Faculty of Chemistry, University of Belgrade, Studentski trg 12-16, P. O. Box 158, YU-11001 Belgrade
Center for Chemistry, ICTM, P. O. Box 815, 11001 Belgrade, Yugoslavia
Abstract
Acid-catalyzed reaction of the steroidal 
1-unsaturated 3,5-epoxyimino compound 2 and 3-unsaturated 1,5-epoxyimino products 3 and 7, results in intramolecular rearrangement
involving the N-CH3 group to give the corresponding perhydro-3,1-oxazine derivatives 9-11.
Under similar reaction conditions, the saturated analogues 4, 6 and 8 remain unchanged. The
difference in reactivity between the unsaturated and saturated compounds is studied and
elucidated by the semiempirical molecular orbital MNDO-PM3 method.
Keywords: isoxazolidines; perhydro-oxazines; seco steroids; molecular rearrangement.
Introduction
In a preliminary communication [1] we described briefly a new type of acid-catalyzed
rearrangement of the steroidal isoxazolidines, i.e., N-methyl-3,5-epoxyimino-5-cholest-1-ene
(2), N-methyl-1,5-epoxyimino-5-cholest-3-ene (3) (Scheme 1) and N-methyl-1,5-epoxyimino-19-nor-5-androst-3-en-17-yl acetate (7) (Scheme 2) to the corresponding
perhydro-3,1-oxazine derivatives 9-11, respectively (Scheme 3). In the present paper we wish to
report more extensively on the observed transformation of these 1- and 3-unsaturated
substrates and also on the behaviour of the respective saturated analogues (compounds 4, 6 and 8,
Schemes 1 and 2), when subjected to similar acid-catalyzed conditions.
The isoxazolidines 2 and 3 were prepared (in 56% and 30% yield, respectively) [2], by heating
(Z)-3-acetoxy-5,10-secocholest-1(10)-en-5-one (1) with N-methylhydroxylamine hydrochloride
in ethanol-pyridine (1:1, v/v) solution at reflux for 24 h (Scheme 1), while the corresponding 19-norandrost-3-ene analogue 7 (Scheme 2) was obtained (in 43% yield) under similar conditions
starting from (Z)-3,17-diacetoxy-19-nor-5,10-secoandrost-1(10)-en-5-one (5); the minor
product of the latter reaction being the saturated 3-acetoxy isoxazolidine 6 (isolated in 29%
yield) [3].
A 3-unsaturated isoxazolidines 3 and 7 are formed by an intramolecular, 1,3-dipolar cycloaddition
of the C(5)-nitrone group of the intermediate ii and iv, respectively, to the corresponding
transannular 1(10)-double bond (the reaction proceeding via i and iii produced by acetic acid
elimination from the (Z)-5,10-secosteroidal ketones 1 and 5), while 1-isoxazolidine 2 arises
from an intramolecular process in which the C(5)-nitrone function and both the 1(10)- and
3-double bond are involved (for details see Ref. 4). On the other hand, isoxazolidine 6 is
product of transannular C(5)-nitrone 1,3-dipolar cycloaddition to the 1(10)-bond of 5 which takes
place without acetic acid elimination.
Acid-catalyzed reactivity of N-methylisoxazolidines 2-4 and 6-8
Acid-catalyzed reaction of isoxazolidines 2, 3 and 6 was carried out in boiling toluene solution (6 mM) with p-toluenesulfonic acid (20 mol%) for 48 h. Under these conditions all three substrates underwent an intramolecular rearrangement involving their N-CH3 group, to give (Scheme 3) the perhydro-3,1-oxazine derivative 9-11 (in 42-54% yield); the residue being the recovered starting material (17-20%) and a complex mixture.
The structure of the obtained products 9-11 was established as follows. In their 1H-NMR spectra the orginal N-CH3 group (singlet at 2.60 ppm) was missing. Instead, two doublets appearing between 4.18-4.32 ppm and 4.65-4.82 ppm, respectively, assingnable to the -CH2- group between oxygen and nitrogen, indicated the presence of the -O-CH2-NH-C(5) fragment. The N-H group appeared in the IR spectra of compounds 9-11 as a new absorption at 3300 cm1. Besides, the number of the primary, secondary, tertiary and H-free C-atoms detectable in the DEPT 13C-NMR-spectra of these compounds [for 9 and 10, 5 CH3, 11 CH2, 9 CH (of which 2 olefinic and 1 bearing oxazine oxygen) and 3 H-free C-atoms, and for 11, 2 CH3, 8 CH2, 8 CH (of which 2 olefinic and 1 bearing oxazine oxygen) and 3 H-free C-atoms] are consistent with the proposed structures. For additional spectral characteristics confirming the structures 9-11 see Experimental.
The results obtained with the unsaturated isoxazolidines 2, 3 and 7 prompted us to investigate the
synthetic possibilities of the described transformation using as substrates the corresponding
saturated analogues, i.e., N-methyl-1,5-epoxyimino-5-cholestane (4) (obtained by diimide
reduction of the 1-unsaturated derivative 2, Scheme 1), N-methyl-1,5-epoxyimino-19-nor-5-androstane-3,17-diyl diacetate (6), as well as the corresponding 3,17-dihydroxy derivative
(8) (obtained from 6 by alkaline hydrolysis, Scheme 2). However, when 4, 6 and 8, respectively,
where subjected to similar acid-catalyzed conditions as above, i.e., heating with p-toluenesulfonic acid in boiling toluene for 48 h, neither of these compounds was transformed to
the expected perhydro-1,3-oxazine (recovery of the starting material being 68-77% (see
Experimental), while the residue was a complex mixture). This indicated that the presence of the
olefinic 1- or 3-double bond in the steroid ring A is indispensable for the reaction to occur.
The general course of the investigated transformation is presented in Scheme 4.
It was anticipated [1] that the reaction is initiated by protonation
of the oxygen(3) in the epoxyimino bridge (to
give oxonium ion B), followed by cleavage of the O-N bond (to
produce species C), and subsequent proton elimination involving, as the final step,
intramolecular cyclization in which participate the imine function of the obtained intermediate D
and the newly formed hydroxyl group.
In an attempt to get a mechanistic explanation of the observed difference in reactivity between the unsaturated isoxazolidinines 2, 3 and 6 and their saturated analogues 4, 7 and 8, semiempirical M.O. methods have been applied.
As model compounds the saturated and unsaturated cis-decaline-3,5- and 1,5-isoxazolidine
derivatives I-IV (Figure 1) were selected.
Method of calculation
The structures of model compounds (I - IV) were generated by PC MODEL, version 4.0 [5], that involves an MMX force field [6,7] and were saved as MOPAC [8,9] files for PM3 semiempirical calculations [10-11].
In our work we used the MNDO-PM3 method that proved to be highly reliable for investigating
molecular properties of molecules and ions [8-20]. We used the MOPAC program package,
Version 7.01. The geometries of all molecular species correspond to the energy minima in a
vacuum and were optimized by the PM3 method. The transition states for all the reactions were
found using the corresponding MOPAC facilities (TS, SADDLE). When needed, the obtained
structures were refined by Bartel's method (Non-Linear Least Squares gradient minimization
routine - NLLSQ), and transition states are further proved by vibrational analysis showing only
one negative vibration. The influence of the solvent to the cations was not studied, because all
the studied reactions were experimentally done in the non polar solvent.
Results and Discussion
In calculations the following intuitive models have been considered:
(i) Bond strain imposed to the isoxazolidine ring upon introduction of the olefinic double bond into the steroid ring A. In the saturated systems epoxyimino bridge is attached to the diaxial 3,5- (in 4) or 1,5-bonds (in 6 and 8), in contrast to the unsaturated molecules in which the -oriented bonds next to the double bond (i.e., 3- in 2 and 5- in 3 and 7) assume pseudoaxial (a') orientation. Therefore, in unsaturated substrates epoxyimino bridge should be connected to the spatially more distant a',a-3,5- (in 2) and a,a'-1,5-positions (in 3 and 7), thus weakening their ON bond.
(ii) Competition between oxygen and nitrogen of epoxyimino bridge to attract proton; as shown in Scheme 4, only protonated oxygen can initiate reaction proceeding in the right direction.
Results given in Table 1 indicate that length of the NO bond in the O-protonated intermediates
(of type B, Scheme 4) in saturated and unsaturated 3,5- and 1,5-isoxazolidines are very
similar. Moreover, stretching of the NO bond in saturated systems is somewhat greater than in
the unsaturated ones, but the former compounds do not open their isoxazolidine ring under acid-catalyzed conditions.
| Table 1.
N-O Bond lenghts for the species protonated at oxygen | ||||
| Protonated isoxazolidine | 3,5-saturated I | 1,5-saturated II | 1-3,5-unsaturated
III |
3-1,5-unsaturated
IV |
| Distance N-O [Å] | 1.814 | 1.816 | 1.781 | 1.807 |
Besides, Table 2 (in which the calculated heats of formation of various reaction intermediates are
given) shows that in all investigated cases, O-protonated intermediates have considerably higher
energies (for 16 - 20 kcal/mol) than the corresponding N-protonated species. Yet, reaction
products of the latter intermediates were not detected in the respective reaction mixtures.
| Table 2.
The calculated energies (Hf in [kcal/mol]) of cis-decaline isoxazolidines I - IV and their respective reaction intermediates leading to the imino form D | |||||
| I Saturated 3,5-isoxazolidine | protonated | ||||
| at N | at O, B | open C | imino form D | ||
| 26.048 | 123.874 | 148.408 | 147.987 | 96.339 | |
| II Saturated 1,5-isoxazolidine | protonated | ||||
| at N | at O, B | open C | imino form D | ||
| 22.454 | 129.481 | 145.965 | 147.741 | 96.332 | |
| III 1- 3,5-isoxazolidine |
protonated | ||||
| at N | at O, B' | open C' | pyrrolidine form | imino form D' | |
| 0.655 | 151.405 | 170.363 | 170.900 | 165.190 | 121.492 |
| IV 3- 1,5-isoxazolidine |
protonated | ||||
| at N | at O, B" | open C" | aziridine form | imino form D" | |
| 2.116 | 153.343 | 169.217 | 175.391 | 163.342 | 119.973 |
Therefore, in order to explain the acid-catalyzed reactivity of unsaturated compounds, some other factors had to be also considered. Namely, it can be safely assumed that olefinic bond present in the molecules 2, 3, and 7, is directly involved in the rearrangement process, by interaction with electrophylic N atom in the intermediate C.
To establish what kind of stabilization by the olefinic double bond in the model compounds III and IV is possible, an extensive calculation of potential energy surfaces for these compounds have been done. As reaction coordinates were chosen the N-O distance in the course of transformation B to C, and the distance between nitrogen and one of the carbons on the double bond in the course of formation D.
Figure 2 shows the potential energy surface for conversion of 1-3,5-isoxazolidine, III, to the
imine D'. The primarily formed O-protonated intermediate of structure B' is transformed via the
completely opened intermediate C' to the most stable intermediate for which the pyrrolidine
structure has been computed. The latter species after proton elimination from the N-methyl
group forms imine D'. Activation energies for the above transformations from B' to C' and to
pyrrolidine intermediate are very low (about 2-3 kcal/mol).
Very similar results are obtained for 1-3,5-isoxazolidine, IV (shown in Fig.3). Primarily
formed, O-protonated intermediate (structure B"), after complete opening of the isoxazolidine
ring gives intermediate structure C", which is stabilized by participation of the 3-double bond to
form aziridine intermediate. By proton elimination from N-methyl group, it is transformed to
imine D". Activation energies for interconversions among these structures are also very low,
ranging from 2 to 3 kcal/mol.
These results suggest that stabilization due to formation of the pyrrolidine structure (for III), and the aziridine structure (for IV) makes feasible the observed isomerization of unsaturated isoxazolidines leading to the corresponding imine intermediates, and further to the perhydro-1,3-oxazine derivatives. As shown by experiment, the saturated systems, in which such a stabilization is not possible, do not react under similar acid-catalyzed condition.
Experimental
General.
Removal of solvents was carried out under reduced pressure. Prep. column chromatography: silica gel 0.063-0.200 mm. TLC: control of reactions and separation of products on silica gel G (Stahl) with benzene/AcOEt 9:1 and 7:3, toluene/AcOEt 9:1, 8:2 and 7:3 or toluene/EtOEt 9:1 and 8:2, detection with 50% aq.H2SO4 soln. M. ps. uncorrected. IR spectra: Perkin-Elmer-337 spectrophotometer; in cm-1. NMR spectra: Brucker AM-360 or Varian Gemini 200 (1H at 360 or 200 MHz, 13C at 90.55 or 50.28 MHz ); CDCl3 soln. at r. t., TMS as internal standard; chemical shifts in ppm as values, J in Hz. Mass spectra: Finnigan-MAT 8230. Light petroleum: fraction boiling at 40-60 C.
Preparation of starting materials:
N-Methyl-isoxazolidines 2, 3, 6- 8 were prepared according to the reported procedures [1,2].
N-methyl-3,5-epoxyimino-5-cholestane (4). - To a stirred suspension of N-methyl-3,5-epoxyimino-5-cholest-1-ene (2) (0.2 g, 0.484 mmol) and potasium azodicarboxylate [5] (0.8 g,
4.12 mmol) in methanol (10 ml, distilled over NaBH4) and dry methylene dichloride (10 ml)
cooled in an ice bath, a solution of acetic acid (0.52 ml, 8 mmol) in dry methanol (2 ml) was
slowly added. The reaction mixture was stirred at room temperature for additional 12 h, than
poored into water and extracted with diethyl ether-methylene dichloride. The organic layer was
washed with water, 5% aq. NaHCO3 soln., water, dried over Na2SO4, and evaporated to dryness.
The residue were recrystallized from acetone, to give N-methyl-3,5-epoxyimino-5-cholestane
4 (0.176 g, 87.5%), m.p. 65 °C. [
]D25= - 4.7 (c=1, CHCl3). IR (KBr): 1470, 1450, 1435, 1380,
1370, 945, 930 . 1H-NMR (360 MHz): 0.67 (s, CH3(18)), 0.88 (2d, J=6, CH3(26), CH3(27)), 0.91
(d, J=6, CH3(21)), 0.97 (s, CH3(19)), 2.01 (fd, J=12.5, H-C(6)), 2.33 (d, J=12.5, H-C(4)), 2.51
(s, CH3-N), 4.41 (t, J= 5, H-C(3)). 13C - NMR (90.55 MHz): 75.0 (d, C(3)), 70.6 (s, C(5)), 56.9
(d, C(17)), 56.6 (d, C(14)), 44.8 (d, C(9)), 43.2 (s, C(13)), 42.5 (q, CH3-N), 41.3 (s, C(10)), 40.7
(t, C(12)), 39.8 (t, C(24)), 37.3 (t, C(4)), 36.5 (t, C(22)), 36.1 (d, C(8)), 35.4 (d, C(20)), 29.6 (t,
C(6)), 29.5 (t, C(2)), 28.7 (t, C(1)), 28.6 (t, C(16)), 28.3 (d, C(25)), 27.9 (t, C(7)), 24.4 (t, C(23)),
24.1 (t, C(15)), 23.0 (q, C(26)), 22.8 (q, C(27)), 22.8 (t, C(11)), 19.2 (q, C(19)), 18.9 (q, C(21)),
12.6 (q, C(18)). MS: m/z = 415 (M+, 100%), 369, 194, 165, 110, 97, 81, 57, 53. Anal. calc. for
C28H49 ON (415.705): C 80.90, H 11.88; found: C 80.59, H 11.82.
Acid-catalyzed reactivity of N-methyl isoxazolidine derivatives 2-4 and 6-8:
Reaction of N-methyl-3,5-epoxyimino-5-cholest-1-ene (2) with p-toluenesulfonic acid. - A
solution of 2 (50 mg) and p-toluenesulfonic acid (5 mg) in toluene (20 ml) was refluxed for 48 h,
then diluted with diethyl ether and washed with 5% aq. NaHCO3 soln. and water. The organic
layer was dried over Na2SO4 and evaporated to dryness. The residue was chromatographed on 2 g
silica gel. Elution with benzene-diethyl ether (98:2) gave the starting isoxazolidine 2 (10 mg,
20%), identified by IR and 1H-NMR spectra (which were identical to those of an authentic
sample). Benzene-diethyl ether (96:4) and (94:6) eluted a complex mixture (8 mg, ~15%), which
was not further investigated. Elution with benzene-diethyl ether (90:10) gave 3,5-epoxymethyleneimino-5-cholest-1-ene (9) (21 mg, 42%), m.p. 118 °C (from acetone-MeOH).
[]D25= +131 (c=0.50, CHCl3); IR (KBr): 3290 (-NH-), 1470, 1430, 1385, 1370, 1135, 965. 1H-NMR (360 MHz): 0.66 (s, CH3 (18)), 0.86 (2d, J=6, CH3(26), CH3(27)), 0.89 (d, J=6, CH3(21)),
1.10 (s, CH3(19)), 1.96 (dt, J=12, 5.4, H-C(6)), 2.23 (fd, J=12, H2C(4)), 4.36 (m, H-C(3)), 4.18
and 4.66 (2d, J=11.4, H2C(28)), 5.53 (dd, J=11.4, 5.4, H-C(2)), 6.20 (d, J=11.4, H-C(1)); 13C-NMR (90.55 MHz) (CDCl3): 144.1 (d, C(1)), 118.7 (d, C(2)), 68.8 (t, C(28)), 66.0 (d, C(3)), 55.9
(d, C(17)), 55.6 (d, C(14)), 51.6 (s, C(5)), 49.4 (d, C(9)), 42.0 (s, C(13)), 41.2 (s, C(10)), 39.6 (t,
C(12)), 39.0 (t, C(24)), 37.1 (t, C(22)), 35.7 (t, C(4)), 35.7 (t, C(6)), 35.3 (d, C(20)), 34.8 (d,
C(8)), 27.7 (t, C(16)), 27.6 (d, C(25)), 26.4 (t, C(7)), 23.8 (t, C(15)), 23.4 (t, C(23)), 22.4 (q,
C(27)), 22.1 (q, C(26)), 21.6 (t, C(11)), 18.2 (q, C(21)), 14.3 (q, C(19)), 11.5 (q, C(18)); MS: m/z
= 413 (M+), 412 (M+-1), 398, 384, 368, 344, 43 (100%); Anal. calc. for C28H47ON (413.689): C
81.29, H 11.45, N 3.40; found: C 80.96, H 11.29, N 3.37.
Reaction of N-methyl-1,5-epoxyimino-5-cholest-3-ene (3) with p-toluenesulfonic acid. - A
solution of 3 (50 mg) and p-toluenesulfonic acid (5 mg) in toluene (20 ml) was refluxed for 48 h
and the mixture worked up as above. The residue was chromatographed on 2 g silica gel. Elution
with benzene-diethyl ether (90:10) gave starting isoxazolidine derivative 3 (9.5 mg, 19%),
identified by m.p., mixed m.p., IR and NMR spectra. Benzene-diethyl ether (85:15) and (80:20)
eluted a complex mixture (8 mg, ca. 15%). Further elution with benzene-diethyl ether (80:20)
gave 1,5-epoxymethyleneimino-5-cholest-3-ene (10) (25.5 mg, 51%), m.p. 145 °C (from
acetone-MeOH). []D25= +3 (c=0.50, CHCl3). IR (KBr): 3310 (-NH-), 1470, 1440, 1375, 1135,
1025, 935. 1H-NMR (360 MHz): 0.66 (s, CH3(18)), 0.86 (2d, J=6, CH3(26), CH3(27)), 0.90 (d,
J=6, CH3(21)), 1.28 (s, CH3(19)), 1.96 (dt, J=12, 6, H-C(6)), 2.32 (m, H2-C(2)), 3.73 (br s, H-C(1)), 4.32 and 4.82 (2d, J=11.4, H2-C(28)), 5.28 (fd, J=11.5, H-C(4)), 6.13 (dt, J=11.5, 5.5 H-C(3)). 13C-NMR (90.55 MHz): 130.2 (d, C(3)), 129,8 (d, C(4)), 73.8 (d, C(1)), 70.5 (t, C(28)),
56.0 (d, C(14)), 56.0 (d, C(17)), 53.2 (s, C(5)), 43.7 (d, C(9)), 42.6 (s, C(13)), 39.6 (t, C(12)),
39.4 (t, C(24)), 38.6 (s, C(10)), 36.0 (t, C(22)), 35.6 (d, C(20)), 35.0 (d, C(8)), 33.4 (t, C(6)), 29.4
(t, C(2)), 28.0 (t, C(16)), 27.9 (d, C(25)), 27.7 (t, C(7)), 23.6 (t, C(23)), 23.9 (t, C(15)), 22.6 (q,
C(26)), 22.4 (q, C(27)), 21.8 (t, C(11)), 18.6 (q, C(21)), 13.8 (q, C(19)), 12.0 (q, C(18)). MS: m/z
= 413 (M+), 412 (M+ -1), 398, 384, 368, 353, 43 (100%). Anal. calc. for C28H47ON (413.689): C
81.29, H 11.45, N 3.40; found: C 81.70, H 11.89, N 3.29.
Reaction of N-methyl-1,5-epoxyimino-19-nor-5-androst-3-en-17-yl acetate (7) with p-toluenesulfonic acid.- A solution of 7 (67 mg) and p-toluenesulfonic acid (7 mg) in toluene (26.8
ml) was refluxed for 48 h and the mixture worked up as above. The residue was
chromatographed on 2 g silica gel. Elution with benzene-diethyl ether (85:15) afforded starting
isoxazolidine derivative 7 (11.4 mg, 17%), identified by m.p., mixed m.p., IR and NMR spectra.
Benzene-diethyl ether (85:15) and (80:20) eluted a complex mixture (10 mg, ~15%). Further
elution with benzene-diethyl ether (80:20) gave 1,5-epoxymethyleneimino-19-nor-5-androst-3-en-17-yl acetate (11) (27 mg, 54%), m.p. 110 °C (from petrolether-acetone), []D25 = + 6.5
(c=0.83, CHCl3). IR (KBr): 3286 (-NH-), 1739, 1451, 1374, 1248, 1046, 795. 1H-NMR (200
MHz): 0.80 (s, CH3(18)), 2.04 (s, AcO-C(17)), 4.18 (br.s, w/2= 13.2, H-C(1)), 4.31 and 4.78 (2d,
J= 10.8, H2-C(21)), 4.57 (dd, J= 9, 7.6 , H-C(17)), 5.33 (d, J=9.8, H-C(4)), 6.15 (dt, J=7, 3.1, H-C(3)). 13C-NMR (50.28 MHz): 171.2 (s, MeCOO); 131.9 (d, C(3)); 128.9 (d, C(4)); 82.7 (d,
C(17)); 70.6 (t, C(21)); 68.5 (d, C(1)); 50.9 (s, C(5)); 50.7 (d, C(10)); 49.3 (d, C(14)); 42.9 (s,
C(13)), 41.0 (d, C(9)), 40.9 (d, C(8)), 38.7 (t, C(12)), 36.5 (t, C(2)), 28.7 (t, C(6)), 27.3 (t, C(16)),
26.9 (t, C(7)), 24.9 (t, C(11)), 23.2 (t, C(15)), 21.1 (q, MeCOO), 12.0 (q, C(18)). MS: m/z = 346
(M+ +1), 345 (M+), 286 (100%). Anal. calc. for C21H31 NO3 (345.483): C 73.00, H 9.05; found: C
72.14, H 9.13.
Reaction of N-methyl-1,5-epoxyimino-19-nor-5-androstan-3,17-diyl diacetate (6) with p-toluenesulfonic acid. - A solution of 6 (200 mg) and p-toluenesulfonic acid (20 mg) in toluene (80 ml) was refluxed for 48 h and the mixture worked up as above. The residue was chromatographed on 8 g silica gel. Elution with toluene-ethyl acetate (80:20) and (75:25) gave a complex mixture (40 mg, 20%), which was not further investigated. Toluene-ethyl acetate (70:30) eluted starting material 6 (134 mg, 67.5%), the IR and NMR spectra of which were identical to those observed for a previously described sample 6 [3].
Reaction of N-methyl-1,5-epoxyimino-5-cholestane (4) with p-toluenesulfonic acid.- A solution of 4 (52 mg) and p-toluenesulfonic acid (5.2 mg) in toluene (20.4 ml) was refluxed for 48 h and the mixture worked up as above. The residue was chromatographed on 2 g silica gel. Elution with toluene and toluene-ethyl acetate (95:5) gave a complex mixture (9 mg, 18%). Further elution with toluene-ethyl acetate (95:5) afforded starting isoxazolidine 4 (34 mg, 68%), the IR and 1H-NMR spectra of which were identical to those observed for an authentic sample 4.
Reaction of N-methyl-1,5-epoxyimino-19-nor-5-androstane-3,17-diol (8) with p-toluenesulfonic acid. - A solution of 8 (46 mg) and p-toluenesulfonic acid (5 mg) in toluene (18.4
ml) was refluxed for 48 h and the mixture worked up as above. After evaporation, the residue
were crystallized from acetone, giving the starting isoxazolidine 8 (34 mg, 73.9%), identified by
m.p., mixed m.p., and spectral data [3]. The residue (6 mg, ~13%) was a complex mixture.
Acknowledgement. The authors are grateful to The Serbian Academy of Science and Arts and to
The Serbian Ministry of Sciences and Technology for financial support.
References and notes
[1] Rajkovic M, Lorenc Lj, Petrovic I, Milovanovic A, Mihailovic MLj. Tetrahedron Lett.1991; 32: 7605.
[2] Mihailovic MLj, Lorenc Lj, Rajkovic M, Juranic I, Milovanovic A. Heterocycles 1989; 28: 869.
[3] Mihailovic MLj, Rajkovic M, Lorenc LjB, Pavlovic VD, Milovanovic A, Tinant B. Declercq J-P. Tetrahedron 1996; 52: 11995.
[4] Pavlovic V, Rajkovic M, Lorenc Lj, Mihailovic MLj. J. Serb. Chem. Soc.1994; 59(12): 903.
[5] Serene Software Box Bloomington IN 45402-3076.
[6] Gajevski KE, Gilbert KE, McKelvey J. Adv. Mol. Mod. 1990; 2: 65.
[7] Burket U, Allinger NL. Molecular Mechanics. Washington, DC: American Chemical Society,1982.
[8] Stewart JJP. J. Comp.-Aided Molec. Des.1990; 4(1): 1.
[9] Stewart JJP. QCPE # 455.
[10] Stewart JJP. J. Comp. Chem. 1989; 10: 109;
[11] Stewart JJP. J. Comp. Chem. 1989; 10: 221.
[12] Bartoszak E, Degaszafran Z, Grundwaldwyspianska M, Jaskolski M, Szafran M J. Chem. Soc., Faraday Trans 1993; 89(12): 2085;
[13] Burk P, Koppel IA. Theoretica Chimica Acta 1993; 86(5): 417;
[14] Zuilhof H, Vertegaal LBJ, Vandergen A, Lodder G. J. Org. Chem. 1993; 58(10): 2804;
[15] Aakeroy CB. J. Mol. Struct. (Teochem) 1993; 100(2-3): 259;
[16] Jurema MW, Shields GC. J. Comp. Chem. 1993; 14(1): 89;
[17] Matsuoka T, Harano K. Tetrahedron 1995; 51: 6451-6458;
[18] Kelly S, Malachowski MR. Analytical Chemistry 1993; 65: 1654-1657;
[19] Kalkanis GH, Shields GC. J. Phys. Chem. 1991, 95: 5085-5089;
[20] Beck B, Clark T, Glen RC. J. Comput. Chem. 1997; 18: 744-756.
[21] Thile J. Ann. 1982; 271: 127;
[22] Adam W, Eggelte HJ. Angew. Chem. Int. Ed. Engl. 1977; 16(10): 713.
1. E-mail: ijuranic@chem.bg.ac.yu
2. Our esteemed teacher, colleague and friend deceased on June 8. 1998.
3. As reported in Ref. 1, protonation of the nitrogen in the epoxyimino bridge of 1,5-isoxazolidines 3 and 7 results in elimination of the CH3NH2 fragment and formation of the unsaturated 1-oxo derivatives.