Nitrogen inversion process in some camphor-based isoxazolidines

A series of isoxazolidines has been prepared via asymmetric nitrone cycloaddition reactions of a camphor-based methylenenitrone with several alkenes. The 1 H NMR spectra of the N(2) and C(5)-disubstituted isoxazolidines in CDCl 3 at low temperatures revealed the predominant or exclusive presence of the diasteromer having trans disposition of the substituents. The presence of hydroxyl substituent on the camphor moiety permits H-bonding with the N or O of the isoxazolidine ring. The effect of H-bonding - intramolecular in CDCl 3 and intermolecular in CD 3 OD - on the population ratio of the invertomers and nitrogen inversion process has been investigated. The nitrogen inversion barriers are determined using complete line-shape analysis, and their dependence on solvent is discussed.


Introduction
The compounds containing nitrone functionality (-C=N + -O -) undergoes cycloaddition reaction with an alkene to produce isoxazolidines having an −N−O− moiety embedded in the ring skeleton. 1The presence of an −N−O− moiety in an organic molecule has a distinctive place in conformational analysis, [2][3][4] oxygen being next to nitrogen raises the barrier to nitrogen inversion (N i ) to such an extent that at temperatures lower than the ambient, individual invertomers can be identified by NMR spectroscopy. 5The asymmetric nitrone cycloaddition reactions involving camphor-derived intramolecularly H-bonded nitrone (Scheme 1) has been found to be very effective in transferring chirality to the newly created stereocenter at C(5) of the cycloadducts isoxazolidines. 6owever, it is often difficult to assign the stereochemistry of the substituted isoxazolidines by spectroscopic analysis owing to complications arising out of the relatively slow nitrogen inversion process as well as pseudorotation in a five-membered ring.We have prepared a number of 5-mono-, and 5,5-disubstituted isoxazolidines 4 and 5 of known configurations 6 using nitrone (2)-alkene (3) cycloaddition reactions to examine the conformational aspects as well as nitrogen inversion process by NMR spectroscopy (Scheme 1).The study would provide us with the opportunity to examine the effect of H-bonding -intramolecular in CDCl 3 and intermolecular in CD 3 OD-on the population ratio of the invertomers.

Results and Discussion
Since nitrone 2 is optically pure, the isoxazolidines 4 and 5 differ only in the configuration of the C(5) substituents.During the course of a structural investigation of the isomeric isoxazolidines 4c and 5c (R 1 =H, R 2 = CO 2 Me) it was interesting to observe the presence of a single invertomer for 4c, while 5c remained as an equilibrating mixture of two invertomers (cis and trans) at -40°C both in CDCl 3 and CD 3 OD (Scheme 2).This is interesting since a look at their structures seems to covey that 2,5-trans-4c and 2,5-trans-5c (or 2,5-cis-4c and 2,5-cis-5c) should have a comparable steric environments, and as such a large discrepancy in the population ratio of the invertomers in 4c and 5c is quite unexpected.The current study helped us to provide a rationale for this observation (vide infra).The nitrogen inversions barriers are determined using NMR band shape analysis.Slow nitrogen inversion in most of the isoxazolidines has been observed to give broadened peaks in 1 H and 13 C spectra recorded at ambient temperature.On lowering the temperature, the spectral lines become sharper and show two distinct forms of the compound.
The 13 C chemical shifts in CDCl 3 and CD 3 OD were assigned on the basis of DEPT experiment results, general chemical shifts arguments and consideration of substituent effects, and are given in Tables 1 and 2, respectively.Around -10°C, the 1 H NMR spectra of these compounds show well separated signals for the two invertomers.Integration of the relevant peaks gives the population trends in these systems.The proton spectra were used in the calculation of barriers in all compounds.The complete band shape analysis yielded the rate constants and the free energy of activation using Eyring equation.The activation parameters ∆H ≠ and ∆S ≠ were calculated from plots of ln(k/T) vs. 1/T.It is well known 7 that NMR band shape fitting frequently gives rather large but mutually compensating errors in ∆H ≠ and ∆S ≠ and as such their values are not reported here.However, band shape fitting is viewed as a method of getting rather accurate values of ∆G ≠ (within ± 0.3 kJ/mol) in the vicinity of the coalescence temperature. 7The ∆G ≠ values calculated at 0 o C are reported in Table 3, along with the invertomer ratios and ∆G o values.
The conformation of 5-membered ring system is indeed very complex to elucidate with some certainty.The complexity arises from the fact that changing the size of the substituent may lead to change in conformation (half chair/envelope/near planar) and the flap of the envelope.Earlier works 8,9 on 2,5-disubstituted isoxazolidines revealed the trans-invertomer as the major isomer.The 2-methyl-, 2-isopropyl-, and 2-t butyl-5-t butyldimethylsiloxymethylisoxazolidines were found to have the transand cis-invertomers in a ratio of 53:47, 55:45 and 63:37, respectively.The compounds studied in this work are sterically similar to the 2-isopropylisoxazolidines since they also contain a secondary alkyl (i.e.camphor moiety) substituent at the N(2)-position (Scheme 2).Note that the isoxazolidines 4a-c in CDCl 3 exist almost as a single invertomer in each case (Table 3), and display sharp NMR signals at ambient as well as lower temperatures, while compound 5c revealed the presence of two invertomers at the lower temperatures.The analysis below will help us to provide a rationale for identifying the major invertomers as having the trans configuration.The likely configuration of the invertomers of the disubstituted isoxazolidines is shown in Scheme 2. The camphor 'H' (attached to the carbon adjacent to N) is placed anti to the nitrogen lone pair in all the configurations since it will have the lower number of gauche interactions (two in these cases) around the C-N bond.The H-bonding with the nitrogen lone pair is possible in all the configuarations having camphor 'H' placed gauche to the nitrogen lone pair, but it will be in a less favourable 5-membered ring structure and will develop considerable eclipsing between the camphor moiety and N-C(3) bond of the isoxazolidine ring.This would leave the possibility of H-bonding with the ring oxygen in 2,5-trans-4 and 2,5-cis-5 invertomers as shown in the Scheme 2. The 2,5-cis-4 and 2,5-trans-5 can not form H-bonded structure with the ring oxygen.The 2,5trans-4 enjoys two advantages -sterically favored trans disposition of the substituents as well as H-bonding -rendering it overwhelmingly favored over the 2,5-cis-4 and as such it exists as the sole invertomer.For the disubstituted isoxazolidines 5, each invertomer enjoys one advantage; while the cis-5 is H-bonded with the ring oxygen, the trans-5 enjoys the trans disposition of the substituents.As a result, both the invertomers exist for the isoxazolidines 5c in a 90:10 ratio for the trans and cis invertomers, respectively.Note that the overall steric environment provided by the camphor moiety is presumably much bulkier than a t-butyl group and as such the steric crowding in 2,5-cis-5 outweighs whatever advantage the H-bonding has to offer.
The situation becomes complicated in the trisubstituetd isoxazolidines 4d and 4e.We believe the H-bonded forms remain to be the exclusive or overwhelmingly predominant invertomers (Scheme 3).While the isoxazolidine 4d exists in H-bonded and non H-bonded forms in a 87:13 ratio, the corresponding stereoisomer 5d remains exclusively in the H-bonded form.The Hbonded form in 5d presumably enjoys an extra stabilization as a result of an additional Hbonding with the cis-disposed C(5) CO 2 Me group (vide infra).

Scheme 3
The stereochemistry of adduct 4c has been earlier confirmed by crystallographic analysis (Scheme 4).In methanol, the intramolecular H-bonding is disrupted; the steric bulk of the solvation shell of the nitrogen lone pair increases in hydrogen-bonding solvents.This should diminish the preference for the trans-invertomers in CD 3 OD as observed for most of the isoxazolidines (Table 3).However, the isoxazolidines, under study, have so many lone pairs on N and O involved in H-bonding that it may sometimes have unpredictable influence on the invertomeric ratios.
The chemical shift difference between the isomers for a particular ring carbon is generally less than 1 ppm for most carbons and as such the C-13 shifts are not very sensitive to the difference in the isomeric conformations (Tables 2 and 3).However the C-3 signals appeared upfield in the trans (Scheme 2) and H-bonded forms (Scheme 3) by over 2 ppm thus providing evidence that the major invertomers have the similar configuration.The corroborative findings of X-ray crystallography, study of invertomeric compositions in different solvents, the consideration of steric crowding in the cis invertomers, as well as upfield C-3 chemical shifts in the trans invertomers provided the evidence in identifying and distinguishing the configuration of the invertomers (vide supra).
The nitrogen inversion barrier is expected to be high when an oxygen atom is directly attached to the nitrogen as in isoxazolidines. 4,10The inversion barrier decreases to some extent in hydrogen bonding solvent CD 3 OD (Table 3).Usually it is known that the inversion barriers in CD 3 OD increase in cyclic system as a result of an extra energy required for breaking of Hbonding prior to inversion. 4However, the inversion in CDCl 3 may also require extra energy to break the intramolecular OH … N hydrogen bond prior to inversion.
The NMR study at lower temperatures has thus successfully led to the determination of the stereochemistry of the nitrogen invertomers in camphor-derived isoxazolidines.The identification of nitrogen lone pair stereochemistry will indeed be a valuable information in per acid induced ring opening of the isoxazolidines. 11

Experimental Section
Compounds studied A total of 7 compounds have been studied in the current work.The structures of these compounds are given in Scheme 1.

Physical methods
All mp are uncorrected.The variable temperature 1 H NMR spectra were recorded on a JEOL Lambda NMR spectrometer operating at 500.0 MHz.Most of the compounds were studied as 25 mg/cm 3 solutions in CDCl 3 and CD 3 OD with TMS as internal standard. 13C NMR spectra were measured in CDCl 3 or CD 3 OD using TMS as internal standard on a JEOL LA 500 MHz spectrometer.Multiplicities of the carbons were determined using DEPT experiments.Elemental analysis was carried out on a EuroVector Elemental Analyzer Model EA3000.I.r.spectra were recorded on a Perkin Elmer 16F PC FTI.R spectrometer.X-ray crystallographic analysis was carried out on a Bruker-AXS Smart Apex system equipped with graphite-monochromatized Mo-Kα radiation (λ = 0.71073 Å).Silica gel chromatographic separations were performed with Silica gel 100 from Fluka Chemie AG (Buchs, Switzerland).Paraformaldehyde from Fluka were used as received.Dimethyl methylenemalonate were prepared as described in the literature. 12he chiral 3-(hydroxyamino)borneol (1) was prepared as described . 13,14oxazolidines All the isoxazolidines except 4e were prepared using procedure as described. 6The compound 4e was prepared by cycloaddition reaction of nitrone 2 with dimethyl methylenemalonate (3e) (vide infra).The low temperature 1 H and 13 C NMR data of these compounds are given below.

Isoxazolidine 4c
Both the 1 H and 13 C NMR spectra in CDCl 3 revealed the presence of a single invertomer.The spectra at +20 °C and -40 °C were almost identical.

Isoxazolidine 5c
The 1 H NMR spectrum in CDCl 3 at -40 °C revealed the presence of two invertomers in a 90:10 ratio as determined by integration of several proton signals.

Isoxazolidine 4d
The 1 H NMR spectrum in CDCl 3 at -40 °C revealed the presence of two invertomers in a 87:13 ratio as determined by integration of several proton signals.

Isoxazolidine 4e
To a solution of the hydroxylamine 1 (556 mg, 3.00 mmol) in CHCl 3 (15 cm 3 ), was added paraformaldehyde (150 mg, 5.0 mmol) and the mixture was stirred using a magnetic stir bar at 45 °C for 4 h.Thereafter, MgSO 4 (~ 2 g) was added followed by dimethyl methylenemalonate (3) (0.506 g, 3.51 mmol).After stirring in a closed vessel at 50 °C for 6 h, the reaction mixture was filtered, concentrated and the residual liquid was chromatographed over silica using ether/hexane In CD 3 OD (-40 °C) the 1 H NMR spectrum revealed the presence of several nonoverlapping minor signals indicating the major/minor invertomers of 4e in a ratio of 94:6.While the methyl singlets of the major invertomer appeared at δ 0.79, 0.94 and 1.08 ppm, the non overlapping methyl singlets of the minor invertomer appeared at δ 0.85, and 1.25 ppm.The major invertomer displayed a proton at δ 2.90 (1H, d, J 6.7 Hz) and another proton at 3.67 (1H, d, J 6.7 Hz) ppm.The corresponding protons for the minor invertomer appeared at δ 2.74 (1H, d, J 6.7 Hz) and 3.53 (1H, d, J 6.7 Hz).Major invertomer.δ C (CD

Inversion barrier calculations
The variable temperature 1 H NMR spectra were recorded on a JEOL Lambda NMR spectrometer operating at 500.0 MHz.Most of the compounds were studied as 25 mg/cm 3 solutions in CDCl 3 and CD 3 OD with TMS as internal standard.Simulations of exchange-affected proton spectra for all compounds were carried out using a computer program AXEX, 15 corresponding to a two non coupled sites exchange with unequal populations.The following signals were utilized: 4a, (CD 3 OD), methyl singlets at δ 1.13 (major) and 1.19 ppm (minor); 4b, (CDCl 3 ), C(5)H at δ 4.97 (minor) and 5.18 ppm (major); 4b, (CD 3 OD), methyl singlets at δ 0.79 (minor) and 0.81 ppm (major); 5c, (CDCl 3 ), methyl singlets at δ 0.78 ppm (major) and 0.81 (minor); 5c, (CD 3 OD), methyl singlets at δ 3.69 (minor), 3.73 ppm (major); 4d, (CDCl 3 ), methyl singlets at δ 3.75 (minor) and 3.79 ppm (major); 4d: (CD 3 OD) methyl singlets at δ 3.66 (major) and 3.72 ppm (minor); 4e, (CD 3 OD), methyl singlets at δ 0.85 (minor) and 0.94 ppm (major).Simulations of exchange affected triplets were carried out by modifying the two-site exchange program. 15The first order coupling to these protons is simply assumed as giving overlapping two site exchanges with the same population ratio and equal rates of exchange.Simulations of exchange affected doublet of doublets were carried out by modifying the two-site exchange program.5a The first order coupling to these protons is simply assumed as giving overlapping two site exchanges with the same population ratio and equal rates of exchange.

Table 1 .
13C NMR Chemical Shifts of compounds Studied in CDCl 3 at -30°C

Table 2 .
13C NMR Chemical Shifts of Compounds Studied in CD 3 OD at -40°C

Table 3 .
Free Energy of Activation (∆G ≠ ) for nitrogen inversion, Ratio of the Invertomers, and standard Free Energy Change (∆G o ) for major ⇌ minor isomerization in CDCl 3 and CD 3 OD a At 0 o C. b At -40 o C. *The other isomer was not detected.