摘要:

The compoundH = 1,4-bis(9-hydroxyfluoren-9-yl)benzene is a versatile host molecule which conforms to Weber’s host design specifications in that it is rigid and bulky and contains hydroxy moieties, which are good hydrogen-bond donors.1We have employed this host for the separation of lutidine (dimethylpyridine) isomers.2While the synthesis, structure and thermal stability of organic inclusion compounds have been widely discussed,3the question of guest exchange has received little attention despite the fact that such processes are important for sensing and catalysis based on inclusion.4–6Recent studies on this subject include the study of single crystal to single crystal transformations induced by guest exchange in inclusion compounds of cholic acid.7Bulky hosts containing the fluorenyl moiety are responsive to volatile guests and the exchange of EtOH and n-PrOH has been monitored by X-ray powder diffraction.8A bisresorcinol derivative of anthracene has been shown to undergo guest exchange that is concomitant with structural adjustment while retaining its crystallinity.9We have crystallised the host (H) with acetone, DMSO and propan-2-ol and elucidated the following structures:1H· acetone;2H·DMSO;3H·MIX (MIX = 58% DMSO + 42% acetone as mole fractions) and4H·2 PriOH.The crystals of1and2are isostructural and crystallise in space groupP1&cmb.macr;, and the guest molecules lie in channels running parallel to [100]. The projection of1is shown inFig. 1(a) and 1(b), the latter with the guest molecules omitted and the host as a van der Waals representation, so that the channels are clearly evident. In both structures the guests are stabilised by two independent (Host)O–H ⋯ O(Guest) hydrogen bonds, details of which are given inTable 1. Complex4also crystallises inP1&cmb.macr; but is different in that the host–guest ratio is 1 ∶ 2 and the channels in which the PriOH guests are located are wider as shown inFig. 2(a) and 2(b), which is viewed along [010]. Each PriOH is stabilised by two independent H-bonds.Hydrogen-bonding parametersCompoundDonor (D)Acceptor (A)D–H/ÅD ⋯ A/ÅD ⋯ H–A/ °−x, 1 − y, 1 − z.1O9AO1G0.93(2)2.866(2)172(2)O9BO1G0.90(2)2.870(2)171(2)2O9AO1G0.91(2)2.791(2)173(2)O9BO1G0.95(2)2.785(2)171(2)3O9AO1G0.93(3)2.795(2)170(2)O9BO1G0.94(3)2.798(2)167(3)4O9O1G0.98(2)2.722(2)166(2)O1G O90.93(2)2.840(2)167(2)(a) Projection of1along [100], H-bonds are shown by dotted lines; (b) space-filling projection of1along [100] with guest molecules omitted, showing the open channels.(a) Projection of4along [010], H-bonds are shown by dotted lines; (b) space-filling projection of4along [010] with guest molecules omitted, showing the open channels.We have carried out selectivity experiments between all three pairs of guests by dissolving the host in mixtures of the guests in different proportions, allowing the inclusion compounds to crystallise and by analysing the guests in the crystals. The results are given inFig. 3. These show that there is virtually no discrimination between acetone and DMSO [Fig. 3(a)] and that PriOH is selected preferentially over both DMSO and acetone [Fig. 3(b) and 3(c)].Results of the competition experiments: (a) DMSOversusacetone; (b) DMSOversusPriOH; (c) PriOHversusacetone.Xguestis the mole fraction of guest in the liquid mixture;Zguestis the mole fraction of this guest in the crystal.The result of the acetone/DMSO competition as well as the fact that their complexes are isostructural led us to attempt a guest exchange experiment. We preferred to work with powdered samples in order to avoid retardation of the process by diffusion barriers in large single crystals. We prepared theH·acetone compound by exposing a powdered specimen of the host to acetone vapour for 12 hours. The resulting inclusion compound is identical to structure1, as demonstrated by X-ray powder diffraction. The guest exchange was carried out by exposingH·acetone powder to DMSO vapour in a closed vessel at 25 °C. We sampled the resulting compound at regular intervals by DSC. The results are shown inFig. 4which displays the movement of the first endotherm fromTon = 81 °C, corresponding to pure acetone guest, toTon = 197 °C for2. The second endotherm,Ton = 261 °C corresponds to the host melt, and remains constant throughout. The reaction is complete after 86 hours and the fact that we obtain a moving but single endotherm of guest release shows that a continuous solid solution of the mixed guests has been formed in the channels of the host.Migration of the desorption endotherm as a function of time for the guest exchange experiment (H·acetone + DMSO→H·DMSO + acetone).We also grew single crystals with mixed guests. GC analysis showed that these contained 58% DMSO and 42% acetone (as mole percentages). We elucidated the structure of this compound,3(H·MIX), which is isostructural with1and2. However, the difference electron density map obtained after the host structure had been refined could only be interpreted as a severely disordered guest, averaging DMSO and acetone. The packing of3, however, is essentially the same as that of1and2. We have carried out similar structure analysis of an inclusion compound with mixed guests, when the host 1,1-bis(dihydroxyphenyl)cyclohexane was shown to enclathrate a mixture of 2,3-xylenol and 3,5-xylenol.10The structure of4is distinctly different from those of1,2and3. We therefore considered the problem of structure change as the ratio of PriOH–acetone and PriOH–DMSO was increased systematically. We therefore obtained X-ray powder diffraction (XRPD) patterns from powders with mixed guests and found that for the acetone–PriOH series the structure adopted is that of1forxPriOH ≤ 0.3, after which it changes to that of4. In the case of DMSO–PriOH, however, structure2prevailed throughout. The XRPD patterns for1and4are shown inFig. 5. We have demonstrated similar effects with the structures of 1,1,2,2-tetraphenylethane-1,2-diol with mixtures of 2,6-lutidine and 3,5-lutidine.11Experimental XRPD traces of1(solid line) and4(dotted line).Lattice energy calculations were performed for1,2and4using the atom–atom potential method. We employed the program EENY,12using a force field of the typeV(r ) = aexp (−br) − c/r6whereris the interatomic distance and the coefficientsa,bandcare those given by Gavezzotti.13We incorporated a hydrogen bonding potential that is a simplified version of that given by Vedani and Dunitz 14and is formulated asV(H-bond) = (A/R12 − C/R10) cos2&thetas;whereRis the distance between the hydroxy hydrogen and the O acceptor,&thetas;is the O–H ⋯ O angle, and the cos2&thetas;term is the energy penalty paid by the bond to take into account non-linearity. We obtained the following values for the lattice energies:1, −533.6 kJ mol−1,2, −533.4 kJ mol−1and4, −540.6 kJ mol−1. The lattice energy values show the stabilities of the inclusion complexes to be in the order4 > 2 ≈ 1. This is in agreement with the results obtained from the competition experiments where the host compound shows no discrimination between acetone and DMSO while PriOH is selected preferentially over both acetone and DMSO. The contributing factor for this is the additional hydrogen bonding which occurs in the PriOH structure. In inclusion compounds lattice energy calculations are only strictly valid for host–guest systems which have the same stoichiometry and where the guests are isomeric. The values obtained are therefore only offered as a guideline.

ISSN:1472-779X    

DOI:10.1039/b009446n

出版商:RSC    

年代:2000

数据来源: RSC

摘要:

IntroductionIn the course of studies on alkaloidal constituents of the South American vinesUncaria tomentosa(Willd.) DC. andUncaria guianensis(Aubl.) Gmel. (Rubiaceae),1N-oxides of oxindole alkaloidsThe numbering system is based on that customarily used for the hetero-yohimbinoid alkaloids.were required as reference compounds. A simple process was desired to yield pureN-oxide samples in aqueous solutions which can be used directly for reversed-phase HPLC. Hydrogen peroxide was chosen as the oxidant. Because of the known tendency of spiro oxindole alkaloids to isomerize in aqueous solution2and due to their poor solubility in water, the effect of organic co-solvents was to be examined. Of course these co-solvents had to be compatible with the requirements of HPLC. During this work the rate-enhancing effect of acetonitrile on oxidations in neutral aqueous solutions was discovered and shown to be a general phenomenon.

ISSN:1472-779X    

DOI:10.1039/b102066h

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

ResultsIn applying the SCI-PCM solvation model, based on the non-spherical cavity continuum model of Tomasi,8one requires a solvent relative permittivity (ϵ) input parameter. The experimental data being modeled in the present calculations were obtained in a variety of solvents, but most of the low-temperature results involved SO2ClF solvent, containing an excess of SbF5. No literature data for theϵof SO2ClF were found, but a value of 9.1 is reported 9for the related SO2Cl2. Relative permittivities are slightly temperature dependent and given all this uncertainty, a value ofϵ = 30.0 was chosen for the present calculations. In part, we are also assuming that an excess of the strong Lewis acid SbF5would create a slightly more polar medium than that of the pure solvent.In practice, the calculation results are not very dependent onϵvalues once one chooses a relatively high value such as 30.0. Doubling this to 60.0 in the calculations results in only small changes to the differential solvation values because the relative permittivity appears as anϵ − 1/ϵfactor in solvation theory. This fact could also explain why our previously reported experimental results2in FSO3H solvent (ϵ = 120–150) 10did not differ very much from the related SO2ClF solvent data, even though it would now appear from the present work that a substantial differential solvation energy (CHisomervs.CCisomer) is involved.Calculated differential solvation energiesInTable 1are shown the calculated relative energies for the cation systems1,2and3, in both theirCHandCCisomeric forms, comparing in particular the differences between the gas-phase results and those using the solvation model. The comparison data for thecis-1,3,5-trimethyl system2are missing because we could not find a gas-phase energy minimum for theCHisomer. However this problem does not exist when the SCI-PCM solvation model is used.Calculated relative energies forCHandCCcarbocation isomersGas phaseSolvation simulationDifferential solvation favoringCHisomerA B A1 B1 B kJ mol−1, relative to the most stable isomer = 0.0.Amethod: B3LYP/6-31G** + 0.98 ZPVE,Bmethod: B3LYP/6-311+G**//B3LYP/6-31G** + 0.98 ZPVE/6-31G**.Basis sets and ZPVE as inbusing the SCI-PCM method. Strictly speaking, the calculated solvation energies are free energies, but for the differential solvation energy between two ions in the same solvent, the entropy term specific to the gas-phase solvation process would effectively cancel, so that ΔΔHsolv ∼ ΔΔGsolv.A geometry optimization of this isomer was unsuccessful, giving instead an isomer of theCCform with the C1 methyl group rotated 60°.1CC0.00+0.716.07.06.31CH0.330.000.000.002CC——2.33.52CH0.000.003CC2.54.08.69.85.83CH0.000.000.000.00The gas phase–solvation simulation data inTable 1can be directly compared for the1and3systems, showing that solvationdifferentially stabilizes theCHisomersby about 6 kJ mol−1. This computational area is still relatively unexplored but the magnitude of this result is quite significant, and this is discussed later. In comparisons with experimental data (Table 2), the solvation results now overestimate the experimental (solution)CHisomer stability by about 2 kJ mol−1. However it should be kept in mind that Møller–Plesset-based theoretical methods show a somewhat larger gas-phase stabilization of theCCform of1(compared to1CH), which even after the differential solvation “correction”, come out on the opposite side of the experimental results to those shown inTable 2. In either case the experimentalvs.solvation model results are now within ±4 kJ mol−1.Comparison of experimental and calculated (SCI-PCM solvation model) energy differences for theCHandCCisomersCC ⇄ CHCation systemCalc. energy diff.Experimental ΔH ΔHcalc. − ΔHexp.Taken from Table 1.Ref.1.1−7.0−4.6−2.42−3.5−1.0−2.53−9.8≥−7.5−2.3Structure comparisons—gas phasevs.solvation modelStructures for the gas-phase1CCand1CHcations have been published.3Comparison of the structures for gas-phase and solvated species shows that these are very similar in both cases. There is a slight reduction in the C2–C3 (C6–C5) bond from 1.604 to 1.596 Å in the solvated structure of1CC, this long carbon–carbon bond being the most important structural parameter characterizing this isomer. In1CH, the hyperconjugating C2–Hax(C6–Hax) bond is almost unchanged (1.114vs.1.112 Å).Similar comparative results are found for2CC, and for3CHand3CC, where the C2–C3 (C6–C5) bond is reduced from 1.616 to 1.606 Å in2CC, and from 1.602 to 1.592 Å in3CC, whereas3CHhas C2-Hax(C6-Hax) bonds which are virtually identical (1.112 Å) in the gas phase and solvation model.Remote methyl substituent effects—calculatedvs.experimentalAs outlined in the Introduction, our previously reported experimental study 1showed that the addition of remote methyl substituents to the cation1system brought about significant changes in the respectiveCH–CCequilibrium constants (assuming of course that the data were being correctly interpreted). Clearly, if one were able to model these substituent effect changes using MO theory, then this in itself would strengthen the original interpretation. As discussed in the Introduction, thecis-1,3,5- (2) and 1,4,4-trimethyl-1-cyclohexyl (3) cation systems were selected. The most meaningful comparison is that shown in column 4 ofTable 2. Even though the individual calculatedvs.experimental results differ in magnitude byca.2 kJ mol−1, this difference is almost constant for the1,2and3sets, so that overall the substituent effectsper seare being very well modeled by the calculations.The opposing effect of the 4,4- andcis-3,5-dimethyl groups when added to the skeleton of1is of some interest. Using isodesmic reactions of the type2CC + cyclohexane → 1CC + 1,3-dimethylcyclohexaneone finds that2CCis stabilized by methyl substituents (14.2 kJ mol−1). This result is intuitively reasonable given the ability of CH3vs.H in stabilizing a positive charge on carbon. A related evaluation of2CHwas not possible since this structure was not found as a minimum in the gas-phase calculations (vide infra). In the 4,4-dimethyl case, the corresponding3CCis destabilized by 12.3 kJ mol−1while3CHis stabilized by 8.58 kJ mol−1. These individual numbers may be unreasonably large but the net effect is a destabilization of3CCby the substituents. This result is not readily predictable, but as discussed later, these observations (and the actual existence of isomericCHandCCcyclohexyl cation structures) are potentially very relevant in the interpretation of solvolysis results involving cyclohexyl systems. The destabilization of3CCalso agrees with results using remote methyl substituents in 2-methyl-2-adamantyl cation equilibria.Relative entropies of the CC and CH isomersThe calculated internal entropy differences between theCCandCHisomers in the1,2and3series, using frequency data, are small and show no particular trends. However, the experimental data 1for this same series,CC ⇄ CH, gives ΔS = −6 to −8 J K−1for each system.In the original study no explanation was offered for the experimental ΔSobservations, but the present calculations now offer a reasonable rationale. This is based on the fact that the calculations show theCHisomer to be more strongly solvated than theCCcounterpart.Larger solvation energies for a solute–solvent interaction should differentially reduce the entropy of the solvent system around the more highly solvatedCHisomer because of a lowered randomness of this specific portion of the bulk solvent. This effect would be part of theexperimentalΔSterm associated with theCC ⇄ CHequilibrium process and we suggest that the measured ΔSof −6 to −8 J K−1is mainly due to this factor (common to all three cation systems).NMR chemical shift calculationsThe13C NMR chemical shifts calculated for the six carbocations involved in this study are listed inTable 3. There is a close similarity between the individual carbon shifts in theCHandCCisomers, except for the C3–C5 chemical shifts, and to a lesser extent the C1 center. In the C3–C5 chemical shift comparison, the difference is an almost constant 30 ppm in the1,2and3series. This same large divergence was originally deduced from the experimental13C NMR results, and in fact was implicit in the original “two species” postulate, since one can only expect large chemical shift changes for a dynamically averaged signal, as a function of temperature, when the individual isomers have at least one quite divergent chemical shift value.Calculated13C chemical shifts for theCCandCHisomers of the cation systems1,2and3Carbon position (δ)CationC1C2–C6C3–C5C6CH3Ring CH3Relative to TMSδ = 0, the absolute shielding for TMS = 207.12. In all calculations the GIAO MP2/6-31G* level was used, employing the SCI-PCM B3LYP/6-31G** optimized geometries as input.1CC317.4760.5256.2727.3843.761CH332.0457.8125.9722.3946.372CC304.7867.5166.2842.1742.5825.782CH330.7164.4536.5438.6746.8922.813CC323.6457.4966.4535.6344.4425.91, 24.913CH330.8757.0137.7931.2645.2524.60, 30.78As noted in the Introduction, theCC ⇄ CHequilibrium could not be “frozen out”, and so experimentally only a single set of average13C NMR chemical shifts can be obtained. In the case of the1and2systems, the dynamic populations of both theCCandCHisomers are significant, and so the averaged13C shift of the C3–C5 carbons does not closely correspond to that of either pure species. However, cation system3(even though dynamic) was deduced to overwhelmingly favor the3CHisomer, and if so the experimental13C NMR spectra for this cation can be compared to the calculated results for this single isomer, as is shown inTable 4. The agreement for all13C peaks is extremely good; on average the calculated results are 1.4 ppm largerδthan experiment, with the C3–C5 peak differing by 2 ppm. In contrast, calculatedδvalues for the C3–C5 peak of3CCdiffer by 31 ppm from the experimental value.Comparison of calculated and experimental13C NMR shifts for cation3CHCarbon positionExperimentalCalc. for3CHCalc. for3CCδ(ppm).Chair–chair interconversion is very fast in this system and is not being “frozen out”.C1328.7330.87323.64C2–C656.457.0157.49C3–C535.837.7966.45C629.531.2635.63CH3at C143.645.2544.44Ring CH3 27.3 (av.)27.69 (av.)25.41 (av.)The calculated NMR chemical shifts inTable 4were obtained with the GIAO-MP2 method, which in this case is much superior to uncorrelated methods, particularly with respect to the chemical shifts computed for the C3–C5 carbons in theCCisomers. The excellent agreement of calculated and experimental data for3CHalso gives one confidence that the other cations in this series are being calculated to a similar precision.Equilibrium isotope effects—calculated differential energy changes for d4-substitution at C2–C6 in the cationsInTable 5are recorded the calculated differential energy changes brought about by d4-substitution at C2 and C6 in the three cation systems. This isotopic substitution is calculated to produce a significant change in the relative enthalpies (or free energies) of theCHandCCisomeric pairsin favor of theCCisomer, a very similar 1.6–1.8 kJ mol−1in the1–3series. From related calculations in which other isotopomers are used, one can show that the main destabilization effect of the d4-substitution involves the two axial hydrogens at C2–C6 in theCHisomer. These are the same hydrogens which are involved in the hyperconjugative delocalization which defines theCHisomers, and deuterium substitution of these would be expected to destabilize this isomer.Calculated enthalpy changes (163 K) produced by d4-deuteration at C2–C6 in theCCandCHisomers of cation systems1,2and3CationΔH(CC ⇄ CH)-h4ΔH(CC ⇄ CH)-d4 Isotope-induced ΔΔHchangeIn kJ mol−1using the SCI-PCM model with B3LYP/6-311+G**//B3LYP/6-31G** energy and 0.98 ZPVE and thermal data, from B3LYP/6-31G** frequencies. The data for the non-deuterated cations reported here are very similar to the ΔHvalues reported in Table 1, since the thermal correction differences are quite small.Calculated for a temperature of 163 K, a mid-range temperature in the experimental work.C2–C6-d4isomer.1CC7.15.12.01CH0.000.002CC3.61.81.82CH0.000.003CC10.38.32.03CH0.000.00Modeling of the experimental NMR spectra—including equilibrium isotope effectsThe combined data fromTables 2, 3 and 5have been used to model the experimental13C NMR spectra of the1,2and3systems, and this result is shown inFig. 1for a temperature of 163 K.Comparison of calculated and experimental averaged13C NMR shifts for the C3–C5 carbons of cation systems1,2and3, and their d4-substituted analogs, at 163 K. The protocol used to generate this figure required the ΔHcalc. − ΔHexp.value from Table 2 to be set to zero for the parent system1. The corrected “calculated” data for2and3thus become −1.1 and −7.4 kJ mol−1, respectively. The ΔGvalue for all three systems was derived by using the experimental ΔSvalue (CC ⇄ CH) of −7.3 J K−1. Incremental ΔHchanges for the d4-cations were taken from Table 5, and13C NMR data from Table 3. The13C data for the d4-cations have not been corrected for “intrinsic” isotope effects since these are known to be small compared with the large changes shown in Fig. 1. Calculated13C NMR shifts for the individual C3–C5 carbons in theCCandCHstructures are also shown in the figure.The excellent overall theoreticalvs.experimental agreement shown inFig. 1for the2and3cation systems is highly dependent on getting three factors right: (1) the13C NMR shifts have to be correctly calculated; (2) the energy changes produced bycis-3,5- and 4,4-dimethyl substitution of1also have to be correctly calculated; and (3) the equilibrium isotope shifts (direction and magnitude) require a very specific frequency calculation result.Factors (1) and (3) are actually very “structure specific” and we believe that the present theoretical modeling represents auniquefit of experimentalvs.calculated results, and that it is nearly inconceivable that any other model would have fitted the combined experimental data.

ISSN:1472-779X    

DOI:10.1039/b102019f

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionIn 1895, Röntgen discovered that a Cookes tube was emitting a new penetrating form of electromagnetic radiation, which he called the X-ray.1Soon thereafter, Grubbé and associates used X-rays in the treatment of breast cancer.2Thus, began the field of radiation oncology. It would, nevertheless, be another forty-eight years before hydroxyl radical (HO&z.rad;) generation was proposed to be the initiating event in radiation-induced tumor cell killing.3Drawing upon these earlier findings,3we, in 1995, reported thein vivoin situspin trapping of a free radical produced by radiation-induced HO&z.rad; in real time using low frequency EPR spectroscopy.4For these initial investigations, we used the spin trapping system consisting of 4-pyridyl(N-tert-butyl)methanimineN,N′-dioxide1in the presence of EtOH. Hydroxyl radical reacts with EtOH, yielding α-hydroxyethyl radical (CH3&z.rad;CHOH). The latter free radical was subsequently spin trapped by nitrone1, affording the corresponding aminoxyl.5Despite this success, we only detected the HO&z.rad; biomarker in the interstitial space of the tumor. While these studies 4defined radiation-inducedfree radical reactions in this aqueous compartment, intracellular free radical events remain to be defined and await the development of spin traps that can compartmentalize within the tissue. One approach is to use readily hydrolyzable esters of, for instance, 5-carboxy-5-methyl-1-pyrrolineN-oxide2.6Once the ester of 5-carboxy-5-methyl-1-pyrrolineN-oxide2has been taken-up by cells, hydrolysis by esterases therein would entrap nitrone2, as the pKaof this compound is 2.95.6aAfter the spin trapping specificity of 5-carboxy-5-methyl-1-pyrrolineN-oxide2is defined, a reasonable preparative scheme for the synthesis of15N and deuterium-labeled esters of nitrone2,6bas was accomplished with 5,5-dimethyl-1-pyrrolineN-oxide3, could be achieved.7First, however, it is important to define the sensitivity and selectivity of 5-carboxy-5-methyl-1-pyrrolineN-oxide2toward HO&z.rad; and O2&z.rad;−. Herein, we report that 5-carboxy-5-methyl-1-pyrrolineN-oxide2reacts with HO&z.rad; at near diffusion controlled rates, yielding the corresponding aminoxyl6. Using X-ray radiation,we were able to detect HO&z.rad; in a phosphate buffer with as little as 300 nM (1 Gy of radiation) of HO&z.rad;. Furthermore, it was found that nitrone2does not react with O2&z.rad;−. Finally, 5-carboxy-5-methyl-1-pyrrolineN-oxide2can easily be prepared from readily available chemicals, suggesting that isotope labeling is cost effective.

ISSN:1472-779X    

DOI:10.1039/b101945g

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionStructurally modified oligonucleotides, the so-called antisense oligonucleotides, constitute a promising class of chemotherapeutic agents, which enable highly selective inhibition of gene expression. Phosphoromonothioates are the most widely studied oligonucleotide analogs introduced for these purposes.1One of the main hurdles for the use of phosphodiester analogs, such as phosphorothioates, as drugs is the poor cellular uptake of the charged oligomers. A possible way to overcome this problem is the masking of the phosphodiester functions with a protecting group that is removed by the action of intracellular enzymes and rapid subsequent chemical reactions. For example, acylthioethyl (SATE groups)2and acyloxymethyl3groups have been shown to be potential candidates for these purposes. In the present contribution we report on alternative phosphoester protecting groups, derived from bis(hydroxymethyl)-1,3-dicarbonylcompounds and their congeners.4The principle of removal of these groups is outlined inScheme 1. Deprotection of the hydroxymethyl function by intracellular enzymes enables base-catalyzed retro-aldol condensation, accompanied by phosphate elimination giving the desired phosphodiester.To study the structural effects on the kinetics of the non-enzymatic steps of the removal of this type of protecting group, phosphotriesters1a–dwere prepared as model compounds. The dimethoxytrityl group was used in these compounds, instead of an enzymolabile protecting group, since the detritylated compounds2a–dare sufficiently stable under the conditions, where dimethoxytrityl is removed. Accordingly, treatment of1a–dwith trifluoroacetic acid in a mixture of dichloromethane and methanol gave the detritylated compounds2a–d(Scheme 2),the hydrolysis of which could then be studied over a wide pH range (pH 2–7) by HPLC. The methoxyethyl group was used as a mimic for the 5′-O-linked nucleoside, since the methoxyethoxide ion is known to be in phosphoester hydrolysis about as good a leaving group as a nucleoside 5′-oxyanion.5Besides the phosphotriesters1a–d, the phosphorothioate analogs4a,bwere synthesized, and the hydrolytic stability of the detritylated thioates5a,bwas compared with that of the phosphate analog2b.

ISSN:1472-779X    

DOI:10.1039/b101754n

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

MethodologyProton affinities are computed in a standard way [eqn. (1)]:PA(Bα) = (ΔEel)α + (ΔZPVE)α(ΔEel)α = E(B) − E(BαH+)(ΔZPVE)α = ZPVE(B) − ZPVE(BαH+)Eqns. (2) and (3)give the electronic and the zero-point vibrational energy contributions to the proton affinity, respectively. Here B and BαH+denote the base in question and its conjugate acid, respectively, while α stands for the site of proton attack. The search of the Born–Oppenheimer energy hypersurfaces was performed at the economical Hartree–Fock level employing the 6-31G* basis set. The minima on the hypersurface corresponding to optimal geometries are verified by vibrational analyses at the same level. The corresponding frequencies are used in deriving the ZPV energies by applying a common scale factor, 0.89, as usual. The final single point calculations take into account the correlation energy effect at the Møller–Plesset (MP) perturbation level of theory, where the series is terminated after the second order correction. This gives rise to the MP2(fc)/6-311+G**//HF/6-31G* + ZPVE(HF/6-31G*) model. A very important detail of the approach is the use of the 6-311+G** basis set in the final calculation to ensure a proper description of the nitrogen lone pair. Although the MP2 formalism is applicable in quite large systems, the corresponding computations sometimes exhibit convergency problems. It is gratifying that a much simpler scaled Hartree–Fock (HFSC) model25performs almost equally well with only a small sacrifice in accuracy [eqn. (4)] where it is tacitly assumed that the proton attacks a nitrogen atom. The difference ΔEel(HF/6-31G*) refers to the difference in the total molecular energy between a base and its conjugate acid. The high quality of the correlation [eqn. (4)] against the MP2 proton affinities is reflected in the high coefficientR = 0.997 and the low average absolute error 1.3 kcal mol−1.25All calculations are performed by using GAUSSIAN 94 and GAMESS programs.28,29PA(B)N = 0.8924ΔEel(HF/6-31G*)N +  10.4(kcal mol−1)

ISSN:1472-779X    

DOI:10.1039/b101213o

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

ExperimentalSample preparationThe synthesis and characterization followedref. 1for peptide1, andref. 2for peptide2. Single crystals were obtained at ambient temperature by slowly evaporating a solution of peptide1in absolute ethanol and of peptide2in acetonitrile at ambient temperature.X-Ray structure determinationA colorless, prismatic single crystal (0.30 × 0.20 × 0.20 mm3for1and 0.40 × 0.60 × 0.70 mm3for2) was used for collecting three-dimensional X-ray data on a RIGAKU AFC7R diffractometer. The structure was solved by direct methods,15and expanded using Fourier techniques.16All the nonhydrogen atoms were refined anisotropically. Hydrogen atoms of peptide1were refined isotropically. The positions of hydrogen atoms of peptide2were not refined, but isotropicBvalues were refined. The final cycle of full-matrix least squares refinement17onF2was based on 3124 observed reflections[I > 2σ(I)] for peptide1, and 2114 reflections [I > 2σ(I)] for peptide2. The refinement was converged withR = 0.073 (for observed data) andRw = 0.232 (for all data) for peptide1, andR = 0.118 (for observed data) andRw = 0.329 (for all data) for peptide2. The crystallographic details are summarized inTable 1. The molecular graphics were illustrated using molecular modeling software.18Crystallographic details for peptides1and2Parameter12Empirical formulaC36H47N5O8·C2H5OHC41H54N6O9Molecular weight723.86774.91Crystal dimensions/mm30.30 × 0.20 × 0.200.40 × 0.60 × 0.70Crystal system/space groupMonoclinic/P21Orthorhombic/P212121a/Å10.812(2)19.099(3)b/Å19.228(4)19.789(2)c/Å19.324(3)11.980(2)α/°9090β/°101.98(1)90γ/°9090V/Å33929(1)4527.8(9)Z44Density calculated/g cm−31.2231.137Radiation usedCuKα (λ = 1.54178 Å)CuKα (λ = 1.54178 Å)F(000)1552.001656.00Temperature/K223296Scan typeω − 2&thetas;ω − 2&thetas;2&thetas;max/°119.2120.2Observed reflections [I > 2σ(I)]31242114Variables652552Refinement methodFull-matrix least squares onF2Full-matrix least squares onF2Final agreement factorsR = 0.073 (observed data)R = 0.118 (observed data)Rw = 0.232 (all data)Rw = 0.329 (all data)Tables of final positional parameters, equivalent thermal factors, bond lengths, bond angles, and van der Waals contacts for peptides1and2have been deposited in the Cambridge Crystallographic Data Bank as a supplementary publication.CCDC reference numbers 157037 and 157038. Seehttp://www.rsc.org/suppdata/p2/b1/b100774m/for crystallographic files in .cif or other electronic format.Conformational energy calculationIn order to estimate the energetically favored conformations of the N-terminalL-Pro residue in peptide2, an empirical conformational energy calculation was carried out using the structural and energy parameters based on the ECEPP system.19The ECEPP parameters of the ΔZPhe residue were determined in our previous study.14The program PEPCON19–21for obtaining a conformational energy calculation and graphics of a given peptide was modified to be applicable to β-aryldehyroalanine-containing peptides.14,22–24On the basis of the present result and many crystallographic data8for ΔZPhe-containing peptides, all amide groups were fixed to thetransconformation (ω = 180°),and each ΔZPhe side chain was fixed to theZ-configuration (χ1 = 0°). Conformational energy was calculated for Ac-L-Pro-(Aib-ΔZPhe)2-Aib-OMe (Ac = acetyl) with varyingψProvalues (−180° to +180°). Here the Boc group was replaced by an Ac group to simplify the calculation. In the ECEPP system, theuppuckering of the Pro ring (&phis;Pro = −67.6°)19was chosen on the basis of the present data of peptide2. The achiral segment -(Aib-ΔZPhe)2-Aib-OMe was set to a standard left- or right-handed 310-helix:i.e., (&phis;,ψ) = (60°, 30°) or (−60°, −30°),25,26respectively.

ISSN:1472-779X    

DOI:10.1039/b100774m

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionCyclopropenones are of special interest as they are known as stable compounds containing an exocyclic carbonyl group attached to a three-membered ring. The parent compound, cyclopropenone, was first synthesized in 1967,1isolated and subsequently characterized by Breslowet al.2,3Various properties of this compound were also later reported.4–6In most cases, following pyrolysis 7or photochemical decarbonylation,8–16cyclopropenones (I) are converted into acetylenes (II) and carbon monoxide as illustrated inequation (1):Although this kind of reaction is of considerable interest in organic synthesis, its mechanism is not yet well understood. Kresge and co-workers 11–15have indeed employed the reaction (1) as a method to generatein situthe novel class of ynols (II, R2 = OH) and to study the reactivities of these compounds. A number of fundamental questions regarding the decarbonylation of cyclopropenones (I) to acetylenes remain, however, unanswered: (i) The main question is whether the decarbonylation of (I) is a concerted or stepwise process. (ii) Does this reaction take place in the electronic ground state or must it be in an excited state, as required in flash photolytic techniques? 3–6,11–16(iii) How important is the solvent effect on this kind of reaction? (iv) Which factors (e.g., substituents) favor the decarbonylation of cyclopropenones to acetylenes?In an attempt to provide some elements of an answer to those questions posed by experimental results and in relation with our recent theoretical studies on analogous reactions of hydrogen isocyanide with acetylenes 17and also with doubly-bonded dipolarophiles,18–22we have carried out quantum chemical calculations on reaction (1) for three simple cases, the parent system and two fluorinated systems (R1 = R2 = H, R1 = H, R2 = F and R1 = R2 = F) in order to determine the effect on various intermediates being formed in those reactions upon fluorine substitution.Moreover, in order to be able to compare our studies with previous works 17–22and in view of the principle of microreversibility, we have also investigated the reverse reactions of (1), namely the carbonylation of acetylenes. We set out to obtain quantitative information on various structural and energetic aspects of both types of reaction. In the simplest parent case, R1 = R2 = H, we have also constructed hardness and polarizability profiles along the intrinsic coordinate (IRC) pathways and searched for correlations between these electronic properties and the position of the transition structures as well as the calculated energy barriers. The solvent effect in three typical solvents with strongly different relative permittivities including water, acetonitrile and benzene, have also been studied.In contrast to the widespread interest in cyclopropenones, both in theoretical and experimental studies, little is known about the higher homologues cyclopropenethiones (III), also a member of the substituted cyclopropenes group. To extend the scope of the study, we have also performed similar calculations for the combination of C&z.dbd;S and acetylenes, as illustrated inequation (2):

ISSN:1472-779X    

DOI:10.1039/b100709m

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

IntroductionDopamine (DA), the endogenous ligand of dopaminergic neurotransmission systems involved in a broad range of both central and peripheral physiological responses, may be expected to quite freely adopt many conformations, differing only slightly in energy and separated by low potential energy barriers, which interconvert rapidly at body temperatures. This behaviour has militated against any definite conclusions being reached regarding which rotamer(s) is (or are) involved in the elicitation of its pharmacologic actions, in spite of a fair number of X-ray,1,2NMR,3,4and theoretical studies.5–9Experimental data on the free energy, enthalpy and entropy of the DA–receptor interaction show that receptor binding is dominated by a favourable enthalpy change and that the entropic change is unfavourable.10These results indicate that an enthalpically favoured binding mode could help the DA molecule to populate relatively energy-rich conformations. The use of semirigid DA analogues has led to the conclusion that one or both of thetrans-coplanar rotamers (cf.Fig. 1) appear to be involved in the activation of different types of DA receptors, but the possible rôle of each of these conformers is not yet clear.11The different rotamers for theN-protonated forms of the compounds: numbering and definition of the dihedral angles (&thetas;and&phis;) used in the conformational study. (a) Perpendicular rotamers correspond to the staggered conformations resulting from rotation about&phis;with&thetas; = 90°, (b)trans-coplanar rotamers, (c)gauche-coplanar rotamers.Studies on structure–activity relationships in open-chain DA analogues and in phenylethylamines generally as dopaminergic ligands have shown that allN-monosubstituted analogues are less active than DA at D1receptors, with the exception ofN-methyldopamine (epinine), which is equipotent with the endogenous neurotransmitter.12In contrast, in these and in theN,N-disubstituted derivatives, the well-known “N-n-propyl phenomenon”—an increase in affinity for the D2receptor when the nitrogen atom bears ann-propyl group—is apparent.13A hydroxy groupmetato the amine side chain, but not a catechol system, is widely believed to be essential for agonist activity.14Ring fluorination of DA at C-2 or C-5 affords analogues which differ little, if at all, from DA as dopaminergic agonists, while 6-fluorodopamine seemed to be somewhat less potent in a dog renal vascular assay.15In binding studies, all three fluorodopamines are as potent as dopamine in displacing [3H]spiperone (a selective D2receptor antagonist) but the 2- and 6-fluoro analogues are less potent than DA or 5-fluorodopamine in displacing [3H]apomorphine (a relatively unselective agonist which is equipotent with DA at D1receptors and several times more so at D2receptors).16The uptake of these analogues into synaptic vesicles has led to the use of [18F]-fluorinated DAs as false neurotransmitters for positron emission tomographic imaging.17It therefore seems that introduction of the small, strongly electronegative fluorine atom at C-6 (and possibly C-2, but not C-5) interferes with the ability of DA to activate one or both of its major receptor types, but has little effect upon the recognition of the DA molecule by its neuronal receptors or transporters.The introduction of the sterically more demanding, but electronically less disruptive, methyl group yields 2-methyldopamine (2-methylDA), equipotent with the parent molecule, and 6-methyldopamine (6-methylDA), two orders of magnitude less active in stimulating DA-sensitive adenylate cyclase from rat caudate nucleus, a clear reflection of D1receptor activation (seeFig. 2).11This different behaviour may be tentatively attributed to a direct steric effect of the C-6 (but not C-2) methyl group hindering the drug–receptor interaction, or to an indirect effect arising from destabilization of the pharmacophoric conformation. α-Methyldopamine (α-methylDA) has greatly decreased dopaminergic agonist effects or is quite inactive in some assays.18This has been explained as either a steric effect forcing the side chain into a presumably unfavourable perpendicular relationship with the catechol ring,18or a direct steric hindrance to interaction with the receptors,19but neither of these hypotheses has been examined with quantitative tools. The issue is further confused by the fact that several hydroxylated 2-amino-1,2,3,4-tetrahydronaphthalene (2-aminotetralin) derivatives, which incorporate an α-methyl group into a semirigid ring system, are potent dopaminergic agonists.11,14Structures of the compounds studied.We have now addressed this problem in terms of a conformational analysis of DA, its α-, 2- and 6-methyl derivatives and theirN-protonated conjugate acids, using the AM1 semiempirical method 20and COSMO (conductor-like screening model) 21to evaluate the effects of an aqueous medium on conformational preferences. A study of theN-protonated forms is justified by the fact that DA exists as its conjugate acid to an extent exceeding 95% at physiological pH.22Both polar and apolar media are relevant to the conformational profiles of ligands interacting with biological targets such as dopaminergic receptors, because drug molecules must go from an aqueous extracellular environment to a relatively apolar binding site in the receptor protein.Fig. 1depicts the rotamers of theN-protonated forms of the compounds studied in this paper and the dihedral angles&thetas;and&phis;that define the aminoethyl chain conformations.

ISSN:1472-779X    

DOI:10.1039/b100614m

出版商:RSC    

年代:2001

数据来源: RSC

摘要:

Results and discussionSynthesis of cavitands7,8,9,10,11,12,13and14Tetrakis(bromomethyl)cavitands4aand4b 10easily obtained from the corresponding tetramethylcavitands4(X = CH3) by NBS bromination (80%) were reacted with an excess of imidazole, pyrazole, indazole or benzimidazole in CHCl3to give hosts7a(55%),7b(77%),8(65%),9(78%), or10(30%) in good yields (Fig. 1). Host11was obtained from a tetraol5 11and 2-(chloromethyl)pyridine (85%). Cavitands12(50%),13(50%), and14(46%) were obtained from tetrakis(p-chloromethyl)calix[4]arene6 12in CHCl3or CH3CN with an excess of imidazole, pyrazole and indazole, respectively. The new hosts were purified by washing exclusively with deionized water several times followed by repeated recrystallizations from a mixed solution.Cavitands4–14.1H NMR spectra of cavitands12–14show that they exist as cone conformers, as structurally rigid cavitands7–11do, but the binding studies of cavitands10and12cannot be pursued due to their low solubility in CHCl3.Binding stoichiometry and picrate extraction experimentsQuite a few salt binders based on resorcin[4]arene have been developed,13because typical container hosts constructed on resorcin[4]arene do not bind salts.14However, quite a lot of those based on calix[4]arene have been reported due to the conformational and chemical diversity of calix[4]arene.15The 1 ∶ 1 complexation between NaClO4and host7bwas confirmed by solid–liquid extraction (Fig. 2) which showed the chemical shift saturations at a 1 ∶ 1 ratio of solid NaClO4to host7bin CDCl3–CD3CN = 3 ∶ 1 (v/v) at room temperature. The1H NMR data for imidazolylcaviplex7b·Na+ClO4−(400 MHz) suggest all four imidazolyl groups are coordinated to sodium. Compared to the free7b, the resonances of the imidazole unit and methylenedioxy protons in the complex showed large upfield chemical shifts (NCHN, 0.28 ppm, from 7.53 to 7.25 ppm; one of NCHCHN, 0.02 ppm, from 6.93 to 6.91 ppm;exo-OCH2O,0.19 ppm, from 6.02 to 5.83 ppm;endo-OCH2O, 0.23 ppm, from 4.20 to 3.97 ppm). A computer-generated (HyperChem with MM+ force field) stereo view of imidazolylcaviplex7a·Na+in the gas phase (Fig. 3) shows its nesting binding mode.1H NMR spectroscopic titration with Na+ClO4−of imidazolylcavitand7bin CDCl3–CD3CN = 3 ∶ 1 (v/v).Stereo view of the energy minimized structure of imidazolylcaviplex7a·Na+(Hyperchem with MM+ force field).The free energies of complexation (−ΔG° in kcal mol−1) were determined at 22 °C in H2O-saturated CHCl3by Cram’s liquid extraction method 16on a 1 or 4 mM scale host solution binding Li+, Na+, K+, Rb+, Cs+, NH4+, CH3NH3+, and (CH3)3CNH3+picrates. Among the cavitands, except cavitands10and12which are very insoluble in CHCl3, only hosts7showed substantial affinities for alkali or ammonium picrate and the results are summarized inTable 1(average of two determinations whose difference was <0.5 kcal mol−1). Imidazolylcavitand7atends to precipitate in the organic phase (CHCl3) on the 4 mM scale, but imidazolylcavitand7bis soluble on the 4 mM scale. At the 1 mM scale imidazolylcavitands7aand7bshowed similar tendencies for alkali and ammonium metal picrate extraction, experimental differences being within 0.5 kcal mol−1.Apparent association constant (Ka/M−1) and binding free energy (−ΔG°/kcal mol−1) for complexation of imidazolylcavitand7bwith alkali metal, ammonium and alkylammonium picrates in CHCl3saturated with H2O at 25 °CLi+Na+K+Rb+Cs+NH4+CH3NH3+t-BuNH3+The values are average values from organic and aqueous phases of two trials at the 4 mM scale whose difference was <0.5 kcal mol−1.Ka/M−13.1 × 1071.3 × 1081.3 × 1071.6 × 1075.9 × 1066.5 × 1062.1 × 1061.7 × 105−ΔG°/kcal mol−110.111.09.59.49.29.28.57.0Fig. 4illustrates the host–guest structure–binding relationships for complexation of eight cations and imidazolylcavitand7b. For a comparison the binding free energies of octacyanand1having two rigidly preorganized binding sites solely composed of four sp nitrogens for each are also shown. It is reported that octacyanand1showed 1 ∶ 1 binding at 10−3M solution (CHCl3) in a perching fashion.Comparison of binding free energy (ΔG°) between octacyanand1and imidazolylcavitand7b.Imidazolylcavitand7bshowed unexpected high binding energies of −ΔG°av = 9.2 kcal mol−1ranging from 7.0 for (CH3)3CNH3+to 10.9 for Na+with a maximum spread of 3.9 kcal mol−1. Imidazolylcavitand7btends to bind spherical cations better: NH4+a little better than CH3NH3+and CH3NH3+somewhat better thant-BuNH3+showing a peak binding for Na+. Imidazolylcavitand7bhas lower binding energies [Δ(−ΔG°) = −0.5 for Li+to −4.6 kcal mol−1for K+] than octacyanand1.Compared to octacyanand1, imidazolylcavitand7bhas a much less preorganized binding site due to the free rotation of the imidazolylmethyl arms, which also caused its relatively low selectivity.1H NMR titration experimentsBinding studies were also carried out using the chemical shift change caused by incremental addition of Na+BPh4−to imidazolylcavitand7ain CDCl3–MeOH-d4(7 ∶ 1, v/v) at 298 K.Fig. 5shows the observed chemical shift changes of three different hydrogens (Hm,exo, H4′and Hb) of7avs.guest–host ratio.Observed chemical shift changes (Δδ) of the three different hydrogens of7avs.[G]/[H] ratio.The protons whose signals were chosen forKadeterminations exhibited a relatively large chemical shift change upon complexation (>0.05 ppm). The sensitivity of H4′and Hm,exowere similar to those from solid–liquid extraction, but the large sensitivity of Hband the weak sensitivity of Hm,endoare unexpected. TheKavalue calculated using a Benesi–Hildebrand plot 17for cavitand7abinding Na+at 298 K was 541 M−1.The probable anion effect was also tested. A small quantity of host7awas dissolved in CDCl3–MeOH-d4(7 ∶ 1, v/v) to be about 1 mM, and its 300 MHz1H NMR spectrum was recorded. To this solution about 20 equivalents of Bu4N+BPh4−was added and then the1H NMR spectrum was recorded. Chemical shift changes did not occur for H-4′ or Hm,exobut were found for Hb. However, its change (<0.05 ppm) is very much smaller than that with Na+BPh4−(>0.26 ppm at 5 eq.), which supports the theory that the chemical shifts were induced by cation binding to host.Affinities for alkaline earth and transition metal picratesSolvent extractions of aqueous alkaline earth or transition metal cations into H2O-saturated organic host solutions were performed at 25 °C. An aqueous solution containing M(NO3)n(10 mM) and picric acid (1.0 mM) was extracted with the host solution (CHCl3, 1.0 mM).The percentage extraction (Ex%) was calculated by measuring the picrate concentration in the aqueous phase. The results are summarized inTable 2andFig. 6. Imidazolylcavitand7ashowed the highest extractabilities for IIA and transition metal cations with the highest for Ag+(100%). The ionic radius of Ca2+(0.99 Å) is very similar to that of Na+(0.96 Å), which may explain the higher Ex% (57vs.43%) for Ca2+than for Mg2+(the ionic radius of Mg2+is 0.65 Å).Percentage extraction of metal picrates by cavitands7a,8,9,11,13and14at 298 KEx%HostAg+Mg2+Ca2+Co2+Ni2+Cu2+Zn2+Cd2+Organic phase (2 mL of CHCl3) contains cavitand (1.0 mM) and aqueous phase (2 mL) contains picric acid (1.0 mM) and M(NO3)n(10 mM). The two-phase mixture was stirred for 1 min and centrifuged for 1 min and then Ex% values were determined spectrophotometrically.Ref.18.7a1004357474755(56)4931893000390092200011400117481052730901320000000014250000000Illustration of percentage extraction of metal picrates by cavitands7a,8,9,11,13and14at 298 K.Pyridinylcavitand11also showed a high affinity for Ag+(74%), but overall it showed lower affinities for other metal cations, compared to those of cavitand7, due to its higher flexibility. The Ex% (<25%) for Ag+by calix[4]arene-based cavitands13or14is insignificant, but that (93%) by pyrazolylcavitand8was exceptional.The remarkably high selectivity by pyrazolylcavitand8for Ag+over Mg2+, Ca2+, Co2+, Ni2+, Cu2+, Zn2+and Cd2+and the CPK molecular model study suggest that the four lone pair electrons of the N-2 atoms of host8can converge to form a binding site for a spherical cation in an inclusion fashion. π-Base receptors which feature cation–π interactions display a high metal affinity,19which is true for8·Na+. Computer-generated (HyperChem with MM+ force field) stereo views of imidazolylcaviplex8·Na+(Fig. 7) also support its inclusion binding mode. Pyrazolylcavitand8showed much more efficient silver extraction ability (93%) than pyridine-containing cage2(19.5%) which complexes with a silver ion by a linear coordination between the two adjacent pyridine donor atoms.7Indazolylcavitand9could bind in a similar binding mode, but compared to8, the rotation of indazolyl arms to permit an inclusion mode seemed to be much more sterically hindered, which resulted in low affinities for the tested metal cations, even for Ag+.Stereo view of energy minimized structure of pyrazolylcaviplex8·Ag+(Hyperchem with MM+ force field).

ISSN:1472-779X    

DOI:10.1039/b100490p

出版商:RSC    

年代:2001

数据来源: RSC