), 1964, 55, 205. 24 M. Yamaguchi, H. Okubo and M. Hirama, Chem. Commun., 1996, 1771. 25 K. Kano, K. Minami, K. Horiguchi, T. Ishimura and M. Kodera, J. Chromatogr. A, 1995, 694, 307. Paper 8/06054A

ISSN:1472-779X    

DOI:10.1039/a806054a

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1998 19 Enantiomer discrimination using lipophilic cyclodextrins studied by electrode response, pulsed-gradient spin-echo (PGSE) NMR and relaxation rate measurements Ayelet Gafni,a Yoram Cohen,*,a Ritu Kataky,b Simon Palmer b and David Parker *,b a School of Chemistry, Sackler Faculty of Exact Sciences, Tel-Aviv University, Ramat Aviv, Tel Aviv 69978, Israel b Department of Chemistry, University of Durham, South Road, Durham, UK DH1 3LE The diastereoisomeric complexes formed between 2,6-di-O-alkyl-·- and ‚-cyclodextrins and the arylammonium ions propranolol, ephedrine and amphetamine have been studied by electrode response and NMR methods.Enantioselectivity in binding propranolol is 3.3 :1 with 2,6-di-O-dodecyl-‚- cyclodextrin in favour of the (1)-enantiomer as revealed by measurement of the association constant using pulsed-gradient spin-echo (PGSE) NMR methods. In all of the cases of enantiodifferentiation studied here, the (1)-enantiomer is more strongly bound by the cyclodextrin.Relaxation rate measurements of the host and guest proton NMR resonances highlight the importance of hydrogenbonding in enantiomer discrimination. The behaviour of selectively alkylated cyclodextrin derivatives as ionophores in ion-selective electrodes (ISEs) has been the focus of considerable activity in the past few years. Not only do these receptors act as chiral ionophores in potentiometric devices for the enantiodifferentiation of size-matched arylammonium ions1,2 but also they serve as size-selective receptors for complementary ’onium ions.3,4 Recent examples include the development of voltammetric sensors for tricyclic antidepressants such as imipramine,5 and the definition of a robust biosensor for acetyl choline.6 Alkylated and acylated cyclodextrin derivatives have also been studied in detail as chiral stationary phases in enantiomer analysis by HPLC7,8 and GC methods.9,10 The origins of enantiodifferentiation in such systems have focused the attention of many groups,7a,2b,11–14 but generally work has been undertaken on comparisons of elution orders and separation factors with computations of the relative stability of the diastereoisomeric complexes, for example using Monte Carlo and molecular dynamics simulations.14 In an attempt to understand better the origins of chiral discrimination in the solution phase, we have carried out some further experiments correlating electrode response studies—which define the most strongly bound enantiomer and give an estimate of the free energy difference between diastereoisomeric complexes in the membrane used—with NMR measurements.Association constants for selected cyclodextrin–arylammonium ion complexes have been determined using NMR diffusion measurements and information on the degree of hydrogenbonding gleaned from 1H NMR shift and relaxation rate determinations. Results and discussion Pulsed-gradient spin-echo measurements of complex stability Measurements of diffusion coefficients using the pulsedgradient spin-echo (PGSE) technique 15 have proved to be very useful in studying the degree of association of complementary organic host–guest systems.16,17 This NMR method is particularly useful when the host has a much bigger molecular volume than the guest. In this situation, there is a large difference in their diffusion coefficients: when conditions of slow-exchange on the NMR timescale prevail, the bound guest will possess a diffusion coefficient that matches that of the host.Under fast exchange, the guest’s measured diffusion coefficient will be a weighted average of free and bound species and thereby reports on the equilibrium constant for complex formation. Association constants for 1 : 1 complex formation † between the trifluoroacetate salts of the enantiomers of amphetamine 4a, ephedrine 5 and propranolol 6 with 2,6-di-O-dodecyl-acyclodextrin 1, and the b analogue 3a have been measured.A representative data set is shown in Fig. 1 showing the decay of the normalised echo intensity as a function of the square of the pulsed gradient strength for free ephedrine (5) and for the ephedrine and 2,6-di-O-dodecyl-a-cyclodextrin (1) in a 1 : 1 O OH NH2Pri O H5 H3 R¢O O OR OR NH2R Me + (R)-(+)-6 NH2Me OH Me n 1 R = C12H25, R¢ = H, n = 6 2a R = C8H17, R¢ = H or C8H17, n = 6 b R = C8H17, R¢ = C8H17 or Me 3a R = C12H25, R¢ = H, n = 7 b R = C8H17, R¢ = C8H17 or H + + (S)-(+)- 4a R = H b R = Me (1S,2R)-(+)-5 † Confirmation of a 1 : 1 binding stoichiometry was provided by electrospray mass spectrometry where singly charged adducts were detected at the appropriate mass in each case with observed and calculated isotope patterns showing good agreement.20 J.Chem. Soc., Perkin Trans. 2, 1998 Table 1 Diffusion coefficients (D/cm2 s21) of the chiral ammonium salts a and of the cyclodextrin systems studied in the free state and in 1 : 1 solutions along with the association constants (Ka/dm3 mol21) derived from these data b–d System 1 — — (S)-(2)-5 — (R)-(1)-5 1 (R)-(1)-5 1 (S)-(2)-5 2b — 2b (S)-(2)-5 2b (R)-(1)-5 3a — — (S)-(1)-4a — (R)-(2)-4a 3a (S)-(1)-4a 3a (R)-(2)-4a — (S)-(2)-6 — (R)-(1)-6 3a (S)-(2)-6 3a (R)-(1)-6 Cyclodextrin/ 1025 cm2 s21 0.32 ± 0.01 — — 0.28 ± 0.01 0.28 ± 0.01 0.39 ± 0.01 0.31 ± 0.01 0.31 ± 0.01 0.33 ± 0.01 — — 0.31 ± 0.01 0.31 ± 0.01 — — 0.30 ± 0.01 0.24 ± 0.01 Ammonium salts/1025 cm2 s21 — 0.66 ± 0.02 0.66 ± 0.02 0.51 ± 0.01 0.54 ± 0.01 — 0.52 ± 0.01 0.52 ± 0.01 — 0.69 ± 0.01 0.69 ± 0.02 0.48 ± 0.01 0.51 ± 0.01 0.59 ± 0.02 0.59 ± 0.02 0.53 ± 0.02 0.45 ± 0.02 Ka/dm3 mol21 — — — 142 ± 21 114 ± 15 113 ± 19 113 ± 19 — — — 163 ± 28 102 ± 16 — — 67 ± 17 222 ± 58 K(1)/K(2) — — — 1.25 ± 0.25 — 1.00 ± 0.24 — — — 1.60 ± 0.37 — — 3.31 ± 1.2 a As a trifluoroacetate salt.b All experiments were performed on 5 mmol CDCl3 solutions at 283 K using a 500 MHz NMR spectrometer as described previously, see refs. 16 and 17. c Diffusion coefficients are the mean of at least three experiments and the reported values are means ± SD. d 1 = 2,6- didodecyl-a-cyclodextrin; 2b = 2,6-didodecyl-3-O-methyl-a-cyclodextrin; 3a = 2,6-didodecyl-b-cyclodextrin; 5 = ephedrine; 4a = amphetamine; 6 = propranolol. solution of 5–1. From this data set the diffusion coefficient may be calculated according to eqn.(1) 15 where Ag and A0 are the lnSAg A0 D= 2(ggd2)2(D 2 d/3)D (1) echo intensities in the presence and in the absence of the pulsed gradients, respectively, g is the gyromagnetic ratio (rad s21 g21), g is the gradient strength (G cm21), D is the self diffusion coef- ficient of the observed spins (cm2 s21), d is the length of the diffusion gradient and D is the time-separation between the edges of the diffusion gradients. The most strongly bound complex was formed between 3a and (R)-(1)-propranolol (Ka = 222 ± 58 dm3 mol21, 283 K, CDCl3), and the Ka value is of similar magnitude to the reported affinities of b-cyclodextrin Fig. 1 Normalised 1H NMR signal attenuation (ln Ag /A0) of free ephedrine (5) (j), of bound ephedrine in a 1 : 1 complex with 2,6- didodecyl-a-cyclodextrin (1) (m) and of the cyclodextrin in the complex (d) as a function of the square of the pulse gradient strength (G2). The diffusion coefficients were calculated from the slopes of such graphs using eqn. 1.for 1-substituted naphthyl derivatives.18 The (S)-enantiomer was bound more weakly (Ka = 67 ± 15 dm3 mol21, 283 K, CDCl3) corresponding to a free energy difference in binding of 2.8 (±1.0) kJ mol21. Both the sense and the magnitude of the measured enantiodifferentiation are in line with values estimated from chiral HPLC methods.7a In the other two cases studied (Table 1), more modest selectivity was observed, but in each case it was the (1)-enantiomer which was the more strongly bound.Using the cyclodextrin host 2b—in which all residual OH groups have been alkylated—no difference was found in the stability of the diastereoisomeric complexes with (R) and (S)-ephedrine. Alkylation of the 3-OH position in cyclodextrins removes the intramolecular H-bonding network, ([3]OH? ? ?O[2]), that determines the conformational rigidity of the host molecule. It had previously been noted, in electrode response studies, that complete alkylation of the 3-OH group in a-cyclodextrin derivatives removed the enantioselectivity in binding ephedrine and its stereoisomers.1,2b Electrode response studies The behaviour of selected lipophilic cyclodextrin derivatives as ionophores in a standard plasticised PVC membrane electrode1,2 was assessed.The response of an electrode incorporating 2,6-di-O-dodecyl-b-cyclodextrin towards (R)-(1)-propranolol and its enantiomer was compared in the absence and presence of potential interferent ions (Table 2).Differences in the cell electrode potential (Einitial) were noted and give a measure of the difference in free energy of complexation of the propranolol guest by the immobilised cyclodextrin ionophore. In the presence of Na1, K1 or in a simulated clinical background, the difference in measured electrode potential (ca. 30 mV) for (1) and (2)-propranolol corresponded to a free energy difference of 2.9 (±0.4) kJ mol21 which is similar to the value measured by the PGSE method in chloroform solution.The most strongly bound enantiomer (giving rise to the largest Einitial value) was the (R)-(1) isomer, also in agreement with the sense observed in solution by NMR. Similar correlations were noted in the response of electrodes based on 2a, 2b, 3a and 3b towards the protonated salts of ephedrine and amphetamine, (Table 3). In each case the (1)-enantiomer was the more strongly bound (Einitial value) and the magnitude of the enantioselectivity was dependent upon the nature of the cyclodextrin used.ForJ. Chem. Soc., Perkin Trans. 2, 1998 21 Table 2 Response of ISEs incorporating 2,6-di-O-dodecyl b-cyclodextrin to R-(1)-propranolol [(S) enantiomer values are in parentheses] in the absence and presence of interfering ions at 310 K. Selectivity coefficients for interferent ions are given as 2 log Ki, j pot; for individual ions the interferent concentration is 0.1 mol dm23 Interferent Calibration Clinical backgrounda Na1 K1 Ca21 Einitial/mVb 251 (239) 301 (270) 258 (239) 293 (261) 327 (264) Limit of detection/ 1025 mol dm23 2.9 (1.1) 1.4 (0.8) 0.9 (0.7) 1.7 (1.6) 3.3 (2.3) Slope/mV 58 (57) 56 (61.5) 56 (57) 61 (62) 33 (31) 2 log Ki, j pot — 4.0 (4.3) 4.0 (4.2) 3.8 (3.8) 3.5 (3.6) a Clinical background is a simulated background of clinical ions as chloride salts (c/mmol dm23: Na1 145; K1 4.3; Ca21 1.26; Mg21 0.9). b Relative to an external Ag/AgCl double junction reference electrode connected by a saturated aqueous KCl salt bridge.Table 3 Response of ISEs incorporating 2,6-di-O-dodecyl-b-cyclodextrin 3a or the poly-O-octyl-a-cyclodextrins 2a and 2b to chiral amine salts at 310 K Enantioselectivity Slope/mV Entry 1 2 3 4 5 6 Analyte ephedrine ephedrinea ephedrineb amphetamine ephedrineb amphetaminea Ionophore 1 2a 2b 3a 3b 2a DE/mV 9(3) 26(2) 3 39 8 (88.5) (1) 55 60 43 61.5 56 50 (2) 54 50 40 62.5 51 37 a Similar behaviour was deserved with ionophore 1; the electrode response with (R)-(2) amphetamine was unstable, so the quoted DE value is unreliable.Unstable electrode responses were also obtained with methamphetamine and 1 or 2a. b Data from ref. 2b. example with (1)-ephedrine as the analyte, the DE value was 26 mV with ‘poly’-O-octyl-a-cyclodextrin, 2a (containing 15.4 octyl groups and 2.6 OH groups, on average 2a ), but reduced to 9 (±3) mV with 2,6-di-O-dodecyl-a-cyclodextrin as the sensing ionophore (Table 3).In addition when the residual 3-OH groups in 2a were capped by methyls, there was little or no enantioselectivity in binding ephedrine (entry 3 compared to 2, Table 3), and use of a larger b-cyclodextrin ionophore (entry 5) also diminished the measured enantioselectivity. With amphetamine and methamphetamine as analytes, super- Nernstian or unstable (in time) electrode responses were observed when using any of the available a-cyclodextrin ionophores (e.g. entry 6). Only with amphetamine and 2,6-di-Ododecyl- b-cyclodextrin was near Nernstian behaviour observed and in this case, while the sense of the enantioselective response was in line with the NMR measurements, the size of the discrimination was larger (3.8 kJ mol21 in favour of (S)-(1)- amphetamine, cf. 0.9 kJ mol21 from the PGSE data).Solution NMR Studies of Hydrogen Bonding (a) Chemical shift effects. Complex formation was monitored by 1H NMR spectroscopy in CDCl3 solution, using the trifluoroacetate salts of the chiral b-arylammonium ions in the presence (and absence) of 1 and 3a.Earlier studies had reported the behaviour of the complexes of ephedrine with 2a,2b and IR and NMR experiments had identified a strong intramolecular NH? ? ? OH interaction in the free and bound state. The diastereotopic NH2 hydrogens in ephedrinium trifluoroacetate are highly anisochronous in CDCl3 (Dd = 1.03 ppm) consistent with the preferential population of a single conformer in which the two hydrogens are in distinct local magnetic environments.Upon addition of 2a to (1)-ephedrine (in molar ratio 1 : 2.5 at 293 K) there was an increase in the chemical shift nonequivalence of the NH protons to 1.17 ppm and the higher frequency NH (resonating at 9.42 ppm in the ‘free’ state) shifted to higher frequency in the bound form (dNHa = 9.53 ppm; NHa is denoted as resonating to higher frequency of NHb). A similar pattern of behaviour was observed with propranolol both before and after addition of 3.In the free state, the shift nonequivalence of the NH protons was 1.07 ppm, and the presence of an intramolecular hydrogen bond was also suggested by the presence of a weak band at 3590 cm21 in the solution FTIR spectrum (293 K, 1021 mol dm23) whose position was independent of concentration upon dilution by a factor of fifty. In the presence of 3a ([propranolol] = 33 mmol dm23, [3a] = 13 mmol dm23, CDCl3, 293 K), the NH2 shift non-equivalence was 1.04 ± 0.01) ppm for both enantiomers, but in the case of (R)- (1)-propranolol, the complexation shifts were 0.08 and 0.09 ppm to higher frequency whereas with (S)-(2)-propranolol Fig. 2 Changes in the 1H NMR spectra of protonated propranolol in the presence of 3a: (a) (S)-propranolol; (b) no cyclodextrin; (c) (R)- propranolol; 293 K, CDCl3, [propranolol] = 33 mM, [3a] = 13 mM22 J. Chem. Soc., Perkin Trans. 2, 1998 Table 4 Comparative relaxation rates (s21) for chiral arylammonium ions in the presence of 2,6-di-O-dodecyl-a- and b-cyclodextrin (293 K, CDCl3, 500 MHz) Chiral complex 3a–4a 3a–6 1–5 1–4b 1–4a Observed proton a NH3 H3 NHa NHb NHa NHb H3 NHa NHb H3 NH3 Free (1) 12.7 0.64 8.6 8.8 7.9 7.4 0.60 11.9 11.6 0.61 11.2 Bound (1) 3.4 0.78 9.6 8.6 3.5 4.5 0.75 6.6 6.6 0.78 1.9 DR1 29.3 10.14 11.0 20.2 24.4 22.9 10.15 25.3 25.0 10.17 29.3 Free (2) 12.7 0.64 8.6 8.8 7.9 7.4 0.60 11.9 11.6 0.61 11.2 Bound (2) 4.7 0.84 5.5 5.4 4.6 5.3 0.80 5.3 5.4 0.76 3.5 DR1 28.0 10.2 23.1 23.4 23.3 22.1 10.2 26.6 26.2 10.15 27.7 DDR1 1.3 0.06 4.1 3.2 1.1 0.8 0.05 1.3 1.2 0.02 1.6 a In each case, Ha resonates to higher frequency of Hb.Measurements were made on degassed samples in 5 nm tubes sealed under argon. The proton H3 is the cyclodextrin hydrogen which is directed into the cavity. shifts of 0.01 and 0.05 ppm to lower frequency were observed for the NHa and NHb protons (Fig. 2). Comparative behaviour of this type is consistent with the formation of a more welldefined hydrogen-bonded structure in the complex for the (1)- enantiomer, with respect to that formed with the (2)-isomer.Although it is not possible to infer any definitive conclusions in the case of methamphetamine, (S)-(1)-4b, in the presence of 1, (again at a ratio of 1 : 2.5, CDCl3, 293 K) the shift difference on inclusion was 0.38 and 0.32 ppm to higher frequency for NHa and NHb and with the (2)-enantiomer the corresponding values were 10.28 and 10.32 ppm respectively.(b) Relaxation rate measurements. Measurements of the proton relaxation rates (R1 = T1 21) using standard inversion recovery methods were carried out on the complexes of 1 with 5, 4a or 4b and on the complexes of 3a with 4a or 6 under identical conditions in an effort to define further structural differences in the diastereoisomeric complexes. It is well known that measured relaxation rates are particularly sensitive to many factors including solvent, temperature, relative conformational population, exchange dynamics and intermolecular distances.In this short study the intention was to compare R1 data for the diastereoisomeric complexes under controlled conditions, seeking out differences in behaviour and paying attention to the NH protons (Table 4). The relaxation rates of both NH2 protons in the more weakly bound (S)-(2)-propranolol (6), guest decreased markedly upon complexation. With the (R)-(1)-enantiomer the R1 value increased by 1.0 s21 for NHa (the one resonating to higher frequency) while NHb remained more or less unchanged.For the complexes of ephedrine 5 and the amphetamines 4a and 4b, there was a large drop in the NH R1 values but in none of these cases was the difference in the change in R1 (DDR1) as marked as was found with propranolol. In all of these cases complex formation may be associated with the suppression of at least one dipolar relaxation pathway whose effect is to decrease the local reorientational correlation time through an increase in local motional mobility.Such a process may be tentatively linked to NH ? ? ? O hydrogen-bond formation in the complex. The large change may be related to the change in hydrogenbonding state of the N+H3 and NH2 1 protons associated with the presence of residual water (from the solvent or the salt): complex formation may be linked to a decrease in the extent of hydrogen-bonding to these water molecules, as hydrogen bonding to the cyclodextrin host becomes competitive.Counteracting this tendency is the general increase in R1 values expected when a small guest associates with a large molecule. A general increase in the guest R1 values is expected as a consequence of the decrease in molecular motion, w, associated with a more slowly tumbling molecule. Given that these two factors have an opposing effect on the measured R1 value, the overall differential effect will be a function of the difference in equilibrium constants for formation of the diastereoisomeric complexes and the change in the degree of NH ? ? ? O hydrogen bonding in these complexes.Conclusions Previous work with ephedrine and its congeners as guest, had established that it was the configuration a to the amino group which determined the sense of enantiodifferentiation in cyclodextrin inclusion complexes.2b The (2R)-(1)-enantiomer was the more strongly bound and in the (2S)-(2)-complex it was proposed that there was an unfavourable steric interaction between the 2-Me group and the H3 proton of the cyclodextrin host that inhibited a favourable NH ? ? ? O hydrogen bonding interaction.In this work, the NMR measurements of complex formation and the electrode response studies—albeit on a limited set of complexes—all show that the (1)-enantiomer is the more strongly bound. This accords with an increased degree of hydrogen-bonding (involving NHa) observed selectively upon cyclodextrin complexation for the (1)-enantiomer.The case of the complexation of propranolol is the most clear-cut: the (R)-(1)-enantiomer is the more strongly bound (by ca. 2.6– 2.9 kJ mol21) and in its complex with 2,6-di-O-dodecyl-bcyclodextrin there is good evidence for a stabilising N+H? ? ?O hydrogen-bond that is absent in the weaker isomeric complex. Such energy differences are not out of line with hydrogenbond matches and mismatches observed in other hydrogenbonded arrays, where a single H-bonded interaction has been highlighted.19 Notwithstanding the fact that this interpretation is in accord with the early model proposed for the interaction of the parent b-cyclodextrin with propranolol 7a— wherein the (1)-enantiomer forms a stabilising NH ? ? ?O contact with the b-cyclodextrin host that is absent in the isomeric complex—this work lends support to the case for detailed solution experimental studies of complex formation, rather than just relying on conclusions from modelling studies.13,14 Experimental Potentiometric studies: membrane preparation The electroactive membranes were prepared containing 1.2% ionophore, 65.6% plasticizer (2-nitrophenyloctyl ether), 32.8% PVC (high molecular weight), and 0.4% lipophilic anion {sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate} in distilled tetrahydrofuran.The membranes were cast by a controlled evaporation method reported previously.2b The polymer membranes were mounted in Philips IS(561) electrode bodies (Philips Analytical, Eindhoven, The Netherlands) with an inner filling solution of 1023 mol dm23 ammonium chloride.TheJ. Chem. Soc., Perkin Trans. 2, 1998 23 electrodes were conditioned for 24 h in 1023 mol dm23 analyte solution. Constant volume dilution measurements The ion-selective electrode was held in a small volume (approximately 2.3 cm3), thermostatted double walled glass cell with inlet and outlet capillaries and a miniature magnetic follower.The reference cell employed was a T-shaped thermostatted liquid junction configuration in which the analyte solution flowed over a capillary containing a saturated KCl salt bridge solution in contact with a saturated calomel reference electrode (Russell pH Ltd.). The solution was drawn through the system by a peristaltic pump (Gilson Minipuls 3). All measurements were made at 310 K, unless otherwise stated. The ISE and reference electrode were connected to a digital multimeter (Keithley 197) and a chart recorder (Kipp & Zonen) via a buffer amplifier.The system was thermostatted using a Techne tempette junior TE-85 thermostat bath. Solution infrared measurements The solution infrared measurements on propranolol trifluoroacetate were recorded in the range 0.1–0.001 mol dm23 (in CCl4) on a Perkin-Elmer 1600 Series FTIR. NMR Measurements All the NMR measurements were performed on a Bruker AMX 500 instrument.Measurements of 1H relaxation times (T1) were made with degassed CDCl3 samples containing 13.2 mmol dm23 host (alkylated cyclodextrin) and 33 mmol dm23 guest (as the trifluoroacetate salt). The changes in the longitudinal relaxation times were measured using standard inversion–recovery methods. NMR diffusion experiments were performed, at 283 K, on a Bruker ARX500 spectrometer equipped with a BGU pulsed gradient unit on a B-VT-2000 temperature control unit. Data were collected using a commercial 5 mm inverse probe equipped with shelf-shielded g-gradients on 5 mmol dm23 samples in CDCl3.Reagents (R)-(1) and (S)-(2)-propranolol hydrochloride, (R)-(1) and (S)-(2)-ephedrine hydrochloride, (R)-(2)-amphetamine sulphate and (S)-(1)-amphetamine were all obtained from Sigma. Analar KCl along with ortho-nitrophenyloctyl ether (ONPOE) and polyvinylchloride (PVC, high molecular weight) were obtained from Fluka-Microselect and resublimed NH4Cl from Fluka.Sodium tetrakis[3,5-bis(trifluoromethyl)phenyl]borate (TKB) was synthesised in this laboratory. All standard solutions were prepared using de-ionised water (Milli Q, Milipore- Waters, Milford, MA, USA). Acknowledgements We thank Durham University, EPSRC and the Israel Academy of Sciences and Humanities for support. References 1 P. S. Bates, R. Kataky and D. Parker, Analyst, 1992, 117, 1313. 2 (a) P. S. Bates, A. F. Patti and D. Parker, J. Chem. Soc., Perkin Trans. 2, 1994, 657; (b) P.S. Bates, R. Kataky and D. Parker, J. Chem. Soc., Perkin Trans. 2, 1994, 669. 3 P. M. Kelly, R. Kataky, D. Parker and A. F. Patti, J. Chem. Soc., Perkin Trans. 2, 1995, 1955. 4 R. Kataky, P. S. Bates and D. Parker, Analyst, 1994, 119, 181. 5 R. Kataky, S. Palmer and D. Parker, Electroanalysis, 1997, in the press. 6 R. Kataky and D. Parker, Analyst, 1996, 121, 1829; R. Kataky and D. Parker, Chem. Commun., 1997, 141. 7 (a) D. W. Armstrong, T. J. Ward, R. D. Armstrong and T.E. Beesley, Science, 1986, 232, 1132; (b) D. W. Armstrong, T. J. Ward, A. Czech, B. P. Czech and R. A. Bartsch, J. Org. Chem., 1985, 50, 5556; (c) D. W. Armstrong, S. Chen, C. Chang and S. Chang, J. Liq. Chromatogr., 1992, 15, 545. 8 Chiral Separations by Liquid Chromatography, ed. G. Subramaniam, VCH, Weinheim, 1994. 9 V. Schurig and H.-P. Nowotny, Angew. Chem., Int. Ed. Engl., 1990, 29, 939. 10 W. A. Konig, B. Gehrcke, M. G. Peter and G. D. Prestwich, Tetrahedron: Asymmetry, 1993, 4, 165. 11 K. B. Lipkowitz, S. Ragholhama and J. Young, J. Am. Chem. Soc., 1992, 114, 1554; K. B. Lipkowitz, J. Chromatogr. A, 1995, 694, 15. 12 C. J. Easton and S. F. Lincoln, Chem. Soc. Rev., 1996, 24, 163. 13 K. B. Lipkowitz, G. Pearl, B. Coner and M. A. Peterson, J. Am. Chem. Soc., 1997, 119, 600. 14 Examples of molecular dynamics simulations include: J. E. Kohler, M. Hohla, M. Richters and W. A. Kaing, Chem. Ber., 1994, 127, 119; A. V. Eliseev, G. A. Iacobucci, N. A. Khanjin and F. M. Menger, J. Chem. Soc., Chem. Commun., 1994, 2051; M. E. Amato, G. M. Lombardo, G. C. Pappalardo and G. Scarlata, J. Mol. Struct., 1995, 350, 71; see also: D. R. Black, C. G. Parker, S. S. Zimmerman and M. L. Lee, J. Comput. Chem., 1996, 17, 931. 15 (a) E. O Stejskal and J. E. Tanner, J. Chem. Phys., 1965, 42, 285; (b) P. Stilbs, Progr. Nucl. Magn. Reson. Spectrosc., 1987, 19, 1; (c) O. Soderman and P. Stilbs, Progr. Nucl. Magn. Reson. Spectrosc., 1994, 26, 480. 16 (a) O. Mayzel and Y. Cohen, J. Chem. Soc., Chem. Commun., 1994, 1901; (b) O. Mayzel, O. Aleksuike, F. Gnynszfan, S. E. Biali and Y. Cohen, J. Chem. Soc., Chem. Commun., 1995, 1183. 17 A. Gafni and Y. Cohen, J. Org. Chem., 1997, 62, 120. 18 Cyclodextrins, ed. J. Szeitli and T. Osa, in Comprehensive Supramolecular Chemistry, ed. J. L. Atwood, J. E. D. MacNicol and F. Vogtle, vol. 3, Pergamon, Oxford, 1996. 19 G. C. Best and P. B. Dervan, J. Am. Chem. Soc., 1995, 117, 1187. Paper 7/05921C Received 13th August 1997 Accepted 17th September 1997

ISSN:1472-779X    

DOI:10.1039/a705921c

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1999, 23–32 23 A thermodynamic, electrochemical and molecular dynamics study on NAD and NADP recognition by 1,4,7,10,13,16,19-heptaazacyclohenicosane ([21]aneN7)† Antonio Doménech,a Enrique García-España,*a José A. Ramírez,a Bernardo Celda,*b Ma Carmen Martínez,b Daniel Monleón,b Roberto Tejero,b Andrea Bencini c and Antonio Bianchi *c a Department of Inorganic Chemistry, University of Valencia, C/ Dr. Moliner 50, 46100 Burjassot (Valencia), Spain b Department of Physical Chemistry, University of Valencia, C/ Dr.Moliner 50, 46100 Burjassot (Valencia), Spain c Department of Chemistry, University of Florence, Via Maragliano 75/77, 50144 Florence, Italy Received (in Cambridge) 30th July 1998, Accepted 30th October 1998 Interaction of the macrocyclic polyamine 1,4,7,10,13,16,19-heptaazacyclohenicosane ([21]aneN7) in its protonated forms with the dinucleotides NAD1 and NADP1 has been followed by pH-metric titration, NMR and cyclic voltammetry.Both dinucleotides interact strongly with [21]aneN7 forming adduct species with protonation degrees varying from 3 to 7 for NAD1 and from 3 to 8 for NADP1. Plots of the overall amount of complexed species show recognition of NADP1 over NAD1 throughout a wide pH range. The extra phosphate group attached to the ribose moiety of NADP1 seems to be the factor controlling the recognition process as proved by 31P NMR studies. Molecular dynamics on the adducts formed between NAD1 and NADP1 and tetraprotonated [21]aneN7 confirm the participation of the phosphate appended to the ribose moiety of the adenosine nucleoside in the electrostatic and hydrogen bonding interactions.Cyclic voltammetric measurements denote significant alterations in the electrochemical behaviour of the NAD1/NADH or NADP1/NADPH couples as a function of the interaction between the dinucleotide and receptor. Introduction Interaction of polyammonium receptors with polycharged anionic species through non-covalent forces often results in the formation of supramolecular adducts displaying wellcharacterised chemical properties.In this respect, over the last two decades, much attention has been devoted to the interactions of macrocyclic polyammonium receptors with phosphate type anions and particularly with nucleotides like ATP, ADP or AMP.1 In many instances it was found that the formation of the non-covalent host–guest complex brought about important rate accelerations of the ATP hydrolysis into ADP and inorganic phosphate.In this respect macrocycles 1,4,7,13, 16,19-hexaaza-10,22-dioxatetracosane (bisdien) and 1,4,7,10, 13,16,19-heptaazacyclohenicosane ([21]aneN7) have proved to be the most eYcient catalysts.2,3 However, studies on the interaction of the ubiquitous nucleotides NAD1 and NADP1 with polyammonium receptors are very scarce.4 One of the most significant works in this sense is a recent paper of Lehn et al.5 in which it is reported that macrocycles derived from bisdien by mono- or bifunctionalisation with acridine moieties were able to recognise nucleotide and dinucleotide anions through a variety of intermolecular forces.NAD1 and NADP1 play a crucial role in a large variety of redox processes involved in most metabolic routes. In spite of the similarities between both coenzymes, NADP1 diVers † Supplementary data (SUPPL. 57457, pp. 1) are available from the British Library.For details of the Supplementary Publications Scheme, see ‘Instructions for Authors’, J. Chem. Soc., Perkin Trans. 2, available via the RSC Web page (http://www.rsc.org/authors). The supplementary data is also available on the RSC’s web server (http:// www.rsc.org/suppdata/perkin2/1999/23/). structurally from NAD1 only by the presence of an additional phosphate group esterified to the 29-hydroxy group of its AMP moiety (see Scheme 1), they are used diVerently in the metabolic redox processes.While NAD1 participates almost exclusively in oxidative degradations that yield ATP, NADP1 is confined with few exceptions to the reductive biosynthesis. It is the presence of the extra phosphate group in NADP1 that permits the proteins to diVerentiate between both coenzymes, which constitutes a clear example of molecular recognition in biochemical reactions. NAD and NADP cofactors are linked to the proteins through a variety of weak intermolecular forces. Among them, electrostatic interactions or/and hydrogen bonding with lateral chains of basic amino acids like arginine or lysine seem to be well defined features in these domains.6 It is certain that these interactions are modulating the redox properties of these dinucleotides contributing to the specificity of a determined enzyme.Here we report on the interaction of the polyammonium receptor 1,4,7,10,13,16,19-heptaazacyclohenicosane ([21]ane- N7) 7 with NADP1 and NAD1 nucleotides (see Scheme 1) and we show that this receptor, in its protonated forms, is able to recognise NADP1 over NAD1 throughout a wide pH range.In order to gain more insight about the factors contributing to the selective recognition of NADP1 over NAD1 by [21]aneN7, a wide set of energy minimisation, torsion angle search and molecular dynamics calculations have been carried out on the interaction of the dinucleotides with H4([21]aneN7)41. Experimental Materials [21]aneN7 was synthesised as described in ref. 7 and handled as24 J. Chem. Soc., Perkin Trans. 2, 1999, 23–32 its heptahydrochloride salt. b-Nicotinamide adenide dinucleotide sodium salt (HPLC > 98%), b-nicotinamide adenine dinucleotide phosphate monohydrate and b-nicotinamide adenide dinucleotide reduced disodium salt (HPLC > 97%) were Fluka reagents. The concentration of the dinucleotides was accurately determined by potentiometric titration. NaClO4 used as supporting electrolyte was purified according to a published procedure.8 All the other chemicals were Merck reagent grade, and were used without further purification.Emf measurements The potentiometric titrations were carried out, in 0.15 mol dm23 NaClO4 at 298.1 ± 0.1 K, by using the experimental procedure (burette, potentiometer, cell, stirrer, microcomputer, etc.) that has been fully described elsewhere.9 The acquisition of the emf data was performed with the computer program PASAT.10 The reference electrode was an Ag/AgCl electrode in saturated KCl solution. The glass electrode was calibrated as an hydrogen-ion concentration probe by titration of well-known amounts of HCl with CO2-free NaOH solutions and determining the equivalent point by the Gran’s method 11 which gives the standard potential, Eo, and the ionic product of water (pKw = 13.73(1)).The computer program HYPERQUAD12 was used to calculate the protonation and stability constants. The DISPO program was used to calculate the distribution plots.13 The titration curves for each system were treated either as a single set or as separated curves without significant variations in the values of the stability constants. The basicity constants of [21]aneN7 were taken from ref. 14. The basicity constants of NAD1 and NADP1 were deter- Scheme 1 Receptor and substrate molecules. mined under the same experimental conditions.15 The titration of the systems containing either NAD1 or NADP1 and [21]aneN7 were performed either from acid to basic pH or in the reverse sense in order to check the reversibility of the systems.The pH range used for computing the stability constants was usually 3.0–9.5 and at least three titration curves were performed for each one of the studied systems (ca. 250 experimental points, concentrations of [21]aneN7 were 1 × 1023 mol dm23 and those of the nucleotides were in the range 1 × 1023 mol dm23 to 3 × 1023 mol dm23). NMR measurements The 1H and 13C NMR spectra were recorded on Varian UNITY 300 and UNITY 400 spectrometers, operating at 299.95 and 399.95 MHz for 1H and at 75.43 and 100.58 MHz for 13C.The spectra were obtained at room temperature in D2O solutions. For the 13C NMR spectra dioxane was used as a reference standard (d = 67.4) and for the 1H spectra the solvent signal was the reference. The 31P NMR spectra were recorded at 121.42 MHz on a Varian UNITY 300 MHz. Chemical shifts are relative to an external reference of 85% H3PO4.Adjustments to the desired pH were made using drops of HCl or NaOH solutions. NOE measurements were not possible due to unfavourable correlation times even with the use of spin-lock techniques like ROESY. The pH was calculated from the measured pD values using the correlation, pH = pD 2 0.4.16 Molecular dynamics calculations The theoretical calculations have been run in a Power Indigo2 SGI computer, and as forcefield CHARMm MSI version has been used. Two diVerent trajectories for molecular dynamics calculations at 1000 K have been calculated, 100 ps for free NAD1 and NADP1 and H4([21]aneN7)41, and 500 ps for the diVerent NAD1–H4([21]aneN7)41 and NADP1–H4([21]ane- N7)41 complexes formed.Adducts with the tetraprotonated macrocycle were selected, since in their pH range of formation, all four protons would be most likely bound to the macrocycle. On the other hand, structural information for this cation was available from the literature,3 and can be used as a standard check for the conformational calculations.In energy calculations a 14 Å cut-oV distance was employed, with a switching function between 11.0 and 14.0 Å for van der Waals terms. The system was prepared by energy minimisation, 1000 steps of conjugated gradient, 10 ps of heating and equilibration at 1000 K velocity rescaling every 50 fs step. Starting from the final structure of this equilibration a set of dynamic simulations, as previously described, were performed.During the molecular dynamic trajectories a structure was relaxed through energy minimisation at 0 K every 10 ps. For each molecular dynamics trajectory, free compound lowest energy structures were selected to build up the complex between H4([21]aneN7)41 and NAD1 or NADP1. Computing was performed either with dielectric constant e = 1 or e = 80 without observing significant diVerences in the results. Electrochemistry Cyclic voltammograms were obtained on a conventional threeelectrode cell with a potentiostat (HQ100), a signal generator (Newtronics 200P) and an x–y recorder (Riken-Denshi F35).A hanging mercury drop electrode (HMDE) (Metrohm AGCH9100), platinum and glassy electrodes were used as working electrodes. A platinum-wire auxiliary electrode and a saturated calomel reference electrode (SCE) completed the standard three-electrode arrangement. The scan rate was varied from 0.01 to 1000 V s21. All experiments were carried out under an argon atmosphere in a cell thermostated at 298.0 ± 0.1 K.In all samples NaClO4 0.15 mol dm23 was used as supporting electrolyte. Before theJ. Chem. Soc., Perkin Trans. 2, 1999, 23–32 25 series of runs, solid electrodes were cleaned and activated. Electrochemical pre-treatment was performed in blank solutions by applying 11.50 V vs. SCE for 10 min followed by 21 V for 1 min.17 Modified carbon paste electrodes were prepared in a conventional manner 18 by thoroughly mixing a 30 mg sample of high purity powdered graphite (Koch Light) and nujol oil in a mortar and pestle.The paste was packed into the tip of a 3 mm diameter holder in contact with the glassy carbon electrode. Results and discussion Basicity constants of NAD1 and NADP1 NAD1 and NADP1 present, in the pH range 3.0–9.5, one and two protonation steps, respectively (NAD1, log KHA = 3.88(2); NADP1, log KHA = 5.86, log KH2A = 3.87(1)). The basicity constant of 3.8 logarithmic units found in both dinucleotides corresponds, as proved by NMR, to protonation of the nitrogen atom labelled as N1 of the adenine moiety. In fact, in the pH range 3–5 13C signals assigned to carbon atoms A6 and A2 show large variations in chemical shift (for the labelling Scheme 1) while below and above this pH range no further significant variations are observed.First protonation of NADP1 (log KHA= 5.86(1)) occurs, however, on the phosphate group attached to the ribose moiety of the adenine domain. 31P NMR spectra recorded at diVerent pH values show a 5 ppm downfield shift of the singlet signal assigned to this phosphorus in the pH range where this protonation takes place (pH 7–5). Interaction of NAD1 and NADP1 with [21]aneN7 Potentiometric studies on the interaction of [21]aneN7 with NAD1 and NADP1 carried out in the pH range 3.0–9.5, show the formation of adduct species of 1 : 1 stoichiometry (HpLA(p 2 n); L = [21]aneN7, A = NAD1 or NADP1) with protonation degrees varying from 3 to 7 for NAD1 and from 3 to 8 for NADP1.The values of the stepwise stability constants (HpLp1 1 An2 = HpLA(p 2 n)1, log Kp) being: NAD1, log K3 = 3.15(2), log K4 = 4.27(3), log K5 = 5.02(2), log K6 = 6.62(3) and log K7 = 7.59(3); NADP1, log K3 = 3.38(2), log K4 = 4.74(3), log K5 = 7.19(2), log K6 = 10.07(4), log K7 = 11.84(3) and log K8 = 9.35(4), K8 refers to the addition of protonated NADP1 to the heptaprotonated macrocycle. Although the values of the constants obtained for NADP1 are clearly higher than those for NAD1, the diVerent basicities of both dinucleotides makes it diYcult to draw conclusions on the selective recognition of one of them over the other and on the pH range where discrimination is achieved.However, as we have proposed3,19,20 a distribution diagram for the ternary system NAD1–NADP1–[21]aneN7 in which the overall amounts of complexed NAD1 and NADP1 Fig. 1 Calculated distribution diagram for the system NADP1– NAD1–[21]aneN7 (1023 mol dm23 for all the reagents).Overall amounts of complexed NAD1, NADP1 and free [21]aneN7 are plotted versus pH. are represented as a function of pH, oVers an unambiguous means of establishing selectivity patterns in these kinds of systems. Such a representation (Fig. 1) clearly shows that NADP1 is selectively recognised over NAD1 by [21]aneN7 throughout the whole pH range studied. The phosphate group attached to the ribose moiety would play an important role since it yields a higher negative charge on NADP1 and can form additional salt bridges with the polyammonium groups of the receptor. 31P NMR proves the participation of this phosphate group in the interaction with protonated [21]aneN7. The resonance of this phosphorus bears an upfield shift of ca. 1 ppm when adding an equimolar amount of [21]aneN7 at fixed pH. Higher ratios of NADP1–[21]aneN7 do not yield further shifts which, on the other hand, supports the 1 : 1 substrate–receptor stoichiometries inferred from the pH-metric titrations.Molecular dynamics The calculated structure of the H4([21]aneN7)41 cation agrees quite well with that found in the X-ray structure of the solid compound [21]aneN7?4HCl?H2O.3 Nevertheless, some slight diVerences between the X-ray structure and the calculated structure are observed which, most likely, arise from the diVerent crystallisation and molecular simulations conditions.21 As from X-ray data, the tetraprotonated macrocycle itself has a boat-shaped ellipsoid form.3 Torsion angles around the ring (Table 1) can be used to compare calculated and X-ray structures.Although with some deviations with respect to standard values, the average dihedral C–C–N–C angles for the structures of lower energy in molecular dynamics are close to the trans configuration around four of the seven nitrogen atoms in the macrocycle (average value 160 ± 208). Depending on the approach direction, there are two nitrogens with a mixture of trans and gauche conformations.Finally, the last nitrogen is more appropriately described as being in the gauche conformation 60 ± 208. For example, Fig. 2A shows the lowest energy conformation of H4([21]aneN7)41 from a 100 ps dynamic simulation. Similarly to the solid state results,3 intramolecular hydrogen bonding between nitrogens of the macrocycle has not been observed. In general, nucleosides, nucleotides and their derivatives are very flexible molecules with torsion angles falling in a wide range.22 For free NAD1 and NADP1 there were not crystallographic data available in the Cambridge Crystallographic Data Base; the only full set of three-dimensional data found corresponds to NAD1 co-ordinated to Li1 and for complexes with proteins of NAD1 and NADP1.In all these complexes both dinucleotides adopt conformations with energies much higher Table 1 C–C–N–C and C–N–C–C torsion angles for tetraprotonated [21]aneN7 Atoms C(3)–C(2)–N(1)–C(21) C(2)–N(1)–C(21)–C(20) C(2)–C(3)–N(4)–C(5) C(3)–N(4)–C(5)–C(6) C(5)–C(6)–N(7)–C(9) C(6)–N(7)–C(8)–C(9) C(8)–C(9)–N(10)–C(11) C(9)–N(10)–C(11)–C(12) C(11)–C(12)–N(13)–C(14) C(12)–N(13)–C(14)–C(15) C(19)–C(15)–N(16)–C(17) C(15)–N(16)–C(17)–C(18) C(17)–C(18)–N(19)–C(20) C(18)–N(10)–C(20)–C(21) RXa 160.7 2164.4 179.9 179.2 171.5 175.3 2172.5 177.6 2167.2 95.5 2175.4 78.6 73.7 61.2 Free b 168.1 172.2 156.6 150.6 2163.1 163.1 2150.6 156.6 168.1 102.3 169.6 246.2 245.8 86.9 Complex c 286.4 2142.4 139.0 162.7 2164.2 2101.2 2174.7 2156.6 88.2 76.3 88.4 244.2 2173.9 170.1 a Data taken from ref. 3. b Average values from 100 ps molecular dynamics calculations for free H4[21]aneN7. c Average values from 500 ps molecular dynamics calculations of NADP1(4)–H4[21]aneN7.26 J. Chem. Soc., Perkin Trans. 2, 1999, 23–32 than those of the generally accepted “standard low energy structures” of nucleotides.23 Therefore, owing to the inherent flexibility of NAD1 and NADP1 a restricted conformational Fig. 2 Lowest energy structures of free compounds from 100 ps dynamic trajectory. (A) H4([21]aneN7)41; (B) NAD1(1); (C) NAD1(2); (D) NADP1(3) and (E) NADP1(4). search analysis of torsional angles of the central part of both molecules was carried out. In Table S1, available as supplementary data,† a comparison of NAD1 and NADP1 in their free forms or as part of protein complexes is presented. Particular attention was focused on the dihedral angles of the adenosine moiety, P1–AO59–AC59–AC49 and AO59–AC59–AC49–AO19, which can be related to the orientation of the third phosphate group in NADP1 (for the labelling see Scheme 1).Either for NAD1 or NADP1 two diVerent structures, representative of a distinct part of conformational space, have been used as free compounds for the initial docking to the macrocycle. Both structures may be considered as the average of two diVerent conformational families.Therefore, four diVerent configurations have been used for the free dinucleotides (NAD1(1), NAD1(2), NADP1(3) and NADP1(4)). The lowest energy conformations of NAD1(1), NAD1(2), NADP1(3) and NADP1(4) from 100 ps dynamic trajectory are shown in Figs. 2B–2E. Although the energy diVerence between the (1) and (2) forms for NAD1 (118.5 kcal mol21) and the (3) and (4) forms for NADP1 (83.3 kcal mol21) is quite significant, conformational analysis using six dihedral angles of the central part of NAD1 and NADP1 indicates that both configurations are always at the regions of minimum energy, or close to them.However there are quite important energy barriers between some of the dihedral angles of both configurations and these may be critical for conformational interconversion, for instance, between the torsion angles P2–NO59–NC59–NC49 vs. P1–AO59–AC59– AC49 for free NAD1 and NADP1 compounds. Pentosa rings are important flexible regions in NAD1 and NADP1 structures, and they should also be considered in the analysis. The furanose conformation of adenosine nucleoside —sugar (A)—in NAD1(1) and NAD1(2) is C19-exo (pseudorotation angle, P, average value is 1358).Whereas the other furanose ring—sugar (N)—has a typical C19-exo (P = 1398) and C29-endo (P= 1668) for conformations (1) and (2) of NAD1, respectively. NADP1 sugar (N) conformations are similar to those in NAD1, C29-endo (P = 1458) and C29-endo (P = 1688) for the forms (3) and (4) of NADP1, respectively.However, NADP1 sugar (A) configurations are quite similar to the NAD1 ones, C49-exo (P = 488) for form (4), and C39-endo (P = 178) for form (3). Furanose conformations for both sets of structures (forms (1) and (2) of NAD1 and forms (3) and (4) of NADP1) fall close to or within the two preferred pseudorotation angle regions.23–25 The P values for sugar (N) in NAD1 and NADP1 are, for forms (1), (2), (3) and (4), very close to the standard C29-endo conformation (1448 < P<1808). However, the inclusion of the phosphate group in the adenosine nucleoside of NADP1 modifies the furanose ring puckering to a C39-endo configuration (08 < P<368) while sugar (A) in NAD1 retains a conformation close to C29-endo. Once the free compounds had been analysed, the next step was to study the four diVerent complexes.The initial docking of the two distinct NAD1 and NADP1 forms to H4([21]aneN7)41 was empirical; electrostatic and van der Waals forces being the only energetic contributions considered.Molecular dynamics with a trajectory of 500 ps were simulated for each one of the four possible complexes. As previously pointed out, 50 structures (each 10 ps) were minimised. The final comparative conformational analysis was done using the lowest energy complexes. The 50 minimised structures from 500 ps dynamic trajectories can be clustered in diVerent conformational families, which can be considered as representative of an important collective of configurations.For the four simulated complexes, the cluster of configurations with the lowest co-ordinate rmsd (root mean square deviation) usually includes those conformations with lower energy values. As an example, the 10 structure cluster for the NADP1(4)–H4([21]aneN7)41 adduct is shown in Fig. 3.J. Chem. Soc., Perkin Trans. 2, 1999, 23–32 27 Table 2 Conformational energy values for 500 ps dynamic trajectory for the more stable structures of NAD1(1)–H4([21]aneN7)41, NAD1(2)– H4([21]aneN7)41, NADP1(3)–H4([21]aneN7)41 and NADP1(4)–H4([21]aneN7)41 adduct species NAD1(1)–H4([21]aneN7)41 NAD1(2)–H4([21]aneN7)41 NADP1(3)–H4([21]aneN7)41 NADP1(4)–H4([21]aneN7)41 EA a 2217.9 299.4 2338.4 297.1 EB 2383.8 2383.8 2383.8 2383.8 EC 2601.7 2493.2 2722.3 2490.9 ED 2588.6 2856.8 2876.4 2906.6 EF 113.1 2373.6 2154.1 2425.7 a Energy in kcal mol21.EA: conformational energy of forms (1), (2), (3) and (4) of free NAD1 and NADP1.EB: conformational energy of free H4([21]aneN7)41. EC: conformational energy = EA 1 EB; ED = conformational energy of the diVerent complexes. EF = conformational energy = ED 2 EC. Apart from the lowest energy configuration, the rest of the conformations are among those more stable from an energetic point of view. The rmsd value for the whole set of atoms and 10 configurations is quite low, 0.87. Thermodynamic stability of NAD1 and NADP1–macrocycle complexes can be inferred from energy diVerences between complex and free compounds.The results for the four complexes NAD1(1)–H4([21]aneN7)41, NAD1(2)–H4([21]ane- N7)41, NADP1(3)–H4([21]aneN7)41, and NADP1(4)–H4([21]- aneN7)41 are shown in Table 2. As can be clearly seen in Table 2, only three of the four complexes studied may be considered as stable (DEcomplex < 0), whereas NAD1(1)–H4([21]aneN7)41 has DEcomplex > 0. In spite of the significantly lower conformational energies calculated for forms (1) and (3) in the free dinucleotides, the complexes between forms (1) and the macrocycle are less favourable than those with forms (2).In other words, complexation with H4([21]aneN7)41 shifts the dinucleotide conformational equilibrium towards configurations (2) and (4). The lowest energy complex structures, NAD1(2)–H4([21]aneN7)41 and NADP1(4)– H4([21]aneN7)41, are shown in Fig. 4. The overall energy diVerence between both simulated complexes (52 kcal mol21) agrees with the experimental results pointing out that NADP1 is better recognised by the macrocycle than NAD1.In general, the macrocycle displays when complexed a more opened and flattened configuration than when free.26,27 In this sense, the C–C–N–C torsion angles in the lowest energy complex (NADP1(4)–[21]aneN7) are quite diVerent to those found in the free macrocycle (Table 1). Only two of the seven nitrogen atoms are close to the trans conformation while the other ones display a mixture between the trans and gauche configurations depending on the approach direction.The rmsd value range for the comparison between the free and the complexed macrocycle is 2.4–3.5, the higher rmsd value corresponding to the less stable complex (NAD1(1)–H4([21]aneN7)41). Likewise, NAD1 and NADP1 conformations are diVerent in their complexed and free forms. The largest torsion angle modifications correspond to the AO59–AC59–AC49–AO19 dihedral angles, particularly for NADP1(4)–[21]aneN7 complex with the lowest total energy (163.48 in the free form and 232.88 in the complex).Similarly to NAD1 and NADP1 free compounds, the search conformational analysis for the six dihedral angles of the cen- Fig. 3 View of a cluster of 10 structures from 500 ps dynamic simulation, including lowest energy conformation, of NADP1(4)–[21]aneN7 complex. tral part of both molecules in the complexes shows that both configurations are always at or close to the regions of minimum energy, especially in the case of forms (2) and (4).Furthermore, for some torsion angles the energy barriers are high enough to avoid a direct interconversion between forms (1) and (2) and between forms (3) and (4). These energy barriers are particularly high for dihedral angles P1–AO59–AC59–AC49 vs. AO59– AC59–AC49–AO1 either in NAD1 or NADP1. Both torsion angles are related to the mobility of the adenosine moiety and to the interaction with the macrocycle.Puckering of furanose rings is significantly modified during complex formation. The smallest changes occur in NADP1 complexes, particularly in NADP1(4)–[21]aneN7 (complex with the lowest energy). The configurations of sugar (A) in NADP1(3)–[21]aneN7 and NADP1(4)–[21]aneN7 complexes are C49-exo (P = 588), and O49-endo (P = 808), respectively. In the case of NAD1(1) and NAD1(2)–macrocycle complexes, sugar (A) conformations are C19-exo (P = 1218), and O49-endo (P = 858), respectively.Sugar (N) presents C49-exo (P = 698), C19-exo (O49-endo) (P = 1088), C19-exo (P = 1368) and C29-endo (P = 1528) conformations for NAD1(1), NAD1(2), NADP1 (3) and NADP1(4)–H4([21]aneN7)41 complexes, respectively. In general, pentose configurations in the four complexes are relatively far from the two preferred pseudorotation angles (C39-endo and C29-endo).21,24 This fact may be partially compensated for by the modification in the hydrogen-bond network, 1c,26,27 and the gain in favourable electrostatic interaction between the phosphate groups and polyammonium groups. The number of intramolecular hydrogen-bonds (H-bonds) is 3 for free NAD1(1), NAD1(2) and NADP1(3), and 6 for NADP1(4).The total number of H-bonds (intra- plus intermolecular) increases considerably in the four simulated complexes. All intramolecular H-bonds for the four complexes are located in the NAD1 and NADP1 dinucleotides. The number of intramolecular hydrogen bonds in NAD1(1)–macrocycle is 3 while there are 10 intermolecular hydrogen bonds between NAD1 and H4([21]aneN7)41.For NAD1(2)–macrocycle the number of intra- and intermolecular hydrogen bonds is 0 and Fig. 4 Structural comparison between NAD1(2)–[21]aneN7 and NADP1(4)–[21]aneN7 lowest energy conformations from 500 ps dynamic simulations.28 J. Chem. Soc., Perkin Trans. 2, 1999, 23–32 12, respectively. The number of intramolecular hydrogen bonds is drastically reduced from 3 in free NADP1(3) to 0 in the NADP1(3)–H4([21]aneN7)41 complex, which may partially be compensated for by 11 intermolecular hydrogen bonds.Finally for the complex NADP1(4)–macrocycle the number of intermolecular hydrogen bonds is 11. Obviously, another contribution to complex stabilisation is electrostatic interaction between phosphate groups and positively charged nitrogens. The distances between P atoms and charged N for 500 ps dynamics minimum energy structures are shown in Table 3 for the four studied complexes.In NAD1(1)–H4([21]aneN7)41 each phosphate group lies close to two adjacent charged nitrogens in a symmetrical orientation (P1 is closer to N4 and N10, and P2 to N16 and N19). In NAD1(2)–H4([21]aneN7)41, P2 is close and symmetrically positioned with respect to the four charged nitrogens, P1 being close to the contiguous charged nitrogens N16 and N19. The phosphorus atom of the phosphate group hanging from the adenosine moiety (P3) is, in NADP1(3)–H4([21]aneN7)41, far away (>9 Å) from the protonated nitrogens of the macrocycle, whereas the orientation of P1 and P2 is similar to that of P1 and P2 in NAD1(1)–H4([21]aneN7)41.However, in NADP1(4)– macrocycle, P3 is close to the contiguous charged nitrogens N16 and N19, P1 is relatively close to the four charged nitrogens, particularly to N4 and N19, and P2 is pointing towards N10 and N16. Considering distances P–N < 5.0 Å as representative of the electrostatic interaction, the larger and the shorter electrostatic contribution correspond to NADP1(4)–H4([21]aneN7)41 and NAD1(1)–H4([21]aneN7)41 complexes, respectively, which, on the other hand, are the complexes with the lowest and highest total energy.NAD1(2) and NADP1(3) complexes with H4([21]- aneN7)41 have a similar number of P–N distances < 5.0 Å, although, in general, it seems that those for the second complex are shorter, indicating a possible higher electrostatic favourable contribution.It should be pointed out that in the initial minimised structures used for the molecular dynamic calculations the distance between the adenosine phosphate group in NADP1 and any nitrogen atom of the macrocyclic receptor was always longer than 5 Å. The global analysis of intermolecular H-bonds formed, and P–N distances (Table 3) can explain the relative stability 1c,26,27 of the four complexes. NADP1(4)–H4([21]aneN7)41 (lowest energy complex) has the largest electrostatic contribution, (interaction of three P with four charged nitrogens) and 11 intermolecular hydrogen bonds.Besides an important electrostatic contribution NAD1(2)–H4([21]aneN7)41 has the largest number of intermolecular hydrogen bonds. Finally, NAD1(1)–H4([21]aneN7)41 presents the lowest electrostatic contribution being, hence, the least stable of the four simulated complexes. Table 3 Distance between phosphorus atoms of phosphate groups and charged nitrogen atoms in NAD1(1)–H4([21]aneN7)41, NAD1(2)– H4([21]aneN7)41, NADP1(3)–H4([21]aneN7)41 and NADP1(4)–H4([21]- aneN7)41 adduct species P–N1 Distance in Å NAD1(1)–H4([21]aneN7)41 NAD1(2)–H4([21]aneN7)41 NADP1(3)–H4([21]aneN7)41 NADP1(4)–H4([21]aneN7)41 P1 P2 P1 P2 P1 P2 P3 P1 P2 P3 N11 3.35 5.12 5.82 3.62 5.41 3.76 9.37 4.87 3.95 9.72 N17 5.46 3.18 3.51 3.83 3.87 4.89 10.96 3.21 5.51 5.00 N110 5.30 3.15 3.74 3.70 3.60 5.44 11.46 3.44 5.74 3.94 N116 4.16 5.08 6.24 3.80 4.96 3.21 10.823 3.90 3.87 8.03 A point that deserves particular attention, is the fact that interconversion between the two conformations considered for both dinucleotides is practically prevented in their complexes by the high energy torsional barriers (P1–AO59–AC59–AC49 vs.AO59–AC59–AC49–AO19). In other words, once the adducts are formed, there are important sterically unfavourable interactions which avoid changes from NADP1(3) to NADP1(4) since this would imply a rotation around torsion angles P1–AO59–AC59– AC49 and AO59–AC59–AC49–AO19.This may be interpreted as if the equilibrium between conformations (3) and (4) was partially shifted to configuration (4) (lowest energy conformation for free compounds) due to the formation of the more stable NADP1(4)–H4([21]aneN7)41. Finally, the global analysis of simulation calculations seems to point out that selective recognition of NADP1 over NAD1 by [21]aneN7 could be largely achieved by the participation of the third additional phosphate group in the adenosine moiety of NADP1.This observation is in good agreement with experimental data and supports the conclusions derived from those results. Electrochemistry The electrochemical pattern of NAD1 and analogue compounds at mercury electrodes in aqueous and non-aqueous media, has been widely studied.28,29 In neutral alkaline unbu Vered solutions of nicotinamide compounds, a first reduction step takes place at 21.08 V vs. SCE followed by a second ill-defined cathodic process at 21.3 V.On reversal of the potential scan, an anodic process at 20.17 V is observed. Unfortunately, mercury electrodes are unable to completely analyse the interaction between nicotinamide compounds and polyammonium macrocycles because addition of stoichiometric amounts of receptor to a nicotinamide solution causes a catalytic eVect on the hydrogen discharge that obscures NAD1 reduction. In addition, new voltammetric peaks appear in the 0.00 to 20.60 V potential range due to the mercury–macrocyclic ligand interaction, which parallels the results reported for analogue macrocyclic compounds.30 Accordingly, the electrochemical redox behaviour of NAD1 and NADP1 has been monitored by cyclic voltammetry at the glassy carbon electrode (GCE).As is depicted in Fig. 5a, in alkaline solutions the main reduction peak (A) appears at 21.28 V and is followed by an ill-defined shoulder (B) near to 21.5 V. An anodic peak (C), poorly defined, was observed at 20.27 V during the subsequent anodic scan.This peak is, however, absent in the initial positive scans, indicating that the oxidation process aVects a species generated during prior cathodic steps. In agreement with the literature, peak A corresponds to the one-electron reversible reduction of the oxidised nicotinamide to a radical intermediate, which rapidly dimerises (eqns. (1) and (2)). The dimerisation process inhibits, at least NAD1 1 e2 NAD? (1) 2 NAD? (NAD)2 (2) partly, the anodic counterpart (A9) of the electrochemical step (A).At slow sweep rates (v = 0.01–0.50 V s21), the current function, ip/Acv1/2 remains essentially constant with the scan rate. The peak potential becomes 30 mV per unit of log v more negative while the peak to half-peak potential diVerence (Ep 2 Ep/2) tends to 40 mV at low scan rates. All these results satisfy the essential diagnostic criteria for a one-electron reversible transfer followed by irreversible dimerisation.31 The electrode process B is attributable to the proton-assisted reduction of the NAD? radical (eqn. (3)), but in part may be NAD? 1 H1 1 e2 NADH (3)J.Chem. Soc., Perkin Trans. 2, 1999, 23–32 29 due to the reduction of NAD1 which diVuses through the depleted layer.28 Finally, the anodic peak C can be assigned to a one-electron oxidation of the dimeric species to the original nucleotide (eqn. (4)). 1/2 (NAD)2 NAD1 1 e2 (4) In neutral and acidic media, the oxidation process C is followed by a second anodic wave (D) at 10.42 V.The intensity of the wave D increases as the switching potential is negatively shifted, suggesting that this electrochemical process involves a product associated to the electrode process B. This is confirmed by the increase in peak current observed after holding the potential at a value close to 21.5 V and then resuming the scan. It is interesting to note that reduction of NAD1 and NADP1 at carbon electrodes takes place at potentials 200 mV more negative than at mercury electrodes.Assuming that peak A consists essentially of a one-electron transfer, relatively large values of the diVusion coeYcient were obtained (4.2 × 1026 and 4.0 × 1026 cm2 s21 for NAD1 and NADP1, respectively). In addition, the anodic counterpart of peak A is absent even at fast sweep rates. These features support the hypothesis that a significant portion of the nicotinamide species is reduced to dihydronicotinamide in the 21.2 to 21.5 V potential range at the GCE.The observed peak potential D agrees with that reported at the pyrolitic graphite electrode in NAD1 solutions,28 but it is slightly less positive than the values for the dihydropyridine oxidation in aqueous solution of NADH.32 Upon repetitive cyclic voltammetry in the extended potential range, the cathodic region remains essentially constant suggesting that NAD1 is regenerated during anodic steps.To elucidate the observed electrochemical pattern, the cyclic voltammetric response of NADH and NADPH solutions was studied. It has been well established that these species undergo a two-electron electrochemical oxidation to yield NAD1 and NADP1 at too positive potentials to allow the use of mercury Fig. 5 Cyclic voltammograms at the GCE in neutral solutions a) NADP1 (1.95 × 1023 mol dm23) in the absence (continuous line) and in the presence (dotted line) of [21]aneN7 (3.80 × 1023 mol dm23); b) NADH (2.00 × 1023 mol dm23); continuous line: initial anodic scan; dotted line: second scan after the potential was extended to 21.5 V vs.SCE. electrodes. However, the electrochemical processes at solid electrodes are complicated by coupled chemical reactions, adsorption of the reactants onto the electrode surface 32–35 and possible involvement of superficial oxides in hydride abstraction from NADH.34 This last phenomenon may be responsible for the large background currents observed at carbon and platinum electrodes.33 In agreement with the literature, neutral solutions of NADH exhibit a prominent anodic peak (E) at 10.76 V as can be seen in Fig. 5b. This is followed, in the subsequent cathodic scan, by a reduction peak at 21.28 V, corresponding to reduction of the previously generated NAD1 (process A). In the second and successive scans, an additional peak was observed at 10.33 V. This must be assigned to a typical prepeak,36 resulting from adsorption of the NAD1 produced in the oxidation of NADH.32–35 The foregoing set of data support the hypothesis that signifi- cant amounts of dihydronicotinamide species are generated during the reduction of NAD1 and NADP1 at carbon electrodes.In addition to surface phenomena, competitive chemical reactions, namely acid catalysed disproportionation of (NAD)2 into NAD1 and NADH, and hydrolysis of NADH may occur.37 Accordingly, the electrochemical process D can be tentatively assigned to the NADH oxidation mediated by an adsorbed layer of NAD1,38 the overall process possibly being described by eqn.(5). NADH NAD1 1 H1 1 e2 (5) The presence of adsorbed NAD1 at the electrode surface may control the overall electrode process in such a way that the electrochemical response is strongly dependent on electrode surface conditioning.30,31,34,35 In our experimental conditions, it can be expected that adsorption of NAD1 facilitates NAD? (or NAD1) reduction to NADH at the expense of the competing radical dimerisation process.Upon addition of increasing amounts of macrocyclic ligand to a solution of NAD1 at pH < 9, the reduction wave A is shifted anodically while the oxidation peak C disappears progressively (see Fig. 5a, dotted line). Peak D is in turn displaced toward less positive potentials. For ligand to metal ratios >1 the cyclic voltammetric curves remain almost unchanged, denoting the formation of 1 : 1 substrate–ligand complex species.The observed shift in the peak potential decidedly increases below pH 10 to a limiting value (typically 50–60 mV) which remains almost pH-independent until pH 3. This result is in accordance with the distribution diagrams calculated from potentiometric data. As expected, addition of a stoichiometric amount of macrocycle to a solution of NADH leads to an anodic shift of the peak potential (typically 50–60 mV). Both the morphology and peak current of the cyclic voltammograms for solutions containing the receptor are quite similar to that of NAD1 or NADH, suggesting that reversible complexation processes occur (eqns.(6) and (7)). Actual data (NAD1) 1 LHq (NAD1)LHq (6) (NADH) 1 LHq (NADH)LHq (7) suggest that adduct formation: a) facilitates the reduction of NAD1 to NADH whereas subsequent electrode oxidation of dihydropyridine species is hampered by the presence of the macrocyclic ligand, b) it inhibits almost completely the dimerisation reaction (eqn.(2)) at the expense of the dihydronicotinamide formation, and, c) eases the oxidation of NADH presumably mediated by adsorbed NAD1. EVect a) can be attributed, at first glance, to an essentially electrostatic eVect: upon adduct formation, the species formed became positively charged facilitating electron uptake. Inhibition of dimerisation of the intermediate radical most likely30 J. Chem. Soc., Perkin Trans. 2, 1999, 23–32 results from steric hindrance and an increase in energy barriers associated with the conformational changes required for dimer formation. Finally, eVect c) may be explained by assuming that adsorbed NAD1 acts as a mediator; i.e.NADH exchanges the electron with adsorbed NAD1 which, in turn, exchanges the electron with the electrode. Formation of NAD1–macrocycle adducts must result in a decrease of the energy barriers associated with such processes, but a detailed study concerning the eVect of adduct formation on the electrochemical oxidation of NADH species is in progress.To minimise adsorption influence, complementary experiments were carried out at modified carbon-paste electrodes (MCPE) immersed into a solution containing supporting electrolyte and eventually macrocyclic receptor. In the last decade, diVerent electrode surface pre-treatments, formation of polymeric films, and modified electrodes have been used as electrochemical sensors.39 The basic idea is that modified electrodes can preferentially concentrate target analytes from solutions.In MCPE the pre-concentrating agent is incorporated into the surface by mixing it with carbon-paste matrices. Among other procedures,34,35 poly(thionine)-modified electrodes 40 and conducted polymer electrodes 41 have been used to optimize the analytical oxidation of NADH. As can be seen in Fig. 6a, after inclusion of powdered NAD1 into a nujol oil–graphite powder matrix, cyclic voltammograms exhibit an increased signal/background ratio with respect to conventional cyclic voltammograms for NAD1 solutions at the GCE.Reduction of NAD1 takes place at similar potentials but the dimer oxidation peak is entirely absent. Fig. 6b presents the cyclic voltammogram for NADH–MCPE. The oxidation process E is clearly observed without adsorption pre-peak D in an initial anodic scan. Subsequent cathodic scans show the reduction peak of NAD1 supporting the electrochemical pattern previously described.Finally, immersion of NAD1–MCPE into a solution of the macrocyclic receptor as well as preparation of ligand–NAD1–MCPE (see Fig. 6c) leads to cyclic voltammograms that present an enhanced NADH oxidation peak after NAD1 reduction, denoting an almost quantitative NAD1 to NADH conversion during the electrochemical turnover. Fig. 6 Cyclic voltammograms at modified carbon-paste electrodes inmersed into a 0.15 mol dm23 NaClO4 solution at pH 7.0. a) NAD1– MCPE, b) NADH–MCPE, c) NAD1–[21]aneN7.Interaction between NAD1, NADP1 and [21]aneN7 has been also monitored on the basis of the competitive eVect of hexacyanoferrate( II) ions. Competitive methods have been proved of interest for polarographic analysis of metal–ligand equilibria.42 These methods were applied by Kimura et al.43 for studying the competitive interaction of mercury(II) ions and carboxylate ions with polyammonium receptors. The use of competitive anions for analysing the coordination equilibria between anionic species and receptors was proposed by Lehn et al.44 Earlier electrochemical studies on the complexation of ATP and carboxylate anions with polyammonium receptors on the basis of the competitive eVect of the hexacyanoferrate(II) ions have been reported.19,20 Cyclic voltammograms of [Fe(CN)6]42 solutions exhibit a one-electron reversible or quasi-reversible couple without adsorption complications at platinum, carbon and gold electrodes.This electrochemical pattern is altered by adding polyammonium ligands, the morphology of the cyclic voltammograms being heavily dependent on the pH and on the ligand to [Fe(CN)6]42 molar ratio (L/M).In solutions containing a small excess of receptor, [Fe(CN)6]42 is fully complexed at pH < 9. As previously described, the complexed species are electroactive and are oxidised almost reversibly at more positive potentials than hexacyanoferrate(II) ions.19,20,44–46 For our purposes, the relevant point to emphasise is that addition of a competitive anion, such as NAD1 or NADP1, produces a shift of the electrochemical parameters from those of the limiting cyclic voltammogram for the hexacyanoferrate(II)– ligand complex to those of uncomplexed [Fe(CN)6]42.This can be seen for example in Figs. 7 and 8, which represent the pH dependence of the apparent formal potentials and mean diffusion coeYcients determined from the cyclic voltammograms for solutions of uncomplexed hexacyanoferrate(II) (curves A), [Fe(CN)6]42 and ligand (curves B), and [Fe(CN)6]42, ligand and NAD1 (and NADP1) in identical total concentrations (curves C and D).Then, analysis of the electrochemical parameters of the [Fe(CN)6]32/[Fe(CN)6]42couple, namely peak potentials and currents, allows an estimate of the complexation of [Fe(CN)6]42 by polyammonium receptor since both [Fe(CN)6]42 and [Fe(CN)6]42–HqL adducts display reversible one-electron electrochemical processes.The cyclic voltammetric pattern corresponds to the general case of a system with diVerent electroactive species, which diVer in their formal potentials and diVusion coeYcients.47 For this type of system a series of procedures for determining equilibrium constants from the Fig. 7 pH-Dependence of the formal potential of the Fe(III)/Fe(II) couple for solutions containing: (A) [Fe(CN)6]42 2.0 × 1023 mol dm23; (B) [Fe(CN)6]42 2.0 × 1023 mol dm23 and [21]aneN7 2.0 × 1023 mol dm23; (C) [Fe(CN)6]42 2.0 × 1023 mol dm23, NAD1 2.0 × 1023 mol dm23 and [21]aneN7 2 × 1023 mol dm23; (D) [Fe(CN)6]42 2.0 × 1023 mol dm23, NADP1 2.0 × 1023 mol dm23 and [21]aneN7 2 × 1023 mol dm23.J. Chem.Soc., Perkin Trans. 2, 1999, 23–32 31 measurement of mean diVusion coeYcients are available.48 The Cottrell-type analysis of the diVusive portion of the cyclic voltammograms,49 as well as direct chronoamperometric measurements enable us to determine the mean diVusion coe Ycient for a mixture of electroactive species in equilibrium.In solutions containing NAD1, [Fe(CN)6]42, and the receptor, calculations based on the generalisation of the molar-ratio are applicable for determining the NAD1–L stoichiometry.50–52 The values of selected stepwise stability constants were calculated and agreed with those determined potentiometrically; NAD1, log K4 = 4.3(2), log K5 = 4.9(2); NADP1, log K4 = 4.6(2), log K5 = 7.0(2). In addition, the Viossat method provides a direct estimate of the stoichiometry and number of protons (w) involved in the first pH-dependent complexation step.53,54 Values of log K and w in close agreement with potentiometric data were found (NAD1; log K3 = 2.9(1); NADP1; log K3 = 3.2(1). Conclusions [21]aneN7 in its protonated forms interacts in aqueous solution with NAD1 and NADP1 forming adduct species of 1 : 1 stoichiometry which predominate in solution below pH 9.The stability constants for the interaction of [21]aneN7 with NADP1 are clearly higher, and mixed distribution diagrams show that it is selectively recognised over NAD1 by the polyammonium receptor throughout a wide pH range. 31P NMR spectra of solutions of the adducts suggest that one of the clues in the selective recognition of NADP1 is the involvement of the phosphate group esterified to the 29-hydroxy group of the AMP moiety of NADP1 in the interaction with the receptor. Molecular dynamics studies carried out on the interaction of tetraprotonated [21]aneN7 with NAD1 and NADP1 confirm the participation of the extra phosphate group of NADP1 in the recognition process.In fact, the most stable family of conformers for the NADP1–H4([21]aneN7)41 shows that this phosphate group interacts strongly with the two adjacent ammonium groups present in the tetraprotonated receptor which would be the factor responsible for the selective recognition of NADP1 over NAD1. Interaction of NAD1, NADP1 with the macrocyclic receptor decreases their observed reduc- Fig. 8 Variation with the pH of the mean diVusion coeYcient calculated from the cyclic voltammograms for solutions containing: (A) [Fe(CN)6]42 2.0 × 1023 mol dm23; (B) [Fe(CN)6]42 2.0 × 1023 mol dm23 and [21]aneN7 2.0 × 1023 mol dm23; (C) [Fe(CN)6]42 2.0 × 1023 mol dm23, NAD1 2.0 × 1023 mol dm23 and [21]aneN7 2 × 1023 mol dm23; (D) [Fe(CN)6]42 2.0 × 1023 mol dm23, NADP1 2.0 × 1023 mol dm23 and [21]aneN7 2 × 1023 mol dm23.tion potentials whereas the observed oxidation potentials of NADH and NADPH are shifted toward more positive values. Formation of receptor–nicotinamide dinucleotide adducts alters the electrochemical mechanism for NAD1 and NADP1 reduction at carbon electrodes, favouring the formation of dihydronicotinamide species rather than that of the dimer species. Acknowledgements We are indebted to the DGICYT (PB96-0792-CO2-01) and Generalitat Valenciana Project No. GV-D-CN-09-140 for financial support.References 1 Some examples of this chemistry are: (a) F. P. Schmidtchen, Top. Curr. Chem., 1986, 132, 101; (b) J.-M. Lehn, Angew. Chem., Int. Ed. Engl., 1988, 27, 89; (c) M. P. Mertes and K.-B. Mertes, Acc. Chem. Res., 1990, 23, 413; (d ) A. V. Eliseev and H.-J. Schneider, J. Am. Chem. Soc., 1994, 116, 6081; (e) B. Dietrich, Pure Appl. Chem., 1993, 65, 1457. 2 M. W. Hosseini, J.-M. Lehn and M. P. Mertes, Helv. Chim. Acta, 1983, 66, 2454. 3 A. Bencini, A.Bianchi, E. C. Scott, M. Morales, B. Wang, E. García-España, T. DeVo, F. Takusagawa, M. P. Mertes, K.-B. Mertes and P. Paoletti, Bioorg. Chem., 1992, 20, 8. 4 A recent study on the interaction of NAD1 and polyphosphate anions with a receptor containing guanidinium groups is: P. Schiessl and F. P. Schmidtchen, J. Org. Chem., 1994, 59, 509. 5 H. Fenniri, M. W. Hosseini and J.-M. Lehn, Helv. Chim. Acta, 1997, 80, 786. 6 Articles describing NAD1 and NADP1 binding sites in enzymes are for example: (a) M.J. Adams, M. Buehner, K. Chandrasekar, G. C. Ford, M. L. Hakert, A. Liljas, M. G. Rossman, I. E. Smily, W. S. Allison, J. Everse, N. O. Kaplan and S. S. Taylor, Proc. Natl. Acad. Sci. USA, 1973, 70, 1968; (b) N. S. Scrutton, A. Berry and R. N. Perham, Nature, 1990, 340, 39; (c) I. Gibbons, B. H. Gibbons, G. Mocz and D. J. Asai, Nature, 1991, 352, 640; (d ) J. Kuriyan, T. S. R. Krishna, L. Wong, B. Guenther, A. Pahler and C. H. Williams, Jr., Nature, 1991, 352, 172; (e) J.-M.Rondeau, F. Tete- Favier, A. Podjarny, J.-M. Reyman, P. Barth, J.-F. Biellman and D. Moras, Nature, 1992, 35, 469. 7 A. Bianchi, M. Micheloni and P. Paoletti, Inorg. Chem., 1985, 24, 3702. 8 M. Micheloni, P. May and D. R. Williams, J. Inorg. Nucl. Chem., 1978, 40 1209. 9 E. García-España, M.-J. Ballester, F. Lloret, J.-M. Moratal, J. Faus and A. Bianchi, J. Chem. Soc., Dalton Trans., 1988, 101. 10 M. Fontanelli and M. Micheloni, Proceedings of the I Spanish- Italian Congress on Thermodynamics of Metal Complexes, Peñíscola, Spain, 1990, published by Diputación de Castellón.Program for the automatic addition of thew burette additions and electromotive force readings. 11 G. Gran, Analyst (London), 1952, 77, 661; F. J. Rossotti and H. J. Rossotti, J. Chem. Educ., 1965, 42 , 375. 12 A. Sabatini, A. Vacca and P. Gans, Coord. Chem. Rev., 1992, 120, 389. 13 A. Vacca. University of Florence. Unpublished work, FORTRAN program to determine from the stability constants and mass balance equation the distribution of species in multiequilibria systems. 14 A. Bencini, A. Bianchi, P. Dapporto, E. García-España, M. Micheloni and P. Paoletti, Inorg. Chem., 1989, 28, 1188. 15 A previous study on the basicity constants of NAD1 and NADP1 can be found in: T. N. Briggs and J. E. Stuehr, Anal. Chem., 1975, 47,12, 1916. 16 A. K. Covington, M. Paabo, R. M. Robinson and R. G. Bates, Anal. Chem., 1968, 40, 700. 17 R. C. Engstrom, Anal.Chem., 1982, 54, 2310. 18 R. N. De Guzmán, Y.-F. Shen, B. R. Shaw, S. L. Suib and C. L. O’Young, Chem. Mater., 1993, 5, 1395. 19 A. Bencini, A. Bianchi, M. I. Burguete, P. Dapporto, A. Doménech, E. García-España, S. V. Luís, P. Paoli and J. A. Ramírez, J. Chem. Soc., Perkin Trans. 2, 1994, 569. 20 A. Andrés, J. Aragó, A. Bencini, A. Bianchi, A. Doménech, V. Fusi, E. García-España, P. Paoletti and J. A. Ramírez, Inorg. Chem., 1993, 32, 3418. 21 W. F. van Gusteren and H.J. C. Berendsen, Angew. Chem., Int. Ed. Engl., 1990, 29, 992.32 J. Chem. Soc., Perkin Trans. 2, 1999, 23–32 22 S. R. Holbrook, J. L. Sussman, R. W. Warrant and S.-H. Kim, J. Mol. Biol., 1978, 123, 631. 23 B. S. Reddy, W. Saenger, K. Mühlegger and G. Weimann, J. Am. Chem. Soc., 1981, 103, 907. 24 C. Altona and M. Sundaralingam, J. Am. Chem. Soc., 1972, 94, 8205. 25 K. Wüthrich, NMR of Proteins and Nucleic Acids, John Wiley and Sons, New York, 1986. 26 G. Papoyan, K.-J.Gu, J. Wiorkiewicz-Kuczera, K. Kuczera and K. Bowman-James, J. Am. Chem. Soc., 1996, 118, 1354. 27 L. Quian, Z. Sun, J. Gao, B. Movassagh, L. Morales and K.-B. Mertes, J. Coord. Chem., 1991, 23, 155. 28 (a) A. L. Underwood and J. N. Burnett, Electroanalytical Chemistry, vol. 6, A. J. Bard ed., Marcel Dekker, New York, 1972, pp. 1–85; (b) J. N. Burnett and A. L. Underwood, Biochem., 1965, 4, 2060; (c) J. N. Burnett and A. L. Underwood, J. Org. Chem., 1965, 30, 1154; (d ) B.Janik and P. J. Elving, Chem. Rev., 1968, 68, 295; (e) C. O. Schmakel, K. S. V. Santhanam and P. J. Elving, J. Am. Chem. Soc., 1975, 75, 5083. 29 K. S. V. Santhanam and P. J. Elving, J. Am. Chem. Soc., 1975, 75, 5482. 30 (a) M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., 1976, 2335; (b) M. Kodama and E. Kimura, J. Chem. Soc., Dalton Trans., 1978, 1081; (c) E. Kimura, A. Sakonaka, T. Yatsunami and M. Kodama, J. Am. Chem. Soc., 1981, 103, 3041. 31 (a) J. M. Saveant and E.C. R. Vianello, Acad. Sci. Paris, 1963, 256, 2597; (b) R. S. Nicholson, Anal. Chem., 1965, 37, 667. 32 (a) J. Moiroux and P. J. Elving, J. Am. Chem. Soc., 1980, 102, 6533; (b) W. J. Blaedel and R. A. Jenkins, Anal. Chem., 1975, 47, 1337. 33 (a) R. D. Braun, K. S. V. Santhanam and P. J. Elving, J. Am. Chem. Soc., 1975, 97, 5083; (b) J. Moiroux and P. J. Elving, Anal. Chem., 1979, 51, 346; (c) J. Moiroux and P. J. Elving, J. Electroanal. Chem., 1979, 102, 93. 34 (a) P. Hapiot, J.Moiroux and J. M. Savéant, J. Am. Chem. Soc., 1990, 112, 1337; (b) D. Lasr and M. Ariel, J. Electroanal. Chem., 1974, 52, 291. 35 S. Fukusumi, S. Koumitsu, K. Hironaka and T. Tanaka, J. Am. Chem. Soc., 1987, 109, 305. 36 A. J. Bard and L. R. Faulkner, Electrochemical Methods, Wiley, New York, 1990; R. H. Wopschall and Y. Shain, Anal. Chem., 1967, 39, 1514. 37 In general, there has been a failure to oxidize NADH cleanly at solid electrodes. However, it is well established that the first step in the electrochemical oxidation of NADH consists of an irreversible electron transfer. The resulting cation radical NADH~1 loses a proton to form the neutral radical NAD? which may participate in a second electron transfer (ECE mechanism) or in a homogeneous reaction with NADH~1 yielding NAD1. This disproportionation or half-regeneration mechanism (DISP1) has been well characterized for model compounds in organic solvents. See refs. 34, 35. 38 A. P. F. Turner, L. Karube and G. S. Wilson, eds. Biosensors, Fundamentals and Applications, Oxford University Press, 1987. 39 W. T. Bresnaham and P. J. Elving, Biochem. Biophys. Acta, 1981, 678, 151. 40 T. Osaka, K. Tanaka and K. Tokuda, J. Chem. Soc., Chem. Commun., 1993, 222. 41 M. Somasandrum and J. V. Bannister, J. Chem. Soc., Chem. Commun., 1993, 1629. 42 G. Schwarzenbach, R. Gut and G. Anderegg, Helv. Chim. Acta, 1954, 37, 937. 43 E. Kimura, A. Sakonaka, T. Yatsunami and M. Kodama, J. Am. Chem. Soc., 1981, 103, 3041. 44 F. Peter, M. Gross, M. W. Hosseini, J.-M. Lehn and R. B. Sessions, J. Chem. Soc., Chem. Commun., 1981, 1967. 45 (a) A. Bencini, A. Bianchi, M. I. Burguete, A. Doménech, E. García-España, S. V. Luís, M. A. Niño and J. A. Ramírez, J. Chem. Soc., Perkin Trans. 2, 1991, 1445; (b) J. Aragó, A. Bencini, A. Bianchi, A. Doménech and E. García-España, J. Chem. Soc., Dalton Trans., 1991, 319. 46 F. Peter, M. Gross, M. W. Hosseini and J.-M. Lehn, J. Electroanal. Chem., 1983, 144, 279. 47 (a) K. B. Oldham, J. Electroanal. Chem., 1991, 313, 3; (b) D. N. Blauch and F. C. Anson, J. Electroanal. Chem., 1991, 309, 313. 48 (a) V. Kacena and L. Matousek, Collect. Czech. Chem. Commun., 1953, 18, 294; (b) D. R. Crow, J. Electroanal. Chem., 1968, 16, 137; (c) D. R. Crow, Talanta, 1982, 29, 733; (d) D. R. Crow, Talanta, 1982, 29, 739; (e) D. R. Crow, Talanta, 1983, 30, 659; ( f ) D. R. Crow, Talanta, 1986, 33, 553. 49 (a) R. S. Nicholson, Anal. Chem., 1965, 37, 667; (b) D. S. Polcyn and I. Shain, Anal. Chem., 1966, 38, 370; (c) G. Cinzburg, Anal. Chem., 1978, 50, 375; (d ) G. Bontempelli, F. Magno and S. Daniele, Anal. Chem., 1985, 57, 1503. 50 A. Beltrán, D. Beltrán, A. Cervilla and J. A. Ramírez, Talanta, 1983, 30, 124. 51 J. H. Yoe and A. L. Jones, Ind. Eng. Chem. Anal. Ed., 1944, 16, 11. 52 A. Bianchi, A. Doménech, E. García-España and S. V. Luís, Anal. Chem., 1993, 65, 3137. 53 B. Viossat, Rev. Chim. Miner., 1972, 9, 737. 54 A. Doménech, E. García-España and J. A. Ramírez, Talanta, 1995, 42, 163. Paper 8/06004E

ISSN:1472-779X    

DOI:10.1039/a806004e

出版商:RSC    

年代:1999

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1998 25 1H and 13C NMR studies of para-substituted benzaldoximes for evaluation of the electron donor properties of substituted amino groups † Ryszard Gawinecki,*,a Erkki Kolehmainen b and Reijo Kauppinen b a Department of Chemistry, Technical and Agricultural University, Seminaryjna 3 PL-85-326 Bydgoszcz, Poland b Department of Chemistry, University of Jyväskylä, PO Box 35, FIN-40351 Jyväskylä, Finland 1H and 13C NMR spectra of seventeen (E)-benzaldoximes and three acetophenone oximes, both carrying substituted p-amino groups, have been recorded and discussed from the point of view of substituent effect.The resonance effect of these substituents is not transmitted strongly to CH]] NOH group. However, it is found that the chemical shift of Cpara depends linearly on sR o values. This dependence has been used to calculate the resonance substituent constants for the less common amino groups and the 1-pyrrolidine group is found to be the most powerful electron donor among the substituents studied.Introduction Substituted amino groups, which are very interesting from the point of view of differentiation of their electron donor properties, comprise a wide range of members. Electronic effects of aromatic amino substituents depend on the geometry of the CAr]NR2 fragment and it is determined by CAr]N distance, R]N]R valence angle, twist (torsion) 2,3 and bend angles,4,5 both referring to mutual orientation of nN and pAr orbitals, and also to the dihedral angle between the NR2 and the ring planes (the pyramidalization or tilt angle 6).The s substituent constants of various amino groups differ almost exclusively due to the extent of benzene ring–nitrogen atom resonance.7 Since the pKa values of aniline derivatives involve contribution of steric inhibition to solvation,8 basicity does not show properly the resonance between such groups and the aromatic ring. Instead, other data such as intensities of the UV 8,9 and IR bands,2,10 dipole moments,11,12 polarographic data 12,13 and exaltations of molar refraction 4,14 are more useful in prediction of the electron donor strength of amino groups.The sR8 constant of the p-dimethylamino group based on the fluorine chemical shift in the spectrum of N,N-dimethyl-p-fluoroaniline, which is sensitive to very small perturbations in the p-charge density at fluorine atom produced by the substituent,15 is equal to 20.52.15 That group in N,N,2,6-tetramethyl-4-fluoroaniline is much twisted out of the ring plane but there is still some resonance interaction between the two parts of the molecule.16 Both the 19F chemical shift in the NMR spectrum of N,N,2-trimethyl-4- fluoroaniline 15 and intensity of the UV spectrum of ethyl N,N,3-trimethyl-4-aminobenzoate 17 show that the ortho methyl group produces a 56% steric inhibition of the resonance in the Ar]N fragment and this results in a reduction of the sR8 constant of this p-dimethylamino group to 20.24.15 The chemical shift of the para carbon predicts well the electron donor strength of the amino groups.8,18 It shows N-phenylaziridine to be the least conjugated of all N-phenyl cyclic amines studied.4,8,19 It seems noteworthy that two ortho methyl groups decrease the N]methyl one-bond 13C]1H coupling constant in the spectra of N-methyl- and N,N-dimethyl-aniline by about 0.5 and 1.5 Hz, respectively.20 UV spectral data of p-aminonitrobenzenes,9 p-aminoazobenzenes, 21 p-aminobenzenes 4 and complexes of amino- † For a preliminary report of this work, see ref. 1. benzenes with 1,3,5-trinitrobenzene,14 IR spectral data of aminobenzenes,2,10 19F chemical shifts of fluoroanilines,15 acidities of p-aminobenzoic acids,13,22 rates of reduction of p-nitroanilines,13 polarographic halfwave potentials of pnitroanilines, 13 polarographic oxidation potentials of aminobenzenes 14 and dipole moments of aromatic amines 11,12 all show the following order of electron-donor strength of amino substituents: (CH2)4N > Et2N>Me2N > (CH2)5N > NHMe > NH2 > N(CH2)2.Alkane bridges between Cortho and Namino, e.g. in julolidine derivatives, cause the nitrogen atom to reveal stronger donor properties than that in the 1-pyrrolidino group.23 The published 1H, 13C, 15N and 17O NMR spectra of aromatic oximes 24–30 have been mainly concerned with the configuration and with transmission of substituent effects to the CH]] NOH group through the benzene ring.As found,30 there are poor or very poor correlations between the 17O NMR chemical shifts of the oximino oxygen and s, s1 and s1 substituent constants for substituted benzaldoximes. Electron acceptor properties of the CH]] NOH group (sR = 0.10) 7 preclude cross-conjugation to be important in p-aminobenzaldoximes. Thus, such compounds seem to be a very good model series to study the substituent effect. Since electronic properties of the less common amino groups are not known, the 1H and 13C NMR chemical shifts of p-aminobenzaldoximes are used in the present paper to evaluate their resonance substituent effects.The compounds under study have the formulae 1–20. Experimental Syntheses All melting points are uncorrected. The boiling points of all reaction products are expressed in the 8C/mmHg units. Most amines used in formylations are commercially available and known procedures were used to prepare others.31,32 Lilolidine‡ bp 139–141/10 (lit.bp 90–100/0.5,33 112–113/5 34) was prepared in overall 36% yield by reduction (LiAlH4, standard procedure) of 4-oxolilolidine which, in turn, was obtained by ‡ The IUPAC name for lilolidine is 1,2,5,6-tetrahydro-4H-pyrrolo[3,2,1- ij]quinoline. The name benzo[h,i]indolizidine has previously been used in the literature for lilolidine.33,3426 J. Chem. Soc., Perkin Trans. 2, 1998 cyclization–acylation of 1-(b-chloropropionyl)indoline.34 To synthesise 1-methylindoline, 1-formylindoline (bp 168–173/15, lit.bp 115–117/2 35) was first prepared in 94% yield from indoline and formic acid following the procedure used in formylation of N-methylaniline.36 It was then reduced (LiAlH4, standard procedure, 85% yield) to a product of bp 91–94/9 (lit. bps of 1-methylindoline are 68–73/1,37 120/12 38). Kairoline (1- methyl-1,2,3,4-tetrahydroquinoline), bp 102–104/4 (lit. bp 160/ 12,38 80–81/0.4 39) was obtained in 92% yield by reduction (LiAlH4, standard procedure) of 1-formyl-1,2,3,4-tetrahydroquinoline which, in turn, was prepared in 80% yield by reductive formylation of quinoline with formic acid according to ref. 40. 1-Methyl-2,3-benzohexamethyleneimine (bp 120–125/16, lit. bp 59–60/0.2,41 160/12 38) was obtained in a five-step synthesis (overall yield 34.5%) involving oximation of a-tetralone, 42 tosylation of a-tetralone oxime,43,44 Beckmann rearrangement of the O-tosyl derivative of a-tetralone oxime (for details see synthesis of its O-benzenesulfonyl derivative 44), methylation of homohydrocarbostyril 41,45 and reduction (LiAlH4) of 1-methyl-homohydrocarbostyril.41,45 Some aldehydes and ketones were commercial products.p-(N,N-Dimethylamino)acetophenone was a gift from Dr Tomasz Ba�k. Other aldehydes were obtained by the Vilsmeier– Haack46 or Duff 47 methods and were purified by vacuum distillation or crystallization from aqueous ethanol. Detailed synthetic procedures will be given in another paper.48 p-Aminobenzaldoxime was obtained in 47% yield by reduction of p-nitrobenzaldoxime, according to the procedure used in synthesis of p-aminoacetophenone oxime.49 The product was purified by crystallization from aqueous ethanol.Other aldoximes were obtained from their respective aldehydes by the standard procedure 50 and crystallized from aqueous ethanol. The yields were 26–91% (no attempts were made to improve the reaction efficiency).The synthetic procedure for ketoximes was slightly different. Thus, the mixture of the appropriate acetophenone (0.08 mmol), hydroxylamine hydrochloride (11.1 g, 0.16 mol), 96% aqueous ethanol (60 ml) and conc. hydrochloric acid (few drops) was refluxed for 1.5 h. The reaction mixture was then diluted with water (1 ml) and extracted with diethyl ether. Ketoximes were prepared in 51–69% yield by evaporation of solvent from the extract and recrystallization of the residue from 96% ethanol.Melting points (8C) of oximes: 1, 126–130 (127–128 51); 2, 96.5–96.9; 3, 147–148 (145–147 52); 4, N R5 R6 C R1 R2 R3 R4 R7 NOH 3 4 5 6 1 7 2 1 2 3 4 5 6 7 8 R1 H H H H H Me Me H R2 H H H H H H H Me R3 H H Me Me Et Me Me Me R4 H Me Me Et Et Me Me Me R5 H H H H H H H H R6 H H H H H H Me H R7 H H H H H H H H 9 10 11 H H H H H H (CH2)4 (CH2)5 (CH2)6 H H H H H H H H H 12 13 14 15 H H H H (CH2)2 (CH2)3 (CH2)3 (CH2)4 Me H Me Me H H H H H H H H H H H H 16 17 H H (CH2)2 (CH2)3 (CH2)3 (CH2)3 H H H H 18 19 H H H H H Me H Me H H H H Me Me 20 H H (CH2)5 H H Me 117–121; 5, 89–91 (93 53); 6, 105–109 (107–109 52); 7, 134–135; 8, 97–98 (97–98 52); 9, 190–192; 10, 161–163; 11, 110–114; 12, 97–99; 13, 147–151; 14, 87–88; 15, 107–109; 16, 104–106.5; 17, 126–128 (127–128 54); 18, 150–152 (153–154 55); 19, 218–222; 20, 163–116. NMR spectroscopy NMR spectra of the saturated solutions of oximes in [2H6] acetone were recorded on a JEOL JNM GSX-270 FT NMR spectrometer working at 270.17 and 67.94 MHz for 1H and 13C observation, respectively.TMS (internal reference) and [2H6] acetone (lock) were used both in 1H and 13C NMR experiments. Other conditions are: 1H: spectral width 3500 Hz, 32 K data points, digital resolution 0.21 Hz/point, pulse width 9.4 ms, flip angle 90 deg, number of scans 4, pulse delay 1 s, pulse sequence SGNON; 13C: spectral width 15 000 Hz (1H decoupled)/10 000 Hz (1H coupled), 32 K data points (1H decoupled)/64 K data points (1H coupled), digital resolution 0.92 Hz/point (1H decoupled) 0.32 Hz/point (1H coupled), pulse width 7.8 ms, flip angle 90 deg, number of scans 100–400 (1H decoupled)/ca. 9000 (1H coupled), pulse delay 4 s, pulse sequence/decoupling SGBCM/continuous bilevel SGNON (1H coupled). In order to analyse accurately the 1H NMR spectra, resolution enhancement was performed by combined exponential and trapezoidal windowing (T2 = 5% and T3 = 50%) and zero filling until the digital resolution was <0.05 Hz.The data matrix for 13C–1H HETCOR was as follows: 10 000 Hz and 1024 points for the 13C-axis and 1800–2000 Hz and 256 points for the 1H-axis. An average value of 1J(C,H) = 125 Hz was used for the correlation between coupled nuclei. Assignments of the signals in the aliphatic part of the spectra were possible from their homonuclear 1H–1H DFQ–COSY experiment. The signals of aromatic quaternary carbon atoms for compounds 9, 13, 15, 16 and 17 were assigned with the help of the 2D 13C–13C INADEQUATE spectra (solutions in [2H6] acetone).Data matrix: 10 000 Hz and 1024 points in the f1-axis and 20 000 Hz and 256 points in the f2-axis. The number of scans was 256 (8*32). An average value of 1J(C,C) = 36 Hz was used for the correlation between coupled nuclei. The 2D 13C–13C INADEQUATE correlation map of oxime 16 was recorded in saturated [2H6]DMSO solution at 30 8C with a Bruker Avance DRX500 spectrometer working at 125.76 MHz equipped with a 5 mm broadband direct detection probehead.The spectral width was 23 000 Hz (180 ppm), number of scans 160 and composite pulse decoupling (WALTZ-16) was used to decouple protons during the pulse sequence. The delay transmitting the correlation between coupled neighbouring 13Cnuclei was set to correspond the direct coupling constants of 1J(13C,13C) = 36 Hz. Results and discussion All oximes studied are the syn, i.e. E isomers. Although a coplanar arrangement of the ring and the CH]] NOH group in compounds carrying two methyls ortho to the oxime group may lead to serious steric repulsion, oxime 7 has also been assigned the E configuration.24 The 1H and 13C NMR chemical shifts for the oximes 1–20 are collected in Tables 1, 2, 4 and 5.As seen, the chemical shift of C7 changes in a narrow range (1.48 ppm) from 148.65 for 6 to 150.13 ppm for 16. The values of dC7 in the spectra of ketoximes 18–20 are >154 ppm. There is no linear relationship between the chemical shift of C7 and sR o substituent constants of the different amino groups for the compounds studied.It is known27,29,30 that the chemical shift of C7 in the spectra of p-substituted benzaldoximes depends mainly on the substituent inductive effect and its resonance is of reduced importance. Moreover, multiparameter correlations of dC7 with inductive and resonance substituent constants 56 and with semiempirical parameters that represent the paramagnetic interactionJ.Chem. Soc., Perkin Trans. 2, 1998 27 Table 1 13C chemical shifts of aromatic and a-methine carbons in the spectra of oximes 1–20 (d in ppm from TMS, in [2H6]acetone) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 16 c 17 18 19 20 C1 122.57 122.05 121.86 121.43 121.01 120.10 119.55 128.02 121.10 123.98 121.01 123.29 121.05 a 123.22 126.09 124.10 122.84 121.00 126.61 126.01 128.05 C2 128.74 128.73 128.52 128.70 128.86 138.07 139.07 130.05 128.67 128.47 128.87 122.55 128.67 127.54 126.34 126.74 125.68 d 126.34 127.58 127.42 127.33 C3 114.91 112.51 112.75 112.58 112.17 114.55 113.14 132.53 112.35 116.11 111.83 131.52 121.56 a 121.41 135.44 129.57 128.45 121.80 114.65 122.62 115.95 C4 150.73 152.06 152.24 150.81 149.35 151.91 151.21 154.92 146.45 153.59 150.59 155.42 147.62 148.46 154.64 b 151.01 144.70 149.97 151.89 153.14 C5 114.91 112.51 112.75 112.58 112.17 110.78 113.14 119.07 112.35 116.11 111.83 106.75 114.25 111.08 116.76 119.16 118.03 121.80 114.65 112.62 115.95 C6 128.74 128.73 128.52 128.70 128.86 128.75 139.07 126.05 128.67 128.47 128.87 128.41 126.43 127.07 128.79 120.90 119.83 e 126.34 127.58 127.42 127.33 C7 149.61 149.62 149.42 149.46 149.50 148.65 149.16 149.20 149.63 149.23 149.46 149.80 149.68 149.62 149.31 150.13 148.88 f 149.61 154.32 154.07 154.01 a Signals may be interchanged. b Due to limited solubility of this compound in acetone, the quality of its INADEQUATE spectra is poor and this signal is not seen in the spectrum. c In [2H6]DMSO at 125.758 MHz.d 1J(C2,H2) = 155.2 Hz. e 1J(C6,H6) = 158.14 Hz. f 1J(C7,H7) = 160.4 Hz. between the substituent and carbon atom57 are of better quality.27 The chemical shift of C1, i.e. the para carbon atom with respect to the amino substituent is the most appropriate from the point of view of electron donor strength of the amino substituent. Since sensitivity of 2D 13C–13C INADEQUATE experiment for compounds 9, 13, 15, 16 and 17 was good enough (see Fig. 1), the respective spectra were recorded to distinguish between the signals of different quaternary carbon atoms. The 13C chemical shifts for both aldoximes 1–17 and ketoximes 18–20 vary from 119.55 for 7 and 120.10 for 6 to 128.02 for 8 and 128.05 ppm for 20. Thus, the chemical shift dispersion of dC1, i.e. 8.50 ppm, is wider than that for dC7. Although, resonance and inductive effects of the para substituent on dC1 in the spectra of benzaldoximes are comparable by their magni- Table 2 13C chemical shifts of the side-chain carbons in the spectra of oximes 1–21 (d in ppm from TMS, in [2H6]acetone) 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 16 g 17 18 19 20 NC 29.83 40.27 46.93 a 37.41 b 44.85 40.20 40.22 44.10 48.12 50.17 49.81 56.13 a 35.41 b 42.17 51.53 a 38.81 b 57.18 a 43.01 b 47.20 c 55.08 e 46.02 c 54.13 e 50.49 — 40.38 50.12 NCC — — 11.44 12.86 — — — 26.01 26.31 28.12 28.67 27.79 22.81 30.12 23.54 d 28.80 f 22.40 d 27.85 f 28.40 — — 26.26 NC2C and NC3C — — — — — — — — 25.07 27.56 — 22.56 28.39 26.13 35.52 24.54 23.45 22.62 — — 25.07 CCH3 — — — — 20.63 22.27 18.68 — — — — — — — — — — 11.31 11.22 11.26 a NCH2.b NCH3. c NCH2CH2CH2. d NCH2CH2CH2. e NCH2CH2. f NCH2CH2. g In [2H6]DMSO at 125.758 MHz. tudes,27 only dCl values in the spectra of p-aminobenzaldoximes studied were found to be linearly dependent on sR o. The correlation obtained is: dC1 = 132.44 1 19.34sR o for compounds 1–3, 5 and 8–10 (correlation coefficient = 0.974, standard deviation = 2.45, standard error = 0.93).From its decreased value of sR o it is clear that the dimethylamino group in 8 is considerably twisted out of the ring plane. In order to determine the resonance effect of the less common amino groups, the above equation was used to calculate the respective sR o values. They are given in Table 3. The chemical shift of C4, i.e. the ipso carbon atom with respect to the amino substituent, for compounds 1–20 changes in a very wide range from 130.88 to 155.42 ppm (Dd = 24.54 ppm) but no simple dependence between dC4 and the type of substituent was found.It should be mentioned that multiparameter correlation between the chemical shift of C4, i.e. the ipso carbon, and the substituent constants in the NMR spectra of p-substituted benzaldoximes was observed.27 Although the contribution of the inductive effect is more than seven times that of the resonance effect, the paramagnetic interaction between the substituent and C4 is that which most contributes to dC4.57 Fig. 1 The 2D 13C–13C INADEQUATE correlation map of oxime 16 in saturated [2H6]DMSO solution at 30 8C28 J. Chem. Soc., Perkin Trans. 2, 1998 The chemical shift of H7 and that of the oxime proton, NOH, in the spectra of oximes studied changes in a similar narrow range, i.e. 0.49 and 0.43 ppm, respectively. Although, dC7 for both E and Z ring-substituted benzaldoximes correlates well with the s constants,29 no linear relationship between the shift of H7 and NOH, and sR o values was found for the compounds studied.The results obtained show that the resonance effect of p-amino groups is not transmitted strongly to the CH]] NOH group which is in disagreement with the X-ray studies on p-dimethylaminobenzaldoxime.58 Since the inductive substituent constants of amino groups are scarce and those that are available differ only slightly from each other, no correlation between the shift of C7 and s1 values can be obtained (such a procedure was used for some other p-substituted benzaldoximes 28).X-Ray determination 58 shows that the molecule p-(N,Ndimethylamino) benzaldoxime is planar with angles /CMeNCMe9 and /CMeNCAr equal to 115.58, and 121.78 (120.78), respectively. The CAr]N distance (1.380 Å) indicates this bond to have significant double bond character. The angle /CMeNCMe shows the amine nitrogen atom to have sp2 hybridization. Other bond lengths and valence angles confirm also that there is an electron Table 3 sR o substituent constants of amino groups R NH2 NHMe NMe2 N(Me)Et NEt2 1-NMe2; 3-Me 1-NMe2; 3,5-Me2 1-NMe2; 2-Me N(CH2)4 N(CH2)5 N(CH2)6 1-N(Me)[2-(CH2)2] 1-NH[2-(CH2)3] 1-N(Me)[2-(CH2)3] 1-N(Me)[2-(CH2)4] 1-N[2-(CH2)2][6-(CH2)3] 1-N{2,6-[(CH2)3]2} sR o 20.48 a 20.52 a 20.53 a 20.57 b 20.57 a 20.64 b 20.67 b 20.24 a 20.63 a 20.47 a 20.59 b 20.47 b — 20.48 b 20.33 b 20.43 b 20.59 b a Literature values.2,10,15 b Calculated from equation sR o = (dC1 2 132.44)/ 19.34.Table 4 1H chemical shifts of aromatic, a-methine and oximino protons in the spectra of oximes 1–20 (d in ppm from TMS, in [2H6] acetone) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 H2 7.32 7.37 7.44 7.42 7.40 — — 7.40 7.41 7.44 7.40 7.34 7.11 7.16 7.35 6.96 6.95 7.42 7.54 7.54 H3 6.66 6.59 6.71 6.69 6.68 6.54 6.44 — 6.56 6.92 6.71 — — — — — — 6.64 6.71 6.90 H5 6.66 6.59 6.71 6.69 6.68 6.56 6.44 7.01 6.56 6.92 6.71 6.41 6.45 6.53 6.88 — — 6.64 6.71 6.90 H6 7.32 7.37 7.44 7.42 7.40 7.49 — 7.35 7.41 7.44 7.40 7.21 7.13 7.22 7.33 7.16 6.95 7.42 7.54 7.54 H7 7.96 7.96 7.99 7.99 7.96 8.23 8.34 8.02 7.97 8.00 7.96 7.98 7.90 7.93 8.03 7.93 7.85 — — — NOH 9.67 9.63 9.71 9.69 9.62 9.73 9.83 9.98 9.63 9.80 9.63 9.73 9.56 9.65 9.95 9.63 9.55 ~9.7 a 9.75 9.87 a Very weak signal.transfer from the amino nitrogen to oximino oxygen. However, the NMR results presented in Tables 1 and 4 show this is not the case for p-(N,N-dimethylamino)benzaldoxime and the other oximes studied.Moreover, those data indicate that hybridization of the amino nitrogen atom in the compounds studied is more sp3 than sp2-like. The conformation of indoline and of its homologs is not known. Absorption bands in the spectra of N-alkylindolines have significantly reduced intensity, and are red shifted, as compared to the spectra of N,N-dialkylanilines.59 Both differences in hybridization of nitrogen atoms in those compounds and conformational equilibria of the five-membered ring in indoline can account for this behaviour but no definitive explanation of extent of the benzene ring–nitrogen resonance in that compound was given.59 On the other hand, molecular models show that hybridization of the N atom in indolines can be both of sp2 and sp3 type, and the five-membered ring can be both planar and puckered.Moreover, deformation of the benzene ring in anilines carrying short CAr]N bridges can also occur.The derived sR o values show piperidino to be the most powerful donor among the studied amino substituents. The propane bridge between the amino N atom and ortho position was found to enable the resonance to be more effective than that in the systems that contain shorter and longer bridges. Finally, in disagreement with earlier work,23 the sR o values show that the amino nitrogen atom in julolidine is a weaker electron donor than that in pyrrolidine.It is noteworthy that the observed order of the substituents resembles, in general, that based on values of EHOMO and positions of the CT bands in the spectra of complexes of aminobenzenes with 1,3,5-trinitrobenzene,14 UV spectral data of p-aminoazobenzenes,21 dipole moments of aromatic amines,11,12 polarographic oxidation potentials of aminobenzenes 14 and polarographic halfwave potentials of p-nitroanilines.13 Other spectral studies, now in progress, are expected to show the accuracy of estimation of the s values obtained.Acknowledgements Financial support provided by the (Polish) Committee for Table 5 1H chemical shifts of the side-chain protons in the spectra of oximes 1–20 (d in ppm from TMS, in [2H6]acetone) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 NH 4.90 5.24 — — — — — — — — — — 5.30 — — — — ~4.4 h — — NCH — 2.83 2.96 3.43 a 2.93 b 3.41 2.95 2.94 2.69 3.28 3.23 3.51 3.31 a 2.74 b 3.29 3.23 a 2.87 b 2.92 a 2.84 b 2.98 d 3.28 f 3.18 — 2.96 3.20 NC2H — — — 1.09 1.15 — — — 2.00 1.64 c 1.78 2.89 1.86 1.91 1.71 2.04 e 2.87 g 1.92 — — 1.63 d NC3H and NC4H — — — — — — — — — 1.64 c 1.53 — 2.71 2.69 1.58 2.74 2.62 2.69 — — 1.63 d CCH3 — — — — — 2.37 2.35 2.29 — — — — — — — — — 2.14 2.15 2.16 a NCH2.b NCH3. c Centre of the multiplet. d NCH2CH2CH2. e NCH2- CH2CH2. f NCH2CH2. g NCH2CH2. h Very weak signal.J. Chem. Soc., Perkin Trans. 2, 1998 29 Scientific Research [KBN, grant 3 T09A 147 09] (R.G.) and the Academy of Finland (E. K.) is gratefully acknowledged. References 1 Preliminary communication: 13th IUPAC Conference on Physical Organic Chemistry, (Inchon, Korea), 1996, p. 74. 2 A. R. Katritzky and R. D. Topsom, Angew. Chem., Int. Ed. Engl., 1970, 9, 87. 3 M. D. Rozeboom, K. N. Houk, S. Searles and S. E. Seyedrezai, J. Am. Chem. Soc., 1982, 104, 3448. 4 A. T. Bottini and Ch. P. Nash, J. Am. Chem. Soc., 1962, 84, 734. 5 C. Cauletti, G.Cerichelli, F. Grandinetti, L. Luchetti and M. Speranza, J. Phys. Chem., 1988, 92, 2751. 6 A. Hastie, D. G. Lister, R. L. McNeil and J. K. Tyler, J. Chem. Soc., Chem. Commun., 1970, 108. 7 O. Exner, Critical Compilation of Substituent Constants in Correlation Analysis in Chemistry: Recent Advances, ed. N. B. Chapman and J. Shorter, Plenum Press, New York, 1978, p. 439. 8 C. P. Nash and G. E. Maciel, J. Phys. Chem., 1964, 68, 832. 9 M. J. Kamlet, J. L. Abboud and R. W. Taft, J.Am. Chem. Soc., 1977, 99, 6027. 10 R. T. C. Brownlee, R. E. J. Hutchinson, A. R. Katritzky, T. T. Tidwell and R. D. Topsom, J. Am. Chem. Soc., 1968, 90, 1757. 11 K. S. F. Beach, J. D. Hepworth, J. Sawyer, G. Hallas, R. Marsden and M. M. Mitchel, J. Chem. Soc., Perkin Trans. 2, 1984, 217. 12 D. Mazet, W. D. Weringa and H. Lumbroso, C. R. Seances Acad. Sci., 1970, 270C, 1537. 13 W. D. Weringa and M. J. Janssen, Recl. Trav. Chim. Pays-Bas, 1968, 87, 1372. 14 F. Effenberger, P.Fischer, W. W. Schoeller and W.-D. Stohrer, Tetrahedron, 1978, 34, 2409. 15 R. W. Taft, E. Price, I. R. Fox, I. C. Lewis, K. K. Andersen and G. T. Davis, J. Am. Chem. Soc., 1963, 85, 3146. 16 M. J. S. Dewar and Y. Takeuchi, J. Am. Chem. Soc., 1967, 89, 390. 17 R. W. Taft and H. D. Evans, J. Chem. Phys., 1957, 27, 1427. 18 P. C. Lauterbur, J. Chem. Phys., 1963, 38, 1415. 19 R. G. Pews, J. Am. Chem. Soc., 1967, 89, 5605. 20 C. H. Yoder, B. A. Kaduk and R. E. Hess, Tetrahedron Lett., 1970, 3711. 21 G. Hallas, R. Marsden, J. D. Hepworth and D. Mason, J. Chem. Soc., Perkin Trans. 2, 1984, 149. 22 B. van de Graaf, H. J. Hoefnagel and B. M. Wepster, J. Org. Chem., 1981, 46, 653. 23 R. W. Castelino and G. Hallas, J. Chem. Soc., B, 1971, 793. 24 M. Christl, J. P. Warren, B. L. Hawkins and J. D. Roberts, J. Am. Chem. Soc., 1973, 95, 4392. 25 C. P. J. Vuik, M. ul Hassan and C. E. Holloway, J. Chem. Soc., Perkin Trans. 2, 1979, 1214. 26 R. E. Botto, P. W. Westerman and J.D. Roberts, Org. Magn. Reson., 1978, 11, 510. 27 G. V. Rutkovskii and V. P. Zmeikov, Zh. Org. Khim., 1987, 23, 142. 28 M. S. Gordon, S. A. Sojka and J. G. Krause, J. Org. Chem., 1984, 49, 97. 29 I. Pejkovic�-Tadic�, M. Hranislavljevic�-Jakovljevic�, S. Nes¡ic�, C. Pascual and W. Simon, Helv. Chim. Acta, 1965, 48, 1157. 30 G. Cerioni and A. Plumitallo, Magn. Reson. Chem., 1993, 31, 320. 31 C. F. van Duin, Rec. Trav. Chim., 1932, 51, 878. 32 L. C. Craig and R. M. Hixon, J.Am. Chem. Soc., 1930, 52, 804. 33 B. D. Astill and V. Boekelheide, J. Org. Chem., 1958, 23, 316. 34 G. Hallas and D. C. Taylor, J. Chem. Soc., 1964, 1518. 35 J. T. Shaw and F. T. Tyson, J. Am. Chem. Soc., 1956, 78, 2538. 36 L. F. Fieser and J. E. Jones, Org. Synth., 1955, Coll. Vol. 3, 590. 37 G. W. Gribble and Ch. F. Nutaitis, Synthesis, 1987, 709. 38 H. Ahlbrecht, E. O. Düber, J. Epsztajn and R. M. K. Marcinkowski, Tetrahedron, 1984, 40, 1157. 39 F. Minisci, R. Galli and M. Perchinunno, Org. Prep. Proced. Int., 1969, 1, 77. 40 A. N. Kost and L. G. Yudin, Zh. Obshch. Khim., 1955, 25, 1947. 41 L. A. Paquette, D. E. Kuhla and J. H. Barrett, J. Org. Chem., 1969, 34, 2879. 42 R. Huisgen, W. Rapp, I. Ugi, H. Walz and E. Mergenthaler, Liebigs Ann. Chem., 1954, 586, 1. 43 G. Schroeter, Ber., 1930, 63, 1308. 44 R. Huisgen, I. Ugi, H. Brade and E. Rauenbusch, Liebigs Ann. Chem., 1954, 586, 30. 45 G. Baddeley, J. Chadwick and H. T. Taylor, J. Chem. Soc., 1956, 451. 46 E. Campaigne and W. L. Archer, Org. Synth., 1963, Coll. Vol. 4, 331. 47 C. Berres and W. Müller, Swiss Patent 430 677/1967; C. Berres and W. Müeller, French Patent 1, 377 226/1964; Chem. Abstr., 1965, 62, P7691f. 48 R. Gawinecki, S. Andrzejak and A. Pucha�a, in preparation. 49 T. Satoh, N. Mitsuo, M. Nishiki, Y. Inoue and Y. Ooi, Chem. Pharm. Bull., 1981, 29, 1443. 50 Vogel’s Textbook of Practical Organic Chemistry, 4th edn., Longman, London, 1978, p. 810. 51 S. Witek, J. Bielawski and A. Bielawska, Polish J. Chem., 1981, 55, 2589. 52 D. J. Gale and J. F. K. Wilshire, Aust. J. Chem., 1975, 28, 2447. 53 J. C. Duff, J. Chem. Soc., 1945, 276. 54 P. A. S. Smith and T.-Y. Yu, J. Org. Chem., 1952, 17, 1281. 55 J. Sykulski, K. Lesz and U. Tomera, Polish J. Chem., 1979, 53, 2013. 56 C. G. Swain and E. C. Lupton, J. Am. Chem. Soc., 1968, 90, 4328. 57 N. B. Smith and T. W. Proulx, Org. Magn. Reson., 1976, 8, 567. 58 F. Bachechi and L. Zambonelli, Acta Crystallogr., Sect. B, 1972, 28, 2489. 59 W. Rettig, K. Rotkiewicz and W. Rubaszewska, Spectrochim. Acta, Part A, 1984, 40, 241. Paper 7/05668K Received 4th August 1997 Accepted 25th Se

ISSN:1472-779X    

DOI:10.1039/a705668k

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1998 31 Molecular dynamics analysis of the dipole moment and conformational properties of 2-(acetyloxy)ethyl-2-(2-naphthyl)acetate Patricia Saez-Torres,a Maria P. Tarazona,b Enrique Saiz,b Evaristo Riande *,a and Julio Guzmán a a Instituto de Ciencia y Tecnología de Polímeros (CSIC), 28006-Madrid, Spain b Departamento de Química Física, Universidad de Alcalá, 28871 Alcalá de Henares, Spain A sample of 2-(acetyloxy)ethyl-2-(2-naphthyl)acetate (ANA), model compound of the side group of poly(2-{[2-(2-naphthyl)-acetyl]oxy}ethyl acrylate) (PNAEA), has been synthesized and its mean-squared dipole moment measured in dioxane solutions at 30 8C, providing an experimental result of ·Ï2Ò = 5.28 D2.Molecular dynamics simulations, performed with the Amber force field, the DL–POLY package and the charge distribution provided by MOPAC, gives a theoretical value ·Ï2Ò = 5.69 D2, in good agreement with experience. The analysis of the MD trajectories indicates that the preferred orientations for the Car–CH2 bond are those in which the plane defined by the pair of bonds Car–CH2–C* is roughly perpendicular to the aromatic group (i.e.� ª ±908) while trans orientation of the CH2–C* bond is strongly disfavoured (by ca. 1.5 kcal mol21) versus gauche. In the O]CH2]CH2]O segment, the O]CH2 bonds show a strong preference for trans while CH2]CH2 prefers gauche and does not require any kind of adjustment a posteriori to account for the so called ‘gauche effect’.Thus, the first-order conformational energies representing the stability of gauche versus trans are EÛ ª 1.2, EÛ1 ª 20.9 kcal mol21 respectively for O]CH2 and CH2]CH2. Second-order interactions E. and E„ respectively controlling the stability of opposite (i.e. g±g+ – ) and identical (i.e. g±g±) combinations of gauche states over the pair of bonds O]CH2]CH2 are both negative (ca. E. ª 21, E„ ª 20.6 kcal mol21). Values of the mean-squared dipole moment computed by adding contributions produced by two ester groups with this set of conformational energies are in very good agreement with the results obtained by the actual MD simulations.Introduction Dielectric relaxation spectra of glass-forming liquids present in the frequency domain, and just above their glass-liquid transition temperatures, a prominent absorption called a relaxation, followed by a fast and relatively weak process called b relaxation. An increase in temperature shifts both relaxations to higher frequencies, and as a consequence of the high activation energy of the a relaxation, a temperature can be reached at which both relaxations merge together forming the ab relaxation. 1 The normalized ab relaxation response function in the time domain exhibits at short times a relaxation process followed by a much more extended time regime in which the relaxation displays a Kohlrausch–Williams–Watts (KWW) stretched exponential decay given by 2,3 eqn.(1) where 0 < b < 1 and t(T) g(t) = expF2S t t(T)DbG (1) is the mean relaxation time whose value is given by the area under the g(t) curve along the positive time axis. The temperature dependence of t in the vicinity of the glass transition temperature is described by the Vogel–Fulcher–Tammann– Hesse (VTHF) eqn. (2) 4–6 in which m and T• are believed to be t = A expF m T 2 T• G (2) related, respectively, to the fractional free volume and to the Kauzmann temperature,7,8 i.e.the temperature at which the conformational entropy vanishes. Transformation of the decay function from the time to the frequency domain allows the evaluation of the strength of the a relaxation process defined as Dea = e0 2 e•, where e0 and e• represent, respectively, the relative permittivity at 0 and • frequencies. In most cases the relaxation strength of the b process is negligible compared with that of the a relaxation and the value of Dea can directly be obtained from the mean-square dipole moment ·m2Ò of the compounds by using Onsager 9 type expressions such as the Fröhlich equation.10,11 The value of ·m2Ò can be computed by either assigning dipoles to bonds or groups of bonds and averaging the square of the vector sum for all the available conformations or from the trajectories of the dipole moment in the conformational space.12–16 In the latter case this gives eqn. (3), where m(ti) is the ·m2Ò = ·Ï(t)?Ï(t)Ò = 1 n o n i = 1 m2(ti) (3) dipole moment of the molecule at time ti in the trajectory, and n is large enough as to make certain that the molecule visits all the conformational space.This latter approach was used in the evaluation of the meansquare dipole moment of 2-(acetyloxy)ethyl-2-(2-naphthyl)- acetate (ANA), model compound of the side group of poly(2- {[2-(2-naphthyl)-acetyl]oxy}ethyl acrylate) (PNAEA), carried out in this work. Hydrolysis of both the model compound and the polymer17,18 gives the naphthalene acetic acid, a compound widely used as a plant growth regulator. An important issue in the study of the conformational characteristics of model compounds containing the OCH2]CH2O moiety in their structure, as occurs in ANA, is that the rotational states population about CH2]CH2 bonds present a gauche effect, i.e.the experimental value of the energy of the gauche states with respect to the alternative trans states is significantly lower than that obtained by using semiempirical potential functions or semiempirical quantum mechanics methods.19–21 The gauche effect expressed in terms of the difference between the experimental and calculated values of the energy of gauche states is ca. 1 kcal mol21 for CH2]CH2 bonds in 1,2-dimethoxyethane (CH3]OCH2]CH2] OCH3).20 Therefore, it is tempting to investigate how the two ester groups flanking the CH2]CH2 bond in 2-(acetyloxy)ethyl- 2-(2-naphthyl)acetate (ANA) may affect the gauche population about this bond.This study forms part of a more general study whose object-32 J. Chem. Soc., Perkin Trans. 2, 1998 ive is to investigate how the mobility of the side groups affects the mean-square dipole moment of acrylic polymers. This study will be extended in a further work to the simulation of the dielectric a relaxations of both the model compound ANA and the polymer PNAEA. Molecular dynamics calculations The structure of the ANA molecule is shown in Fig. 1 where all the rotating bonds on the molecule have been drawn in their planar trans conformation for which the value of the rotational angles f were set to be 1808. Dipole moments of the molecule were calculated by means of partial charges assigned to each atom. They were computed with the MOPAC program and the AM1 procedure.22 Fig. 2 indicates the charges assigned to each atom. Molecular dynamic (MD) simulations were performed at T = 300 K with the DL–POLY package 23 employing the Amber force field.24–26 A time step d = 1 fs (i.e. 10215 s) was used for the integration cycle that was repeated 107 times to cover a total time span of 10 ns (i.e. 1028 s) for the whole simulation. Conformations obtained after each 1000 steps were recorded, thus providing a total of 104 conformations for posterior analysis. The geometry of the molecule was first optimized with respect to all bond lengths, bond angles and rotations by minimizing its conformational energy. This optimized geometry was then used as starting point for the MD simulation whose first assignment was to warm up the molecule from 0 K to the working temperature with increments of 20 K and allowing a relaxation period of 500 fs at each intermediate temperature.Once the sample was stabilized at 300 K, the collection of data started. No scaling of the 1–4 interactions was performed. The Coulombic term of the potential energy was computed as the sum of pairwise interactions according to eqn.(4), where Eij indicates the Eij = 332 qiqj erij (4) interaction between atoms i and j that are parated a distance rij and whose partial charges are qi and qj. The numerical constant 332 renders Eij in kcal mol21 when charges are given in electron units and distances in Å. A value e = 4 was used for the relative permittivity of the system. The temperature of the sample was controlled by scaling the atomic velocities after each integration cycle with a factor l defined in eqn.(5), where T0 represents the working temperli = F1 1 S d 2dDST0 Ti 2 1DG��� (5) Fig. 1 A schematic sketch of the 2-(acetyloxy)ethyl-2-(2-naphthyl)- acetate (ANA) molecule shown in its planar all trans conformation for which the rotatable bonds on the acyclic residue were set to be f = 1808. Dipole moments of the two ester groups are represented by arrows pointing from negative to positive centres of charges. C C* O O* C H H C O H H H H C* C O* H H H t t m m 2 1 Fig. 2 Partial charges, in electron units, assigned to each atom of the 2-(acetyloxy)ethyl-2-(2-naphthyl)acetate (ANA) molecule with the MOPAC program and the AM1 procedure –0.082 –0.082 –0.072 –0.025 –0.026 –0.072 –0.065 –0.041 –0.075 –0.071 0.083 0.083 0.084 0.084 0.084 0.085 0.084 0.050 0.246 –0.326 –0.292 0.056 0.056 0.066 0.078 0.078 –0.292 0.237 0.066 0.078 0.078 0.013 –0.327 0.051 0.051 0.051 ature (i.e. T0 = 300 K in the present work) while Ti indicates the temperature evaluated with the atomic velocities after step i.The damping factor was set to d = 1000 fs. This is a rather high value for a damping factor, but it gently forces Ti towards the target temperature, thus ensuring that the averaged value for the whole simulation matches T0, and, at the same time, allows relatively high thermal fluctuations that facilitate the passage over the rotational barriers. Fig. 3 represents the time evolution of the dipole moment m during the first ns of the simulation.The instantaneous values oscillate between a maximum of ca. 4 D and a minimum of roughly 0, while the averaged results for the whole simulation are ·|Ï|Ò = 2.25 D and ·m2Ò = 5.69 D2, this latter value being in good agreement with the experimental result, 5.28 D2. The probability distribution of each rotating bond on the acyclic fragment, taken as independent of its neighbours, was computed by counting how many times along the simulation the studied angle reached a given value with a tolerance of ±50, (for instance, the results indicated for f = 40 represent the fraction of conformations in which 35 < f < 45).The results are depicted in Figs. 4–6, while some more quantitative information is presented in Table 1 that summarizes the averaged positions, probabilities and relative energies of the rotational isomers for all these angles, excluding the O]C* which are always in trans and the final C*]CH3 which is irrelevant. As Fig. 4 indicates, the preferred states for bond Car]CH2 are those in which the plane defined by the pair of bonds Car]CH2]C* is roughly perpendicular to the aromatic group, i.e.f ª 90, 2708. The staggered orientations of this bond produce strong repulsions of the H atoms on the ring with the H of the CH2 and the carbonyl O*. On the other hand, the trans orientation of the CH2]C* bond is strongly disfavoured versus Fig. 3 Dipole moment of the sample, m, as a function of time during first ns of the simulation.Averaged results are: ·|Ï|Ò = 2.25 D, ·m2Ò = 5.69 D 2. Table 1 Averaged positions, probabilities and relative energies of the rotational isomers for rotatable bonds of the acyclic residue of the ANA molecule Bond Car]CH2 CH2]C* O]CH2 CH2]CH2 State perpen. perpen. trans gauche gauche2 trans gauche gauche2 trans gauche gauche2 ·fÒ/8 86.6 272.6 178.6 294.1 66.5 179.8 285.3 76.4 179.4 295.0 64.5 Probability at 300 K 0.500 0.500 0.040 0.480 0.480 0.618 0.191 0.191 0.070 0.465 0.465 E (relative)/kcal mol21 0.00 0.00 0.00 21.48 21.48 0.00 0.70 0.70 0.00 21.13 21.13J.Chem. Soc., Perkin Trans. 2, 1998 33 gauche (i.e. gauche is ca. 1.5 kcal mol21 below trans according to Table 1). The reason for this difference lies in the interaction between the carbonyl O* and the Car attached to the acyclic residue that are separated ca. 2.6 and 3.0 Å, respectively, for trans and gauche orientations. Fig. 5 shows the results obtained with the two equivalent C*]O bonds of the molecule for which trans is the only allowed orientation.Thus, even if they fluctuate along the MD simulation within a ±308 range, they never cross the barrier to the alternative cis conformation. The analysis of the O]CH2]CH2]O segment, whose results are summarized in Fig. 6, is more interesting, since the relative orientation of these three bonds is the main feature that controls the value of the dipole moment for the whole molecule (see below).Our results indicate that trans is the preferred orientation for the O]CH2 bonds, while CH2]CH2 prefers gauche, as was found before in the analysis of similar compounds employing a different procedure.27 It is interesting to notice that in the present calculation the preference of bonds OCH2]CH2O for gauche states is not introduced a posteriori by means of a socalled gauche effect that has been invoked many times for the analysis of polyoxides.21 The reason is that the Amber force Fig. 4 Probability distributions for rotations over Car]CH2 (—) and CH2]C* (– – –) bonds. Points indicate the actual probabilities computed along the MD simulation while the lines represent least squares fittings and are drawn to show up the shape of the distributions. Fig. 5 Probability distributions for rotations over C*]O bonds. See legend for Fig. 4. field contains a better parametrization of this kind of bond, mainly as regards the intrinsic rotational barrier. The distributions of a priori probabilities for correlated pairs of bonds are represented in Fig. 7 for the pair of bonds Car]CH2]C* and in Fig. 8 for O]CH2]CH2 and its mirror image CH2]CH2]O (with interchange of x and y axes for this last pair). Fig. 7 shows four roughly equivalent maxima, each one representing a probability of ca. 24%, at the combination of orientations f1 = ±908 for bond Car]CH2, f2 = ±608 for CH2]C*. However, the orientations placing CH2]C* bond trans only represent a probability of ca. 4%. Thus, the information provided when this pair of bonds is taken to be correlated is about the same as that which could be obtained by the analysis of the same bonds taken independently. The probabilities for the pair of bonds O]CH2]CH2, represented in Fig. 8, show two main maxima, each representing a Fig. 6 Probability distributions for rotations over O]CH2 (—) and CH2]CH2 (– – –) bonds. See legend for Fig. 4. Fig. 7 Distribution of the a priori probabilities for the Car]CH2]C* pair of bonds 0 0 360 300 240 180 120 60 60 120 180 240 300 360 Rotational angle(C –CH )/degrees2 ar 2 *Rotational angle(CH –C )/degrees Fig. 8 Distribution of the a priori probabilities for the O]CH2]CH2 pair of bonds 0 0 360 300 240 180 120 60 60 120 180 240 300 360 Rotational angle(O–CH )/degrees2 2 2Rotational angle(CH –CH )/degrees34 J. Chem. Soc., Perkin Trans. 2, 1998 Table 2 Probabilities of occurrence (computed at T = 300 K), energies (in kcal mol21) relative to the all trans conformation, first- and second-order contributions to the conformational energies for the allowed conformations of the O]CH2]CH2]O segment of the ANA molecule.The last two columns were computed with the following values of conformational energies: Es = 1.20; Es1 = 20.87; Ew = 20.13; Eg = 20.57 kcal mol21 From MD From contrib. States a ttt (1) ttg (4) tgt (2) tgg (4) tgg2 (4) gtg (2) gtg2 (2) ggg (2) ggg2 (4) g2g (2) Prob. 0.0420 0.0065 0.1358 0.0838 0.0593 0.0007 0.0009 0.0135 0.0129 0.0031 Erel 0.00 1.11 20.70 20.41 20.21 2.44 2.29 0.68 0.70 1.55 Contributions 0 Es Es1 Es 1 Es1 1 Eg Es 1 Es1 1 Ew 2Es 2Es 2Es 1 Es1 1 2Eg 2Es 1 Es1 1 Ew 1 Eg 2Es 1 Es1 1 2Ew Prob. 0.0469 0.0063 0.2018 0.0701 0.0335 0.0008 0.0008 0.02117 0.0057 Erel 0.00 1.20 20.87 20.24 0.20 2.40 2.40 0.39 0.83 1.27 a Symmetrically equivalent conformations listed below are not included in the body of the table for simplicity. The number of conformations represented by each one of those included is given in brackets.ttg = ttg2 = gtt = g2tt, tgt = tg2t, tgg = tg2g2 = ggt = g2g2t, tgg2 = tg2g = g2t = g2gt, gtg = g2tg2, gtg2 = g2tg, ggg = g2g2g2, ggg2 = g2g = g2g2 = g2g2g2, g2g = g2g2. probability of ca. 0.25, for the tg and tg2 conformations [i.e. f(O]CH2) = 1808, f(CH2]CH2) ª ±608], which are a combination of the preferred states for both kind of bonds. The next pair of relevant maxima are gg and g2g2, each having a probability of ca. 0.14. The probability for each one of the combinations gg2 and g2g is ca. 0.06. The conformations placing the second bond trans have low probabilities, thus tt amounts to ca. 0.06 while each of the remaining gt and g2t possibilities that place both bonds in their disfavoured states have probabilities of ca. 0.02. A more detailed analysis of the conformations adopted by the O]CH2]CH2]O segment is summarized in Table 2. The possible conformations adopted by these three bonds are collected in the first column, although only one of each possible series of symmetrically equivalent orientations is represented in order to simplify the Table.The multiplicity, i.e. the number of symmetrical conformations for each row, is given in brackets. The probabilities computed at T = 300 K for each conformation are given in the second column. They were obtained by counting how many conformations along the MD simulation placed simultaneously the three rotational angles in the desired orientations and averaging the results for each series of equivalent conformations.These probabilities were used to calculate conformational energies as Ei = 2RT ln pi where i represents each of the studied conformations. Values of these energies, normalized to give E = 0 for ttt, are collected in the third column. The total energy of any given conformation is customarily separated into a sum of first- and second-order interactions. First-order interactions are those produced among pairs of atoms whose separation depends on just one rotational angle.In the present case, first-order interactions can be represented by two parameters, namely Es that represents the energy of gauche orientations relative to trans for the O]CH2 bonds, and Es1 that has the same meaning for the CH2]CH2 bond. Secondorder interactions are those produced by pairs of atoms whose separation depend on two consecutive rotations and we represent by Ew the interactions produced in pairs of different gauche states, i.e.gg2 or g2g conformations and by Eg the interactions raised in gg or g2g2 orientations. The fourth column in Table 2 indicates the combination of these parameters required to represent each one of the conformations. The values of these contributions are assigned in such a way that they produce the best possible fitting of the actual conformational energies indicated in the previous column, and in this sense, the combination: Es = 1.20, Es1 = 20.87, Ew = 20.13, Eg = 20.57 kcal mol21 provides very good results, as shown in Table 2 whose last two columns were computed with these values of the conformational energies. The values of the first-order parameters seem to be reasonable since they indicate that the O]CH2 bonds prefer trans orientations while gauche are favoured in the case of CH2]CH2 bonds.However, the results obtained for the second-order interactions Ew and Eg are quite amazing because they are both negative, whereas in most polymeric systems Ew is large and positive (in many cases Ew is taken to be infinity thus forbidding the gg2 and g2g conformations) while Eg is usually set equal to zero (i.e.second-order interactions produced in gg and g2g2 conformations are supposed to be rather weak). Nevertheless, some more recent analysis based on ab initio calculations provide negative values for these second-order energies.28 The reason for the present results is that the most important second-order interactions for the O]CH2]CH2 pair of bonds are produced by the O* on the first ester group and the O atom on the second CH2O group. Taking into account that the C*]O bond is always close to trans (see Fig. 5), the O* atom lies in the plane defined by the C*]O]CH2 pair of bonds and this fact makes the differences from g2, g2g to g, g2g2 conformations to be much smaller than in polymers where the groups involved are for instance two CH2. Thus, in the g2 orientation, the O* of the first ester and the O of the second one would be placed at a distance of ca. 1.7 Å if the molecule were in the undistorted geometry (i.e. all the rotations in the perfectly staggered positions and all bond lengths and bond angles in their main value). Such a short distance between these two atoms would produce a strong repulsion that should render a large and positive value of Ew. However, it is quite easy to produce a noticeable increase in that distance through small adjustments of rotations and bond angles so that the interaction becomes attractive.On the other hand, the g conformation, in the undistorted geometry, places the first O* atom at a distance of 1.8 Å from one of the H atoms of the second CH2 group, so that the situation is not too different from the previous case, in particular, this distance could also be increased by small adjustments producing a negative value of Eg. The most important contributions to the total dipole moment of this molecule come from the two ester groups, and since the orientation of the O]CH2]CH2]O segment controls the relative orientations of these two contributions and therefore their possible reinforcement or cancellation, it seems reasonable to presume that the dipole moment of the complete molecule will depend almost exclusively on the conformational characteristics of this segment.Some calculations were performed by adding the dipole moment of two ester groups (represented as m1 and m2 in Fig. 1). In these molecules 29 m1 = m2 = 1.75 D, and each dipole forms an angle t = 1218 with the respective C]C* bond.30 The results obtained with different values of the conformational energies and location of the rotational isomers are summarized in Table 3. The first line in this Table contains the ‘main set’ of parameters, i.e. the values of energies indicated above and rotational isomers located at f(t, g±) = 1808, ±458 for C]O bonds and 1808, ±558 for CH2]CH2, as suggested by Fig. 7J. Chem. Soc., Perkin Trans. 2, 1998 35 and Table 1 which indicate that trans conformations are located at 1808 while gauche are displaced by ca. D(C]O) ª 158, D(CH2] CH2) ª 58 from the perfectly staggered positions. This main set of parameters gives ·m2Ò = 5.73 D2, in excellent agreement with the result obtained from the actual MD simulation. The remaining lines on Table 3 indicate the variation produced in ·m2 Ò by small modifications to the conformational parameters.It is interesting to note that the total dipole moment is more sensitive to the second-order energies Ew and Eg than to the firstorder parameters Es and Es1. The reason is that tgg2, g2g and all their symmetrically equivalent conformations, which are controlled by Ew, produce the highest total dipole moment so that ·m2Ò increases with increasing value of Ew. On the contrary, ggg and g2g2g2, which are controlled by Eg, are, together with ttt, the orientations of lowest polarity so that ·m2Ò decreases when the value of g increases.It is also noteworthy that the dipole moment of the molecule is rather sensitive to the location of gauche states, especially for the CH2]CH2 bond, with ·m2Ò increasing when both states approach their staggered positions (i.e. with decreasing values of the D displacement), where m1 and m2 contributions are roughly parallel for conformations such as tgg2. Discussion The results presented in this analysis, as well as many others reported elsewhere, show that MD simulations give a good account of the polarity of molecules.The use of MD simulations in the determination of dipole moments has the enormous advantage over other more traditional procedures that it does not require a priori knowledge of the polarity of molecular compounds containing groups of atoms with polarity similar to that of the molecule under investigation. At first sight, MD simulations do not seem suitable to predict the polarity of molecules, such as polymers, with a high number of internal degrees of freedom.However, contributions of cross correlation terms to the polarity of molecular chains fall rather sharply as the distance between dipoles increases, so that MD simulations performed on relatively short chains can give a good account of the polarity of larger ones.21 The value of 1.2 kcal mol21 obtained in this work for the conformational energy of gauche states about CH2]O bonds of the ester residue of ANA is nearly 1 kcal mol21 higher than the value of this quantity currently used in the analysis of the conformational properties of polyesters.19,21 It should be pointed out, however, that this energy has a relatively low effect on the polarity of the molecule.As occurs in polyoxyethylene (POE), where the orientation of the dipoles of two consecutive repeating units are separated by one CH2]CH2 bond, the polarity of ANA is very sensitive to the conformational energy of that Table 3 Dipole moments of the ANA molecule computed at 300 K by adding two contributions arising from the ester groups, (m1 = m2 = 1.75 D, t = 1218, see Fig. 1), with several combinations of conformational energies (in kcal mol21), and location of gauche rotational isomers (trans, were always kept in 1808) Es 1.20 1.00 1.40 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20 1.20 Es1 20.87 20.87 20.87 20.67 21.07 20.87 20.87 20.87 20.87 20.87 20.87 20.87 20.87 Ew 20.13 20.13 20.13 20.13 20.13 20.13 10.07 20.13 20.13 20.13 20.13 20.13 20.13 Eg 20.57 20.57 20.57 20.57 20.57 20.57 20.57 20.77 20.37 20.57 20.57 20.57 20.57 fg(C]O) ±45 ±45 ±45 ±45 ±45 ±45 ±45 ±45 ±45 ±40 ±50 ±45 ±45 fg(C]C) ±55 ±55 ±55 ±55 ±55 ±55 ±55 ±55 ±55 ±55 ±55 ±50 ±60 ·m2Ò/D2 5.73 5.78 5.62 5.62 5.81 6.09 5.44 5.45 5.93 5.65 5.81 5.51 5.94 bond.In the case of POE gauche states about CH2]CH2 bonds give rise to strong repulsive interactions between the negative residual charges of two oxygen atoms.As a consequence of the Coulombic contributions, most semiempirical calculations reported to date for the energy of gauche states about CH2]CH2 bonds in 1,2-dimethoxyethane and POE suggest that the energy Es1 of these states is ca. 0.5 kcal mol21 above that of the alternative trans state. However, in order to reproduce the dipole moment of 1,2-dimethoxyethane and both the dipole moments and the unperturbed dimensions of POE it is necessary to assume that Es1 is ca. 0.5 kcal mol21 below the energy of the corresponding trans state.19,20 The difference between calculated and experimental conformational energies is commonly called the gauche effect in the literature. This effect appears in the cases in which electronegative atoms (O, S, Cl, etc.) intervene in the first-order interactions. Therefore, the rather good agreement between the value of Es1 and the value of this energy in both 1,2-dimethoxyethane and POE is remarkable despite the fact that in the former case rather strong repulsive Coulombic interactions also intervene.The value of Es1 is also in very good agreement with that found by NMR techniques for the energy of gauche states about CH2]CH2 bonds of molecular compounds which give rise to interactions between an oxygen atom of an ester group and an oxygen atom of an ether group.31 In view of these results one can conclude that the potential energy surfaces calculated by using the Amber force field give a good account of the experimental values of the rotational energies about CH2]CH2 bonds which produce first-order O? ? ? O interactions without the need to invoke a gauche effect a posteriori, as is commonly made.Most of the force fields, though conveniently parametrized for conformational analysis of paraffins, polyethylene and polyolefines, fail for polar molecules. In the latter case, the use of the Amber force field seems more convenient to obtain the true values of conformational energies. Experimental Synthesis and characterization of 2-(acetyloxy)ethyl-2- (2-naphthyl)acetate Esterification of naphthyl acetic (NAA) acid with ethylene glycol (EG) (molar ratio NAA/EG = 35/88) was carried out in refluxing toluene, under a nitrogen atmosphere, using ptoluenesulfonic acid as catalyst.The reaction was allowed to proceed for 24 h utilizing a Dean–Stark distillation trap to separate the water formed in the esterification process.The solvent was eliminated at reduced pressure and the monoester, 2- hydroxyethyl-2-(2-naphthyl)acetate (HNA), and diester, ethylene glycol dinaphthyl acetate (EGDA), formed in the reaction were separated in successive silica gel columns using the following eluents: hexane–ethyl acetate (3 : 1), hexane–ethyl acetate (2 : 1) and hexane–ethyl acetate (1 : 1). Characterization of the monoester: 1H NMR (200 MHz) (CDCl3): d 7.85–7.15 (m, 7H, protons of the naphthyl group), 4.05 (m, 2H, ]COO]CH2]CH2]OH), 3.95 (s, 2H, naphthyl- CH2]), 3.80 (s, 1H, OH), 3.50 ppm (t, 2H, ]CH2]OH). 13C NMR (50.4 MHz) (CDCl3): d 171.8 (]COO]), 133.8– 123.6 (carbon atoms of the naphthyl group), 66.5 (]O]CH2] CH2]OH), 61.0 (]O]CH2]CH2]OH), 39.0 ppm (naphthyl] CH2]). Acetyl chloride was added dropwise to a solution of HNA in benzene with stirring. Triethylamine, Et3N, was present in the solution to neutralize the HCl formed. The reaction medium was washed three times with water and the organic phase was separated.In the same way, the aqueous phase was treated with benzene to extract the organic product present in this phase, and the washings were added to the first set. The organic phase was dried with sodium sulfate, filtered and the solvent separated under reduced pressure. The diesterified product, 2-(acetyloxy)-36 J. Chem. Soc., Perkin Trans. 2, 1998 ethyl-2-(2-naphthyl)acetate, was purified using a silicagel column and hexane–ethyl acetate (4 : 1) as eluent.Characterization of ANA: 1H NMR (200 MHz) ([2H6]- DMSO): d 7.96–7.44 (m, 7H, protons of the naphthyl group), 4.30–4.16 (m, 6H, ]CH2]), 1.93 ppm (s, 3H, ]CH3). 13C NMR (50.4 MHz) ([2H6]DMSO): d 171.1, 170.1 (carbonyl groups), 133.3–123.8 (carbon atoms of the naphthyl group), 62.2, 61.8 (O]CH2]CH2]O), 37.9 (naphthyl]CH2]), 20.4 ppm (]CH3). Dipole moment of 2-(acetyloxy)ethyl-2-(2-naphthyl)acetate Static relative permittivities e of solutions of ANA in benzene were measured, at 30 8C and at 10 kHz, with a capacitance bridge (General Radio, type 1620 A) coupled with a threeterminal cell.By plotting the increments of the relative permittivity of the solutions with respect to that of the solvent against the weight fraction of solute, w, a straight line was obtained whose slope, de/dw, is proportional to the total polarization of the solute. In the same way, increments of the index of refraction of the solution, n, with respect to that of the solvent were measured with a differential refractometer (Chromatix, Inc.).The plot of these increments vs. w gave a straight line whose slope, dn/dw, is proportional to the electronic polarization. The atomic polarization is in most cases less than 10% of the electronic polarization and in this case was considered negligible. The value of the mean-square dipole moment, ·m2Ò, of ANA was determined by means of the method of Guggenheim ·m2Ò = 27kBMT 4prNA(e1 1 2)2 Fde dw 2 2n1 dn dwG (6) and Smith,32,33 eqn.(6), where kB and NA are, respectively, the Boltzmann’s constant and Avogadro’s number, M is the molecular weight of the solute and r is the density of the solvent. The subindex 1 in eqn. (6) refers to the property of the solvent. The experimental values of de/dw and dn/dw at 30 8C were 2.18 and 0.046. By substituting these results into eqn. (4) one determines that the mean-square dipole moment of ANA in vacuum is 5.28 D2.The uncertainty of this value was estimated to be ca. 2.5%. Acknowledgements This work was supported by the DGICYT through Grants PB94-0364 and PB 95-0134-C02-01. References 1 G. Williams, Dielectric Spectroscopy of Amorphous Polymer Systems: The Modern Approaches, in Keynote Lectures in Selected Topics of Polymer Science, ed. E. Riande, CSIC, Madrid, 1995. 2 R. Kohlrausch, Ann. Phys., 1847, 12(3), 3931. 3 G. Williams and D. C. Watts, Trans. Faraday Soc., 1970, 66, 80. 4 H. Vogel, Z. Phys., 1921, 22, 645. 5 G. S. Fulcher, J. Am. Ceram. Soc., 1925, 8, 339. 6 G. Tammann and W. Z. Hesse, Anorg. Allgem. Chem., 1926, 156, 245. 7 F. H. Stillinger, Science, 1995, 267, 1935. 8 K. Kauzmann, Chem. Rev., 1948, 43, 219. 9 L. Onsager, J. Am. Chem. Soc., 1936, 58, 1486. 10 H. Fröhlich, Trans. Faraday Soc., 1948, 44, 238. 11 H. Fröhlich, Theory of Dielectrics, Oxford University Press, London, 1958. 12 R. C. Nunes, M. R. Pinto, E. Saiz and E. Riande, Macromolecules, 1995, 28, 211. 13 E. Saiz, C. Alvarez, E. Riande, M.R. Pinto and C. Salom, J. Chem. Phys., 1996, 105, 8266. 14 E. Saiz and E. Riande, J. Chem. Phys., 1995, 103, 3832. 15 E. Saiz, E. Riande, J. Guzmán and M. T. Iglesias, J. Phys. Chem., 1996, 100, 3818. 16 E. Saiz, J. Guzmán, M. T. Iglesias and E. Riande, J. Phys. Chem., 1996, 100, 18 345. 17 C. L. McCormick and K. Kim, J. Controll. Release, 1988, 7, 101. 18 C. L. McCormick, Ann. N.Y. Acad. Sci., 1985, 76, 446. 19 P. J. Flory, Statistical Mechanics of Chain Molecules, Interscience, New York, 1969. 20 A. Abe and J. E. Mark, J. Am. Chem. Soc., 1976, 98, 6468. 21 E. Riande and E. Saiz, Dipole Moments and Birefringence of Polymers, Prentice Hall, Englewood Cliffs, NJ, 1992. 22 QCPE, Department of Chemistry, Indiana University, Bloomington, IN, 47405. 23 T. R. Forester and W. Smith, DL–POLY (Ver. 2.0), Daresbury Laboratory, Daresbury, Warrington, WA4 4AD, England. 24 S. J. Weiner, P. A. Kollman, D. A. Case, U. C. Singh, C. Ghio, G. Alagona, S. Profeta Jr. and P. Weiner, J. Am. Chem. Soc., 1984, 106, 765. 25 S. J. Weiner, P. A. Kollman, D. T. Nguyen and D. A. Case, J. Comput. Chem., 1986, 7, 230. 26 S. W. Homas, Biochemistry, 1990, 29, 9110. 27 F. Mendicuti, M. M. Rodrigo, M. P. Tarazona and E. Saiz, Macromolecules, 1990, 23, 1139. 28 G. D. Smith, D. Y. Yoon and R. L. Jaffe, Macromolecules, 1993, 26, 5213. 29 A. L. McClellan, Tables of Experimental Dipole Moments, Freeman, San Francisco, 1963, Vol. I; Rahara Enterprises, El Cerrito, Cal, 1974, Vol. II; 1989, Vol III. 30 E. Saiz, J. P. Hummel, P. J. Flory and M. Plavsic, J. Phys. Chem., 1981, 85, 3211. 31 J. San Román, J. Guzmán, E. Riande, J. Santoro and M. Rico, Macromolecules, 1982, 15, 609. 32 E. A. Guggenheim, Trans. Faraday Soc., 1949, 45, 714. 33 J. W. Smith, Trans. Faraday Soc., 1950, 46, 394. Paper 7/05566H Received 31st July 1997 Accepted 5th September 1997

ISSN:1472-779X    

DOI:10.1039/a705566h

出版商:RSC    

年代:1998

数据来源: RSC

摘要:

J. Chem. Soc., Perkin Trans. 2, 1999, 33–38 33 Theoretical density functional and ab initio computational study of vertical ionization potentials, dipole moments and 13C and 14NNMR shifts of the 2-mercaptopyridine system. A model for thiated nucleobases † Victor Martinez-Merino * and Maria J. Gil Departamento de Quimica Aplicada, Universidad Publica de Navarra, E-31006 Pamplona, Spain. E-mail: merino@upna.es Lower valence vertical ionization potentials (VIPs) of pyridine-2(1H)-thione (1), 1-methylpyridine-2(1H)-thione (3) and their tautomers (2 and 4) from B3LYP/6-3111G(2d,p) calculations were in very good agreement with experimental values when the SCF calculated first VIP was added to the relative energy of the corresponding Kohn–Sham orbitals.Except for the first VIP, the valence VIPs were poorly reproduced by HF or MP2 calculations following Koopmans’ theorem. Both MP2 and B3LYP electronic correlation methods using 6-3111G(2d,p) basis sets gave good predictions for the dipole moments of 3 and 4 in benzene solution.Dipole moments from calculations including solvent eVects by SCRF methods were very large. Relative 14N and 13C-NMR shifts of 3 and 4 tautomers from the GIAO method applied at the B3LYP/6-3111G(2d,p) level were in good agreement with experimental values using a scaling factor of 0.96. Introduction An understanding of the physico-chemical properties of the nucleobases is of fundamental importance in relation to physical organic chemistry.Many papers have tried to model their physico-chemical properties as can be seen in a very recent review.1 Rotational constants and IR frequencies in the gas phase are correctly reproduced by electronic correlation methods, such as post-HF methods (MP2) or DFT methods (B3LYP), using double-x basis sets 6-31G(d,p).1,2 At the same level of theory, the predicted MP2 and B3LYP dipole moments of NA bases and their model systems without sulfur atoms agree with experimental data, when available, within 0.1–0.3 D; but when sulfur atoms are involved as thiol or thione forms the agreement of B3LYP dipole moments is not so good (e.g., overestimation of 0.3–0.5 D for thiouracil derivatives).1 Ultraviolet electronic spectra and polarizations have also been studied by INDO/S-CI,3 CIS/6-31G* 3 and CASPT23,4 methods for NA bases and by CNDO/S5 for their thio analogs. pÆp* transition energies predicted by semiempirical methods from ab initio geometries agree well (±0.3 eV) with experimental values generally, but polarizations needed ab initio predictions in some cases (e.g., in adenine).In the present work we use the unequivocal physico-chemical properties of ionization potentials from photoelectron spectra in the gas phase,6 dipole moments in benzene 7 and 13C and 14N-NMR shifts in acetone 8 of N- and S-methyl derivatives of pyridine-2(1H)-thione to check the electronic treatment of theoretical methods for thiated nucleobases.The pyridine- 2(1H)-thione (1)–pyridine-2-thiol (2) system is usually recognized as a model of the thiol–thione rearrangement of thio analogs of NA bases,1 whereas the well characterized methyl derivatives 1-methylpyridine-2(1H)-thione (3) and 2-methylthiopyridine (4) allowed us to assign the physico-chemical properties unambiguously. The solvent eVect was modeled using Onsager 9 and SCIPCM10 self-consistent reaction fields † Supplementary data (SUPPL.NO. 57453, pp. 13) is available from the British Library. For details of the Supplementary Publications Scheme, see ‘Instructions for Authors’, J. Chem. Soc., Perkin Trans. 2, available via the RSC web page (http://www.rsc.org/authors). For direct electronic access see http://www.rsc.org/suppdata/perkin2/1999/33/. (SCRF) for nonspecific solute–solvent interactions. The results may provide valuable information on the theoretical study of thiated nucleobases, related heterocyclic systems and DNA.Results and discussion The molecules were first studied by conformational analysis to find the most stable structures in the gas phase at B3LYP, HF and MP2(full)/6-3111G(2d,p) levels of theory. As Table 1 shows, we found two minima for thioimidate forms, one when the N1–C2–S7–X dihedral angle equals zero (conformers 2a and 4a) and another when it is 1808 (conformers 2b and 4b). In methyl derivatives 3 and 4, the methyl group adopts a C2–N1– C–H dihedral angle of 1808 in the tautomer 3 and a C2–S7–C–H dihedral angle of 1808 in conformers 4a and 4b.When these dihedral angles of the methyl group in 3 and 4a were zero, the resulting rotamers were saddle points 6.3 and 3.7 kJ mol21 higher, respectively. These B3LYP/6-3111G(2d,p) calculations agree with estimations for analogous rotational barriers of the methyl group in 1-methylpyridin-2(1H)-one (6.3 kJ mol21) 11 and 2-methoxypyridine (5.0 kJ mol21),11 and they were slightly better than from B3LYP/6-31G(d) calculations (7.0 and 2.6 kJ mol21, respectively).Relative stabilities between conformers (2a/2b and 4a/4b) calculated by ab initio methods without electronic correlation (HF) were essentially the same as by the MP2 or DFT method (B3LYP) within 0.9 kJ mol21 (Table 1). Vibrational zero point energies (DZPE) favoured thioimidate (2a and 4a) vs. thioamide (1 and 3) forms by 210.7 and 26.2 kJ mol21 respectively. When those DZPE were added to electronic energies (DEelec in Table 1), ab initio [HF or MP2(full)] and B3LYP methods predicted thioimidate forms as the more stable tautomers in the gas phase in agreement with experimental data.12,13 In contrast to hydrogen tautomers, electronic correlation methods [B3LYP and MP2(full)] diVered very much from the HF method in the relative stabilization of the thioamide form of methyl derivatives (3 vs. 4a). Hyperconjugation between the methyl group and the p system seems to be responsible in part for that disagreement.As reported recently, MP2 overestimates the eVects of hyperconjugation,14 and the B3LYP method does it slightly.15 The highest OMO of p type for methyl thioamide34 J. Chem. Soc., Perkin Trans. 2, 1999, 33–38 Table 1 Gas-phase electronic energies (Eelec) and dipole moments (m) using the 6-3111G(2d,p) basis set at B3LYP, HF and MP2(full) levels of theory for geometries and wavefunctions N S X N S N S X X 1, X = H 3, X = CH3 2a, X = H 4a, X = CH3 2b, X = H 4b, X = CH3 B3LYP HF MP2 Compound 1 2a 2b 3 4a 4a Eelec a 2646.5831157 2646.5802752 2646.5785763 2685.9011909 2685.9057534 2685.9025129 DEelec b 0.0 7.5 11.9 0.0 212.0 23.5 mc 5.61 2.00 3.11 5.38 1.14 3.24 Eelec a 2644.3012904 2644.3024091 2644.3009121 2683.3366624 2683.3455242 2683.3419624 DEelec b 0.0 22.9 1.0 0.0 223.3 213.9 mc 6.47 2.21 3.36 6.20 1.44 3.49 Eelec a 2645.5998796 2645.6009017 2645.5989723 2684.8209185 2684.8243627 2684.8210467 DEelec b 0.0 22.7 2.4 0.0 29.0 20.3 mc 5.52 2.10 3.24 5.31 1.35 3.34 a In atomic units (1 a.u.= 2625.5 kJ mol21). b In kJ mol21. c In Debyes (1 D = 3.33564 10230 C m). Dipole moments of N-methyl derivative 3 and Smethyl derivative 4 in benzene were 5.22 and 1.68 D respectively (ref. 7). Fig. 1 Five last valence OMO of the methylated tautomers 3 (left) and 4a (right) calculated by means of the B3LYP/6-3111G(2d,p) method. 3 (1p in Fig. 1), very close to its HOMO, exhibited a strong polarization of the methyl group as one would expect from hyperconjugation.Thus, the calculation of charge distribution in 3 and 4a, carried out by means of Natural Population Analysis,16 showed a transfer of 0.28 electrons from the methyl group to the p system (at the sulfur atom mainly) in 3 but not in 4a (20.04 electrons) (Fig. 2). Vertical ionization potentials (VIPs) Katritzky and co-workers 6 used photoelectron spectroscopy to study protomeric equilibria of mercaptopyridines. With this technique, they determined the first valence VIPs for compounds 1–4 in the gas phase, and they also made the corresponding orbital assignments.Table 2 contains the experimental Fig. 2 Charge distribution in 3 and 4a from Natural Population Analysis at the B3LYP/6-3111G(2d,p) level of theory.J. Chem. Soc., Perkin Trans. 2, 1999, 33–38 35 (±0.03 eV) and calculated first VIPs of the compounds. From ab initio (HF and MP2) methods, these are simply sign-reversed orbital energies according to Koopmans’ theorem17 (which does not apply to the density functional formalism18).For DFT methodology, the first VIP was calculated by subtracting the total energy of the neutral molecule from the total energy of the cationic doublet state, spin-unrestricted B3LYP calculations being performed for both the neutral and cationic states. The HF or MP2(full)/6-3111G(2d,p) methods aVorded values of first VIP (negative HOMO energy) close to experimental ones (error <0.15 eV).Except for 2a, ab initio first VIPs were slightly greater than experimental values, and electronic correlation diminished the errors. The MP2 VIPs for thioamide derivatives 1 and 3 were in very good agreement with photoelectron spectra (error <0.04 eV), but for thioimidates 2a and 4a, the errors were 0.08–0.13 eV. The DSCF B3LYP/6-3111G(2d,p) methodology yielded good VIPs for 1, 3 and 4a (error <0.07 eV) giving values slightly less than experimental values, but for 2a the VIP was abnormally low (error of 0.27 eV).Less stable conformers 2b and 4b aVorded VIPs of 0.03 and 0.08 eV higher than for the corresponding 2a and 4a using ab initio calculations. Adiabatic ionization potentials of hydrogen derivatives 1 and 2a, calculated at the B3LYP level, were 0.10 eV smaller than VIPs, as expected. Table 3 shows the relative energies of the five last valence orbitals from HOMO and corresponding orbital assignments for methylated derivatives 3 and 4a.Hydrogen derivatives had a similar energy distribution (data not shown). Fig. 1 displays the five highest valence OMOs for 3 and 4a which correspond to Kohn–Sham orbitals using the B3LYP method as in Table 3. Hartree–Fock orbital types were similar with small diVerences. There is some controversy about whether the Kohn–Sham orbitals have a physical meaning although it is generally accepted that they are good approximations to the Hartree– Fock orbitals. However, recently it has been established 19 that the Kohn–Sham orbitals are physically sound and are expected to be more suitable for use in qualitative molecular orbital theory than either Hartree–Fock or semiempirical orbitals.Our work on OMOs, and another recent work on LUMOs,20 present positive arguments for this assertion. As Table 3 shows, the relative energies of valence OMO between B3LYP/ 6-3111G(2d,p) and experimental values were in good agreement. Deviations increased with descending energy of the OMO, but it was <0.3 eV for the last five OMOs.The B3LYP assignments of the orbital types also agreed with estimations from ref. 6 (Table 3). Only the order of the two highest OMOs of 3 diVered from the previous estimation. It is known that, for the majority of cyclic unsaturated thiocarbonyl compounds, the p ionization is generally at lower energy or at least equal to that for the lone pair on the sulfur atom.21 The B3LYP method gave a VIP from the nS OMO 0.07 eV smaller than for the p type OMO of 3.The proximity of the nitrogen and sulfur atoms could stabilize the non-bonding nS OMO,6 but it seems that an overestimation of hyperconjugation by B3LYP decreased nS energy below that of p OMO. Basis sets in DFT methodology did not aVect the orbital assignments and only slightly aVected the energies. The absolute energies of the last five OMOs from Table 2 First VIP (in eV a) of compounds 1 to 4 Methodb Exper.c HF MP2(full) B3LYPd 1 7.80 7.90 7.84 7.74 (7.65) 2a 8.79 8.75 8.71 8.52 (8.42) 3 7.69 7.78 7.72 7.61 4a 8.24 8.39 8.37 8.20 a 1 eV = 96.5 kJ mol21.b Calculations using 6-3111G(2d,p) basis set for geometry and energy. c From photoelectron spectra, ref. 6. d Adiabatic IPs in parentheses. B3LYP/6-31G(d) were 0.2–0.3 eV less than those from B3LYP/ 6-3111G(2d,p), whereas diVerences in energies relative to HOMO were within ±0.08 eV. As shown from the above data, B3LYP/6-3111G(2d,p) may be a useful tool to obtain precise energies of valence OMO by adding Kohn–Sham relative energies from HOMO (Table 3 data) to the sign-reversed first VIP calculated by the DSCF procedure (Table 2 data).The Koopmans’ energies of the last five valence OMO from ab initio methods agreed poorly with the experimental VIPs of compounds (Table 3). Ab initio methods agreed with the estimation of orbital assignment in the case of thioamide 3. However, HF or MP2/6-3111G(2d,p) methods assigned the nN orbital of thioimidates 2a or 4a as HOMO-3 instead of HOMO-1 like estimations or B3LYP; with a less large basis set like 6-31G(d), it was assigned as HOMO-2.Solvent eVects on dipole moments Dipole moments (DM) of 3 and 4 were previously determined in benzene solution,7 but values in the gas phase remain unknown for methyl (3 and 4) and hydrogen (1 and 2) derivatives. However, experimental dipole moments for oxo analogs of 1 and 2 have been reported from microwave spectroscopy,22 and the DM in benzene of oxo analogs of 3 and 4 have been reported.7 Oxo analogs of 1 and 2 showed gas-phase DM of 0.26–0.31 D greater than the corresponding methylated oxo analogs of 3 and 4 in benzene solution.This agrees with the 0.22–0.27 D increase in dipole moment of 1 vs. 3 from either HF, MP2(full) or B3LYP/6-3111G(2d,p) gas-phase calculations (Table 1). It seems that dipole moments in benzene solution are not much diVerent from gas-phase values. The calculated gas-phase DM of 3 is in good agreement with the experimental value in benzene (5.22 D) when electron correlation is taken into account (entries for B3LYP and MP2 in Table 1).On the other hand, the 2-methylthio derivative has two conformers, 4a and 4b, separated by 8–9 kJ mol21 in the gas phase and with very diVerent DM. The benzene eVect calculated by the Onsager B3LYP/6-31G(d) method for the 4a/4b rotamers was 1.9 kJ mol21 in favor of the larger DM (4b). So the relative concentration of 4a vs. 4b in benzene should be 14 or 16 : 1, using gas-phase diVerences from MP2 or B3LYP (Table 1) and including the above benzene eVect. Consequently, and applying the Eliel equation,23 the calculated DM for 4 in benzene would be about 1.37 or 1.54 D from B3LYP or MP2 methods, respectively, rather close to the experimental value of 1.68 D. Note that the calculated DM of methylated thioamide 3 was Table 3 Relative energies (DE, in eV) of valence orbitals to HOMO and assignments for methylated 3/4a tautomers Compound Methoda DE [Assignment] b 3 4a Exper.c Estimated c B3LYP HF MP2(full) Exper.c Estimated c B3LYP HF MP2(full) 0.00 [p] 0.00 [nS] 0.00 [1p] 0.00 [1p] 0.00 [p] 0.00 [1p] 0.00 [1p] 0.00 [1p] 20.17 [nS] 20.07 [1p] 20.71 [nS] 20.82 [nS] 21.32 [nN] 21.28 [nN] 22.13 [2p] 22.08 [2p] 22.81 [p] 22.88 [2p] 23.39 [2p] 23.45 [2p] 22.01 [p] 21.91 [2p] 22.70 [3p] 22.67 [3p] 22.97 [p] 22.98 [3p] 23.76 [3p] 23.68 [3p] 22.38 [p] 22.23 [3p] 22.96 [nN] 23.10 [nN] 24.42 n.j.d 24.09 [s 1 nS] 25.51 [s 1 nS] 5.68 [s 1 nS] 23.23 [nS] 22.96 [nS] 23.99 [nS] 24.05 [nS] a Calculation using 6-311 1 G(2d,p) basis set for geometry and energy. b Calculated assignments following Fig. 1. c From ref. 6. d Not justified in ref. 6.36 J. Chem. Soc., Perkin Trans. 2, 1999, 33–38 Table 4 14N and 13C-NMR chemical shifts relative to nitromethanea and TMSb respectively Compound 3 4a Method Experimental d HF/6-3111G(2d,p) Scaled e B3LYP/6-3111G(2d,p) Scaled e Experimental d HF/6-3111G(2d,p) Scaled e B3LYP/6-3111G(2d,p) Scaled e N1 189 ± 2 259.2 243.1 194.1 187.1 79 ± 3 122.9 117.9 80.2 76.9 C2 181.1 210.0 197.0 193.4 186.4 160.4 171.7 164.7 173.5 166.4 C3 134.6 137.5 129.0 143.1 137.9 121.8 122.0 117.0 125.4 120.3 C4 135.7 141.6 132.8 135.1 130.2 136.5 145.2 139.2 139.1 133.4 C5 113.3 108.8 102.1 113.3 109.2 119.7 118.1 113.3 121.2 116.2 C6 142.6 146.6 137.5 145.2 140.0 150.0 157.5 151.0 155.7 149.3 Me 45.7 43.8 41.1 48.0 46.3 12.9 16.5 15.8 19.5 18.7 a cCompd 2 cNO2Me in ppm.b cTMS 2 cCompd in ppm. c Wavefunction and geometry as indicated. d Determined in acetone solution from ref. 8. e Following ref. 30. 0.2–0.3 D less than that of the hydrogen derivative 1, whereas the DM of the methyl derivative of the thioimidate 4a was 0.7– 0.9 D less than that of the hydrogen derivative 2a (Table 1). Furthermore, the methyl rotamer 4b had a DM 0.1 D bigger than the hydrogen rotamer 2b.Thus the common practice in DM analysis to attribute similar values for methyl and hydrogen derivatives should be used with care for this type of heterocyclic system. The SCRF methods overestimated the DM. The Onsager B3LYP/6-31G(d) method assigned 6.24 D to 3 in benzene, 1.02 D bigger than the experimental value. The SCIPCM B3LYP/6- 31G(d) method gave 6.59 D for 3, whereas the B3LYP/6- 3111G(2d,p)//B3LYP/6-31G(d) calculation yielded 6.40 D. Moreover, SCRF methods gave greater diVerences from the gas phase as well as in the DM of the molecule.Thus Onsager B3LYP/6-31G(d) increased the DM of 4a, 4b and 3 in benzene solution by 0.17, 0.29 and 0.66 D respectively from the gasphase value. 13C and 14N-NMR chemical shifts A comparison of the experimental and theoretical NMR spectra can be very useful in making correct assignments and understanding the basic chemical shift–molecular structure relationship. Semiempirical methods 24 provide a correct qualitative understanding of magnetic shielding tensors (c), but they are not quantitatively accurate.Ab initio methods have been remarkably successful by applying the gauge factors to atomic orbitals (GIAO method25) or to localized molecular orbitals (IGLO26 or LORG27 methods). A new implementation of the GIAO method was claimed to be better for studying molecules with delocalized electron structure,28 as in the case of 1 to 4. The minimum recommended model for predicting NMR properties is GIAO HF/6-31G(d)//B3LYP/6-31G(d), but the triple-x basis set at HF level reproduces much better chemical shifts.29 We have used this method, and we have also analyzed electron correlation by means of DFT (B3LYP) theory.Table 4 shows the relative chemical shifts for both methylated tautomers 3 and 4a. The GIAO HF/6-3111G(2d,p) method properly reproduced the relative order of 13C-NMR shifts in 4a, and it explained up to 99% of experimental shift variance. When a scaling factor of 0.959 was introduced,30 then the error of calculated shifts diminished from 4.8 ± 5.0 ppm to 0.0 ± 4.5 ppm.The GIAO ab initio method also predicted nearly 98% of experimental shift variance in thioamide 3, but the C2 shift was placed further downfield than the experimental value. On the contrary, the GIAO HF method shifted 14N to much higher fields than experimental values for both 3 and 4a. In spite of that, GIAO HF/6-3111G(2d,p) could be considered an acceptable method to assign correct 13C-NMR shifts.The GIAO B3LYP/6-3111G(2d,p) method introduced electronic correlation through an impure DFT method, but it was a very eYcient tool for the quantitative reproduction of both 14N and 13C-NMR shifts, as Table 4 shows. Scaling factors of 0.959 for 4a and 0.964 for 3 were statistically deduced as the best to fit GIAO B3LYP shifts to experimental ones.30 The errors obtained for 4a and 3 13C-NMR shifts in acetone using scaled GIAO B3LYP/6-3111G(2d,p) data in the gas phase were 0.5 ± 4.3 ppm and 20.5 ± 4.3 ppm respectively, and they explained more than 99% of experimental variance.Relative 14N-NMR shifts were reproduced exactly by scaled GIAO B3LYP methodology, even within experimental error (Table 4). Scaling factors were essentially the same for HF and B3LYP/6- 3111G(2d,p) methods and also for 4a and 3 tautomers,31 and could be approximated by 0.96. Geometries of the compounds are an important factor to be considered in the reproduction of NMR shifts.Thus GIAO HF/6-3111G(2d,p)//B3LYP/6- 31G(d) gave relative 4a 13C-NMR shifts 1–2 ppm higher than from B3LYP/6-3111G(2d,p) geometry. Conclusions Both 1 and 2 Koopmans’ ionization potentials from HF, or MP2(full)/6-3111G(2d,p) were accurate within ±0.10 eV of experimental values. Except for 2, vertical IPs of 1 through 4 from B3LYP/6-3111G(2d,p) were more exact than Koopmans’ IPs from HF and MP2 calculations. However, DFT methodology gave a more accurate description of known relative energies of valence molecular orbitals for compounds 1 through 4 than ab initio methodology. Dipole moments in benzene solution of 3 and 4 were better reproduced by calculations in the gas phase than by SCRF methods.Methods which include electronic correlation, like MP2(full) or B3LYP/6- 3111G(2d,p), were necessary to approximate dipole moments. However, HF/6-3111G(2d,p) gave poor results. The 14N-NMR shifts of 3 and 4 in acetone were exactly reproduced by B3LYP but poorly by HF calculations using GIAO methodology and a 6-3111G(2d,p) basis set.Except for C2, which was overestimated (>10 ppm), experimental 13C-NMR shifts of 3 and 4 in acetone agreed well with calculated shifts from both HF and B3LYP/6-3111G(2d,p) methods in the gas phase. A scaling factor of 0.96 is recommended to fit calculated relative 14N and 13C-NMR chemical shifts to experimental ones at the above levels of theory.The results from the B3LYP/6-3111G(2d,p) method indicate that it could be a very eYcient tool for modeling properties derived from electronic distribution of thiated nucleobases. Theoretical methods and computational details All calculations were performed using the GAUSSIAN94 (g94) 32 suite of programs. Geometrical optimizations of neutral molecules in the gas phase were carried out at three diVerent theoretical levels: restricted Hartree–Fock (RHF),33 secondorder Møller–Plesset 34 including all electrons in the correlation calculation (MP2-full), and DFT using the three-parameter hybrid function developed by Becke (B3LYP).35 The basis sets used were Pople’s 6-3111G(2d,p) for H, C, N and S.33,36 TheJ.Chem. Soc., Perkin Trans. 2, 1999, 33–38 37 analytical harmonic vibrational frequency calculations were done to characterize the nature of stationary points on the potential energy surface and to estimate the zero-point vibrational energy (ZPE) at the HF and B3LYP levels.The ZPE values were scaled by 0.89 at the HF level and by 0.98 at the B3LYP level to eliminate known systematic errors.10,37 The valence ionization potentials (IPs) were obtained from ab initio spin-restricted HF and MP2 calculations following Koopmans’ theorem: simple sign-reversed orbital energies. However, DFT IPs were obtained by an DSCF procedure using spin-unrestricted calculations. In more detail, the diVerence between the calculated energies of the neutral and cation at their respective equilibrium geometries, corrected for the ZPE, gave the adiabatic ionization potential (AIP); calculation for the cation at the geometry of the neutral gave the vertical ionization potential (VIP) (no correction for ZPE was made in this case).38 In all cases, when unrestricted wave functions were used, the spin contamination was small (·S2Ò < 0.76).Magnetic shielding tensors 13C and 14N, calculated at the HF or B3LYP/6-3111G(2d,p) levels from the optimized structures, were obtained following the gauge-independent atomic orbital (GIAO) method.25,28,39 The diVerences between magnetic shielding tensors (c) of nuclei in the reference molecule (tetramethylsilane for 13C-NMR and nitromethane for 14N-NMR) and the studied molecule (3 and 4) were considered as relevant chemical shifts.25 Nonspecific solvent eVects on the geometry and physicochemical properties of the molecules and relative stability at the tautomeric equilibrium were studied using the self-consistent reaction field (SCRF) with the Onsager model9 and with the self-consistent isodensity polarized continuum model (SCIPCM)10 developed from a reaction field based on the polarized continuum model proposed by Tomasi and co-workers.40 In the SCRF calculations, the solute is placed in a uniform electric field of solvent with a dielectric constant e or a reaction field.In the SCRF Onsager model, the solute is assumed to occupy a spherical cavity of radius a0 in the medium.The cavity radius for each conformer, a0, was determined by performing a singlepoint calculation with the keyword VOLUME of g94 programs at the optimized geometry of the B3LYP/6-31G(d) level (gas phase). The resulting values of a0 were 4.45 Å for compound 3, 4.32 Å for its tautomer 4a (dihedral angle N1–C2–S–Me = 08), and 4.19 Å for the conformer 4b (dihedral angle N1–C2–S– H = 1808). In the SCIPCM model, the cavity is defined as an isosurface of the molecule, and the coupling of the isosurface and the electron density are taken fully into account.This procedure solves for the electron density that minimizes the energy, including the solvation energy, which itself depends on the cavity which depends on the electron density. In this case, the eVects of solvation are folded into the iterative SCF computation rather than comprising an extra step afterwards. All SCRF geometry optimizations in solution were carried out at the B3LYP/6-31G(d) level of theory.Single-point energy calculations were also performed using SCRF HF or B3LYP/6- 3111G(2d,p) wavefunctions and above SCRF B3LYP/6- 31G(d) optimized structures. Acknowledgements We gratefully acknowledge financial support for this project by the UPNA Research Comission, Pamplona, Spain. We are also grateful to Dr J. I. Garcia for careful and constructive criticism. References 1 M. J. Nowak, L. Lapinski, J. S. Kwiatkowski and J.Leszczynsky, Computational Chemistry: Reviews of Current Trends, ed. J. Leszczynsky, World Scientific Publishing, Singapore, 1997, vol. 2, ch. 4. And references cited therein. 2 J. S. Kwiatkowski and J. Leszczynsky, Int. J. Quantum Chem., 1997, 61, 453. 3 A. Broo and A. Holmen, J. Phys. Chem., 1997, 101, 3589. 4 J. Lorentzon, M. P. Fülscher and B. O. Roos, J. Am. Chem. Soc., 1995, 117, 9265. 5 M. J. Stewart, J. Leszczynsky, Y. V. Rubin and Y. P. Blagoi, J. Phys. Chem., 1997, 101, 4753. 6 M. J. Cook, S. El-Abbady, A. R. Katritzky, C. Guimon and G. Pfistar-Guillonzo, J. Chem. Soc., Perkin Trans. 2, 1977, 1652. 7 H. Lumbroso and D. M. Bertin, Bull. Soc. Chim. Fr., 1970, 1728. 8 L. Stefaniak, Org. Magn. Reson., 1979, 12, 379. 9 (a) M. W. Wong, M. J. Frisch and K. B. Wiberg, J. Am. Chem. Soc., 1991, 113, 4776; (b) M. W. Wong, K. B. Wiberg and M. J. Frisch, J. Am. Chem. Soc., 1992, 114, 523; (c) M. W. Wong, K. B. Wiberg and M. J. Frisch, J. Am. Chem.Soc., 1992, 114, 1645. 10 J. B. Foresman and Æ. Frisch, Exploring Chemistry with Electronic Structure Methods, 2nd edn. Gaussian Inc., Pittsburgh, PA, 1996. 11 P. Beak, J. Bonham and J. T. Lee, Jr., J. Am. Chem. Soc., 1968, 90, 1569. 12 M. J. Nowak, L. Lapinski, H. Rostkowska, A. Les and L. Adamowcz, J. Phys. Chem., 1990, 94, 7406. And references cited therein. 13 P. Beak, J. B. Covington and J. M. White, J. Org. Chem., 1980, 45, 1347. 14 J. W. Gauld, M. N. Glukhovtsev and L.Radom, Chem. Phys. Lett., 1996, 262, 187. 15 M. Sodupe, A. Oliva and J. Bertran, J. Phys. Chem., 1997, 101, 9142. 16 E. D. Glendening, A. E. Reed, J. E. Carpenter and F. Weinhold, NBO program (version 3.1), included in GAUSSIAN94 (ref. 32). 17 T. A. Koopmans, Physica, 1933, 1, 104. 18 S. B. Trickey, Conceptual Trends in Quantum Chemistry, ed. E. S. Kryachko and J. L. Calais, Kluwer Academic Publishers, Dordrecht, 1994. 19 E. J. Baerends and O. V. Gritsenko, J. Phys. Chem., 1997, 101, 5383. 20 J. I. Garcia, V. Martinez-Merino and J. A. Mayoral, J. Org. Chem., 1998, 63, 2321. 21 (a) C. Guimon, G. Pfister-Guillouzo, M. Arbelot and M. Chanon, Tetrahedron, 1974, 30, 3831; (b) C. Guimon, G. Pfister-Guillouzo and M. Arbelot, Tetrahedron, 1975, 31, 2769; (c) C. Guimon, G. Pfister-Guillouzo and M. Arbelot, J. Mol. Structure, 1976, 30, 339. 22 L. D. Hatherley, R. D. Brown, P. D. Godfrey, A. P. Pierlot, W. Caminati, D. Damiani, S. Merlandri and L. B. Favero, J.Phys. Chem., 1993, 97, 46. 23 E. L. Eliel, N. L. Allinger, S. J. Angyal and G. A. Morrison, Conformational Analysis, Wiley, New York, 1967. 24 H. Fukui, Magn. Reson. Rev., 1987, 11, 205. 25 R. Ditchfield, Mol. Phys., 1974, 27, 789. 26 (a) M. Schindler and W. Kutzelnigg, J. Chem. Phys., 1982, 76, 1919; (b) M. Schindler and W. Kutzelnigg, J. Am. Chem. Soc., 1983, 105, 1360. 27 (a) A. E. Hansen and T. D. Bouman, J. Chem. Phys., 1985, 82, 5035; (b) A. E. Hansen and T. D. Bouman, J.Chem. Phys., 1989, 91, 3552. 28 K. Wolinski, J. F. Hinton and P. Pulay, J. Am. Chem. Soc., 1990, 112, 8251. 29 J. R. Cheeseman, G. W. Trucks, T. A. Keith and M. J. Frisch, J. Chem. Phys., 1996, 104, 5497. 30 Scaling factors of each compound were deduced by means of linear regression between experimental and calculated relative 13C-NMR shifts forcing the intercept through the origin. 31 Note that scaling factor for GIAO HF/6-3111G(2d,p) in 3 from regression (ref. 30) was 0.938, but it was for the very high shift of C2.When this atom was excluded from analysis, a scaling factor of 0.983 was obtained. 32 M. J. Frisch, G. W. Trucks, H. B. Schlegel, P. M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Cheeseman, T. A. Keith, G. A. Petersson, J. A. Montgomery, K. Raghavachari, M. A. Al-Latham, V. G. Zakrzewski, J. V. Ortiz, J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayakara, M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W. Wong, J. L. Andrés, E. S. Replogle, R. Gomperts, R. L. Martin, D. J. Fox, J. S. Binkley, D. J. Defrees, J. Baker, J. J. P. Stewart, M. Head-Gordon, C. González and J. A. Pople, GAUSSIAN94, Revision C.3, Gaussian Inc., Pittsburgh, PA, 1995. 33 W. J. Hehre, L. Radom, P. v. R. Schleyer and J. A. Pople, Ab initio Molecular Orbital Theory, Wiley, New York, 1986. 34 (a) C. Moller and M. S. Plesset, Phys. Rev., 1934, 46, 618; (b) M. Head-Gordon, J. A. Pople and M. J. Frisch, Chem. Phys. Lett.,38 J. Chem. Soc., Perkin Trans. 2, 1999, 33–38 1988, 153, 503; (c) M. J. Frisch, M. Head-Gordon and J. A. Pople, Chem. Phys. Lett., 1990, 166, 275; (d) M. J. Frisch, M. Head-Gordon and J. A. Pople, Chem. Phys. Lett., 1990, 166, 281. 35 A. D. Becke, J. Chem. Phys., 1993, 98, 5648. 36 (a) M. J. Frisch, J. S. Binkley and J. A. Pople, J. Chem. Phys., 1984, 80, 3265; (b) A. D. McLean and G. S. Chandler, J. Chem. Phys., 1980, 72, 5639; (c) R. Krishnan, J. S. Binkley, R. Seeger and J. A. Pople, J. Chem. Phys., 1980, 72, 650. 37 J. Urban, P. R. Schreiner, G. Vacek, P. v. R. Schleyer, J. Q. Huang and J. Leszczynsky, Chem. Phys. Lett., 1997, 264, 441. 38 E. P. F. Lee and T. G. Wright, J. Phys. Chem., 1997, 101, 1374. 39 J. Gauss, J. Chem. Phys., 1993, 99, 3629. 40 (a) S. Miertus, E. Scrocco and J. Tomasi, Chem. Phys., 1981, 55, 117; (b) J. Tomasi, R. Bonaccorsi, R. Cammi and F. O. Valle, J. Mol. Struct., 1991, 234, 401; (c) J. Tomasi and M. Persico, Chem. Rev., 1994, 94, 2027. Paper 8/07577H

ISSN:1472-779X    

DOI:10.1039/a807577h

出版商:RSC    

年代:1999

数据来源: RSC