Resolution and conformational analysis of diastereoisomeric esters of cis- and trans-2-(aminomethyl)-1-carboxycyclopropanes

Article abstract of DOI:10.1016/S0957-4166(98)00250-X

(1R,2S)-, (1S,2R)-, (1R,2R)- and (1S,2S)-2-(Aminomethyl)-1- carboxycyclopropanes, conformationally restricted analogues of the neurotransmitter γ-aminobutyric acid (GABA), have been resolved by chromatographic separation of the corresponding diastereoisomeric esters which were formed between the cis- and trans-2-(acetamidomethyl)1- carboxycyclopropanes with (R)-(-)-pantolactone. ¹H NMR, semi-empirical conformational analysis, ab initio (DFT) structure and NMR shielding tensor calculations of the cis-diastereoisomers allowed the absolute configuration assignments of the cis-amino acids.

Full text of DOI:10.1016/S0957-4166(98)00250-X

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 2533–2548
Resolution and conformational analysis of diastereoisomeric
esters of cis- and trans-2-(aminomethyl)-1-carboxycyclopropanes
Rujee K. Duke,∗ Robin D. Allan, Mary Chebib, Jeremy R. Greenwood and
Graham A. R. Johnston
Adrien Albert Laboratory of Medicinal Chemistry, Department of Pharmacology, The University of Sydney, Sydney,
NSW 2006, Australia
Received 2 June 1998; accepted 22 June 1998
Abstract
(1R,2S)-, (1S,2R)-, (1R,2R)- and (1S,2S)-2-(Aminomethyl)-1-carboxycyclopropanes,conformationallyrestricted
analogues of the neurotransmitter γ-aminobutyric acid (GABA), have been resolved by chromatographicseparation
of the corresponding diastereoisomeric esters which were formed between the cis- and trans-2-(acetamidomethyl)-
1-carboxycyclopropanes with (R)-(−)-pantolactone. 1H NMR, semi-empirical conformational analysis, ab initio
(DFT) structure and NMR shielding tensor calculations of the cis-diastereoisomers allowed the absolute configur-
ation assignments of the cis-amino acids. © 1998 Elsevier Science Ltd. All rights reserved.
1. Introduction
γ-Aminobutyric acid, GABA 1, the major inhibitory neurotransmitter in the mammalian central
nervous system, activates three major classes of GABA receptors, the GABAA, GABAB, and GABAC
receptors. GABAA receptors are neurotransmitter gated chloride ion channels which are blocked by
the alkaloid bicuculline, and modulated by barbiturates, steroids and benzodiazepines.1 In contrast,
GABAB receptors operate via second messenger systems and are coupled to calcium and potassium
channels through G-proteins. GABAB receptors are insensitive to bicuculline and GABAA modulators,
are activated by (R)-baclofen and antagonised by the phosphonic and sulfonic analogues of baclofen,
phaclofen and saclofen respectively.2 GABAA and GABAC receptors share some common features;
both are ionotropic, conducting chloride ions, and show marked differences from GABAB receptors
such as being insensitive to (R)-baclofen. However, unlike GABAA receptors, GABAC receptors are
not antagonised by bicuculline, nor are they modulated by barbiturates and benzodiazepines.3,4 Z-4-
Aminobut-2-enoic acid (cis-aminocrotonic acid, CACA) 2 (Fig. 1) is a selective partial agonist,3,4 and
∗
Corresponding author. E-mail: rujeek@pharmacol.usyd.edu.au
0957-4166/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00250-X

2534
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
Fig. 1. GABAC receptor agonists and partial agonists
(1,2,5,6-tetrahydropyridin-4-yl)methylphosphinic acid (TPMPA) is a selective antagonist5,6 at GABAC
receptors.
CACA 2 and E-4-aminobut-2-enoic acid (trans-aminocrotonic acid, TACA) 3 are unsaturated ana-
logues of GABA with the four carbon atoms of the GABA chain constrained to lie in a plane, 2 in
a folded conformation and 3 in an extended conformation. While 2 is a selective partial agonist at
GABAC receptors, 3 is a potent partial agonist at both GABAA and GABAC receptors. The saturated
carbocyclic analogues of GABA, cis-2-(aminomethyl)-1-carboxycyclopropane (CAMP) 4 and trans-2-
(aminomethyl)-1-carboxycyclopropane (TAMP) 5 were synthesised7 to evaluate the effects of constrain-
ing the carbon atoms of the GABA chain in a non-planar configuration by the cyclopropyl ring. While
the folded conformation is retained in 4 and extended in 5, the introduction of the cyclopropyl moiety on
the GABA skeleton induces chirality, resulting in racemic mixtures which complicate pharmacological
evaluation. (ꢀ)-CAMP 4 was found to show selectivity similar to that for CACA 2 and (ꢀ)-TAMP 5 was
similar to TACA 3 in depressing neuronal firing at the bicuculline sensitive site of cat spinal neurones
and in receptor binding assay.7 The conclusion which may be drawn from these observations is that
the folded conformation of the GABA chain was favoured for selectivity at GABAC receptors. When
compounds 2–5 were tested on cloned GABAA and GABAC receptors, the trans-isomers, TACA 3 and
(ꢀ)-TAMP 5, were shown to be more potent at both receptors but less selective for GABAC receptors
than the corresponding cis-isomers.8,9 Since 4 and 5 were tested as racemates, the activities observed
were the combined effect of the enantiomeric pair. The resolved enantiomers were required in order to
associate activity directly with the stereochemical arrangement of binding groups required for GABA
receptor activation.
Our initial attempts to obtain enantiomers of 4 and 5 using a chiral HPLC column were not successful.
No detectable separation between the enantiomeric pairs of 4 and of 5 was observed by HPLC using
a chiral Crown column which is commonly effective for the resolution of α-amino acids. However,
separation via diastereoisomeric esters proved successful. This paper describes the resolution of 4
and 5 via (R)-(−)-pantolactonyl esters, as well as the prediction of the absolute configuration of the
1
diastereoisomers of 4 by H NMR, conformational analysis, ab initio density functional theory (DFT)
and NMR shielding tensor calculations.
2. Discussion
Enantiomers 10, 11 and 16 (Schemes 1 and 2) have been prepared previously by asymmetric syntheses
using the Simmons–Smith reactions of Z- and E-allyl alcohol derivatives obtained from (R)-2,3-O-
isopropylideneglyceraldehyde.10 Enantiomer 10 has also been synthesised from the diastereoisomerically
pure derivative of 3-aza-2-oxobicyclo[3.1.0]hexane prepared by an intramolecular cyclisation.11 Our
resolution has the practical advantage of providing all of the four enantiomers 10 and 11, 16 and 17
which were required for pharmacological evaluation although in low yield. (R)-(−)-Pantolactone, which

R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
2535
was chosen for the formation of the diastereoisomeric esters, has the advantage of having three singlet
resonances in the 1H NMR spectrum which can be used in determining the purity of diastereoisomers.12
Scheme 1. (i) 6 M HCl/AcOH/RT, (ii) (R)-(−)-pantolactone, DCC, acetonitrile, (iii) chromatographic separation, (iv) 6 M
HCl/AcOH/reflux
Scheme 2. (i) 6 M HCl/AcOH/RT, (ii) (R)-(−)-pantolactone, DCC, acetonitrile, (iii) chromatographic separation, (iv) 6 M
HCl/AcOH/reflux
The synthetic route for the preparation of 10 and 11, the enantiomers of 4, is shown in Scheme 1. The
cis-amido ester 6 was prepared by catalytic hydrogenation of cis-2-cyano-2-ethoxycarbonylcyclopropane
in the presence of acetic anhydride.7 Partial hydrolysis removed the ester and gave the amido acid 7 in
good yield. Carbodiimide coupling of amido acid with R-(−)-pantolactone gave low to moderate yields
of diastereoisomeric esters 8 and 9 which were separated using vacuum chromatography.13 The pure
diastereoisomers were cleaved with acid to give the corresponding amino acids 10 and 11.
Compounds 16 and 17, the enantiomers of 5, were prepared by a similar procedure, as shown in
Scheme 2, from the trans-amido ester 12. 12 was prepared by catalytic hydrogenation of trans-2-cyano-
2-ethoxycarbonylcyclopropane in the presence of acetic anhydride.7
1
The diastereoisomeric esters have different 13C NMR and H NMR resonances and therefore are
completely distinguishable. The differences for particular peaks can be very large, up to 1 ppm, and all

2536
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
resonances are separable when chemical shifts are measured to greater than two decimal places. 1H and to
a lesser extent 13C NMR resonances also provide additional means for determining the diastereoisomeric
composition and purity. As anticipated, the patterns and the peak heights of the geminal dimethyl singlets,
0
as well as the singlet from the proton attached to the chiral carbon (H2 ) of the pantolactonyl moiety,
proved to be excellent measures of the diastereoisomeric purity. Contamination above 0.3% by the other
diastereoisomer could be detected. The intensity of the readily distinguishable geminal methyl singlet
0
pair, acetyl singlet and H2 singlet of the diastereoisomers [(8, δ 1.261 and 1.228, 1.986, 5.44) (9, δ
1.234 and 1.134, 1.982, 5.47); (14, δ 1.207 and 1.134, 2.014, 5.36) (15, δ, 1.212 and 1.117, 2.003,
5.34)] allowed facile assessment of diastereoisomeric ratios. In addition, the signals of the methylene
protons adjacent to the amide nitrogen were widely separated in three of the four diastereoisomers [(8,
δ 4.32–4.23, m, 2.68–2.59, m), (9, δ 3.95, ddd, 2.95, ddd); (14, δ 3.52–3.44, m, 3.07–2.98, m), (15,
δ 3.35–3.17, m)]. When the amide hydrogen was saturated or exchanged, these signals resolved to the
simpler AM part of the AMX spin system in diastereoisomers 8 and 9 and the AB part of the ABX
spin system in 14 and 15. The strong couplings between the NH and these methylene protons were also
clearly seen in the COSY spectra of all the diastereoisomers. The AM spin system indicates that one of
the methylene protons is strongly shielded and the other strongly deshielded. To better understand the
causes of the large difference in their magnetic environments, diastereoisomers 8 and 9 were investigated
more rigorously.
The infrared spectra of the diastereoisomers recorded neat are very similar, with the exception of
that of 8 which differs in the region of the NH and carbonyl stretching frequencies. Apart from very
strong NH absorption at 3379 cm−1, the amide and the lactone carbonyl stretching frequencies of 8
are approximately 10 cm−1 higher and lower, respectively, from the corresponding carbonyls of the
other diastereoisomers. All four diastereoisomers show NH stretching vibrations varying in strength
from 3610–3557 cm−1, as well as multiple bands in the 3310–3082 cm−1 region in the infrared spectra,
recorded neat, indicating that the amide groups are involved in various degrees and forms of hydrogen
bonding. These multiple bands were absent in the spectra recorded with less concentrated samples (0.1
M, CHCl3). The band at 3400 cm−1 in 8 was unaffected by dilution, but in 9 it was moderately affected,
being weaker at 0.01 M, and weaker again on further dilution to 0.001 M. The bands at 3460 cm−1 in 14
and 15 were severely affected by dilution, becoming very weak at 0.01 M and almost absent on further
dilution to 0.001 M. The results suggest that intramolecular hydrogen bonding is strong in 8, weaker in
9 and not significant in 14 and 15. The lowering in the lactone carbonyl stretching frequencies, 1780
cm−1 (0.1 M, CHCl3) in 8 and 1785 cm−1 in 9 as compared with 1795 cm−1 in 14 and 15 are strongly
indicative of intramolecular hydrogen bonding between the lactone carbonyls and the amide hydrogens.
The amide carbonyl stretching frequencies in 8 and 9 are also lower, appearing at 1665 cm−1, whereas
those of 14 and 15 are at 1670 cm−1. The NH exchange rates in the 1H NMR of the diastereoisomers also
correspond qualitatively to the relative strength of their intramolecular hydrogen bonding. On addition
of D2O, the NH resonances of the trans-diastereoisomers, 14 and 15 completely exchanged within 6
hours whereas complete exchange for the cis-diastereoisomers 9 and 8 requires up to 24 and 72 hours
respectively. The slow exchange rate of 8 and 9 is possibly a result of the amide hydrogen being encased
within the folded conformation of the cis-structure and therefore not readily accessible for interaction
with D2O. The comparative exchange rates of the two diastereoisomers may partly reflect the relative
strength of intramolecular hydrogen bonding.
Each of the diastereoisomers exhibits a strong absorption in the region of 1544–1549 cm−1 which is
characteristic of trans-geometry of the amide bond. Assuming the amide linkage to adopt a near-planar
geometry, diastereoisomers 8 and 9 each possess five rotatable bonds which need to be considered, as

R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
2537
depicted in Fig. 2. Using the package Spartan 5.0,14 a systematic search of the conformational space of
the trans-amide forms was carried out.
Fig. 2. Five rotatable bonds of diastereoisomers 8 and 9 to be considered when mapping conformational space
Three initial dihedral angles about each rotatable bond were selected, for a total number of 243
conformations per isomer. Each conformer was then geometry optimised using the semi-empirical AM1
method. One conformer of 8 was clearly favoured energetically, whereas three conformations of 9 lay
close in energy. The lowest energy conformers of 8 and 9 shown in Fig. 3 possess an intramolecular
lactone carbonyl–amide hydrogen bond, consistent with the lowering of the pantolactone carbonyl
stretching frequencies observed in the infrared spectra. These conformers also show that H5B in 8 is
closer to the carbonyl oxygen of the acyclic ester while H5B in 9 is closer to the saturated oxygen.
Fig. 3. Conformational global minima of diastereoisomers 8 and 9 as determined by exhaustive conformational analysis using
AM1 theory and geometry optimisation using B3LYP/6-31G(d) theory. Atom labels in red and inside the brackets refer to
chemical nomenclature used throughout the text
Conformers having cis-geometry about the amide bond were also subjected to the same calculation.
According to AM1 semi-empirical method, the global minima of the cis-geometry structures were found
to be higher than that of the trans-geometry, by 3.1 and 1.6 kcal mol−1 for 8 and 9, respectively.
Therefore, the amide bonds in 8 and 9 are likely to exist mainly in the trans-geometry consistent with
the results obtained from the infrared spectra. Consequently, conformers with the cis-geometry were not
subjected to further analysis.
The lowest energy conformer of 8, as well as the five best conformers of 9, as determined by the AM1
method, were submitted to geometry optimisation by high-level gas-phase ab initio density functional
theory calculations (DFT) using the Gaussian 94 package.15 Geometry optimisation was performed

2538
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
using B3LYP theory and the basis set 6-31G(d), a theoretical method known to accurately reproduce
experimental intramolecular hydrogen bonding lengths.16 The results agreed qualitatively with the AM1
predictions, and, for 9, the predicted global energy minimum conformer was 0.92 kcal mol−1 favoured
over the next lowest energy local minimum, indicating the contribution of minor conformers in the
approximate ratios of major:minors=82:17:1. The presence of the predominant minor conformer is
consistent with an extra set of methylene resonances at δ 3.62 and δ 3.37 each integrating for slightly
1
greater than 0.1H in the H NMR spectrum and the corresponding 13C–1H resonance at δ 38.6 in the
HETCOR spectrum of 9.
The gas-phase structure of 8 and the lowest energy structure of 9 are shown in Fig. 3, and their
geometries are given in Tables 1 and 2. Single-point energy calculations were then calculated with
this geometry using Hartree–Fock theory, and the 6-311G(d,p) basis set. NMR shielding tensors were
calculated using the gauge-independent atomic orbital (GIAO) method.17,18 This model incorporating a
triple ξ basis set is recommended by Cheeseman et al.19 as providing accurate gas-phase geometries and
reliable estimates of NMR chemical shifts. The calculated chemical shift values for the 13C resonances
shown in Table 3 and for 1H resonances shown in Table 4, agreed qualitatively with observed values with
a few minor exceptions.
While the calculated and observed chemical shifts of H5A and H5B in 8 are in excellent agreement, a
large discrepancy is evident in the chemical shift value of H5B in 9. According to the calculated chemical
shift, H5B is strongly shielded (δ 2.22) which is inconsistent with the observed value of δ 2.95 ppm. In
addition, H5B of 9, shown in Fig. 3, does not appear to be deep in the carbonyl shielding cone (dihedral
angle OC4OH5B=147.5°, angles C4OH5B=19.6° and H5BC4O=152.9°), and is a long way away from
the C4 carbonyl (distances H5BO=4.22 Å and H5BC4=3.11 Å), and therefore it is not expected to be
greatly affected by the carbonyl shielding. In contrast, the carbonyl shielding of H5B in 8 was predicted
from the gas-phase structure and was observed experimentally. In the lowest energy conformer of 8 also
shown in Fig. 3, H5B lies well within the carbonyl shielding cone of the acyclic ester (dihedral angle,
OC4OH5B=83.4°, angles C4OH5B=92.2° and H5BC4O=62.1°). This, together with its close proximity
to the C4 carbonyl group (distances H5BO=2.47 Å and H5BC4=2.80 Å), results in H5B being strongly
shielded by the C4 carbonyl consistent with the observed resonance for H5B in 8 at δ 2.64 ppm. The large
discrepancy between the observed and calculated chemical shifts of H5B in 9 can be explained in terms of
the higher conformational flexibility of 9 as indicated by the earlier calculations. Although the dielectric
constant of chloroform is low, some conformational changes from the gas-phase structure may occur
when the other conformations are close in energy. Present limitations in computing capability preclude
the performance of higher level calculations on NMR shielding tensors.
The influence of carbonyl diamagnetic anisotropy could also be used to explain the difference between
the chemical shifts of H5A in 8 (δ 4.28) and 9 (δ 3.95). H5A atoms in 8 and in 9 both lie in the deshield-
ing cone of the C_O bond of the amide (8, dihedral angle, OCNH5A=15.4°, angles COH5A=78.3°,
H5ACO=72.1°; 9, dihedral angle OCNH5A=−22.9°, angles COH5A=78.3°, H5ACO=74.8°). Since the
deshielding effect induced by the carbonyl is greatest along the transverse axis of the C_O bond, it is
expected that H5A in 8 would be more affected by the deshielding process than H5A in 9. Moreover, the
distances between H5A and the amide carbonyl, H5AO=2.52 Å, H5AC=2.62 Å in 8 and H5AO=2.63 Å and
H5AC=2.66 Å in 9 also indicate that the deshielding is stronger for H5A in 8. These combined effects are
likely to be the cause of the large difference observed in their chemical shifts.
There were no significant changes in the 1H NMR spectra of 8 and 9 when the temperature was raised
(CDCl3, 20–40°C, in 10°C steps) indicating that 8 and 9 either exist in CDCl3 solution mainly in the
conformations predicted by the calculations or interconversion between conformations is not detectable
using the standard pulse delay over the temperature range investigated. In 8, the observed coupling

R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
2539
Table 1
Cartesian co-ordinates of conformational global minimum of diastereoisomer 8 as determined by
geometry optimisation at B3LYP/6-31G(d) DFT theory
constants between the protons of the more flexible bonds are in complete agreement with the dihedral
angles of the calculated structures. The coupling constants of 8.7 and 4.1 Hz which were clearly removed
from the multiplets of H5A and H5B on addition of D2O or irradiation of the NH peak are consistent
with the calculated dihedral angles between NH and C5H5A of −152.8° and NH and C5H5B of −35.4°
respectively. The dihedral angles, H2C2C5H5A=−55.4° and H2C2C5H5B=−174.4° are also consistent
with the couplings of 3.3 and 9.8 Hz observed between H2 and H5A, H2 and H5B respectively. In 9, the
observed coupling constants between NH and H5A, NH and H5B (JNH,H5A=6.1 Hz and JNH,H5B=6.3 Hz)
and between H2 and H5B (JH2,HB=9.6 Hz) are in accordance with the respective calculated dihedral angles

2540
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
Table 2
Cartesian co-ordinates of conformational global minimum of diastereoisomer 9 as determined by
geometry optimisation at B3LYP/6-31G(d) DFT theory
(HNC5H5A=132.6°, HNC5H5B of 15.5°, H2C2C5HB=166.3°). However, the coupling constant between
5A and H2 of 5.4 Hz is larger than the value of about 4 Hz predicted by the dihedral angle of 48.2°. This
H
discrepancy may be accounted for in terms of conformational flexibility of 9. The calculated dihedral
angle is located in the steep part of the vicinal Karplus correlation curve and therefore a small deviation
from this angle would result in a large change in the predicted coupling constant.
The cis-configuration of diastereoisomers 8 and 9 allows formation of intramolecular lactone
carbonyl–amide hydrogen bond, thereby adding stability to the conformation. In the case of 8 and 9,
intramolecular hydrogen bonding also induces large differences in the chemical shift of the neighbouring

R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
2541
Table 3
Relative 13C NMR chemical shifts of diastereoisomers 8 and 9 from magnetic shielding tensors
0
calculated at HF/6-311G(d,p)/6-31G(d) in ppm relative to the observed chemical shift values of C3 , 8
(δ 39.7), 9 (δ 40.0)
Table 4
Relative 1H NMR chemical shifts of diastereoisomers 8 and 9 from magnetic shielding tensors
0
calculated at HF/6-311G(d,p)/6-31G(d) in ppm relative to observed chemical shift values of H2 , 8
(δ 5.44), 9 (δ 5.47)
methylene protons. Conformational stability and large differences in the magnetic environment of the
methylene protons, together with known chirality of pantolactone, enabled the absolute configuration
to be predicted. The prediction was proven to be correct by comparison of the optical rotations of
the enantiomers 10 and 11 with the literature values.10,11 Using the same procedure as for 8 and 9,
diastereoisomers 14 and 15 were subjected to lowest energy conformer searches using the package

2542
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
Spartan 5.0.14 As anticipated for the trans-configuration, the low energy optimised conformers of 14
and 15 assume partially extended or extended conformations, and, consistent with the results obtained
from infrared spectrometry, intramolecular carbonyl–amide hydrogen bonding was not observed.
3. Conclusion
cis- and trans-2-(Aminomethyl)-1-carboxycyclopropanes 4 and 5 were successfully resolved
by the separation of diastereoisomeric esters prepared from the corresponding cis- and trans-2-
(acetamidomethyl)-1-carboxycyclopropanes and (R)-(−)-pantolactone. The chirality of the cis-
diastereoisomers, assigned by 1H NMR, conformational analysis, and ab initio (DFT) calculations, was
consistent with the absolute configurations of the corresponding amino acid enantiomers determined by
comparison of the optical rotation and reference to literature values.
4. Experimental
4.1. General procedures
Melting points were determined on a Reichert hot-stage apparatus and are reported uncorrected.
Elemental analyses were carried out by the Microanalysis Unit, Department of Chemical Engineering,
the University of Sydney. High and low resolution chemical ionisation mass spectral determinations
using CH4 as the reagent gas were carried out on a VG Scientific ZAB-E high resolution magnetic sector
mass spectrometer and a Finigan-MAT TSQ46 MS/MS instrument respectively. Solution infrared spectra
were acquired on a Perkin–Elmer 177 grating spectrophotometer. Reflection infrared spectra of the neat
material were acquired on a Bruker IFV66V FT-IR fitted with a microscope. Optical rotations were
recorded on a Perkin–Elmer 241 polarimeter.
1D 1H (300 MHz), 13C (75 MHz) and 2D NMR experiments were performed at 20°C using a 300 MHz
Varian–Gemini 300 spectrometer. For 1H NMR, chemical shift values are given in ppm relative to internal
TMS for spectra measured in CDCl3, and HOD (4.75 ppm) for spectra measured in D2O respectively. For
13C NMR, chemical shift values are given in ppm relative to CDCl3 (77.0 ppm) for compounds measured
in CDCl3 and relative to dioxan (67.4 ppm) for compounds measured in D2O. Standard pulse sequences
were used for COSY, DEPT and HETCOR experiments.
AM1 semi-empirical conformational analyses were performed by Spartan 5.014 on a SGI O2 R5000
computer running under IRIX 6.3. Ab initio Hartree–Fock and density functional theory (DFT) calcula-
tions were performed by Gaussian 9415 on an SGI Power Challenge R10000 network under IRIX 6.2.
4.2. (ꢀ)-cis-2-(Aminomethyl)-1-carboxycyclopropane 4
Crude (ꢀ)-CAMP 4 prepared by the method of Allan et al.7 was adsorbed on a cationic exchange
resin (Dowex 50) and eluted with 1 M pyridine. Removal of the eluant under reduced pressure gave a
light brown solid which was further purified by gel filtration on Sephadex G10 (H2O). Fractions were
analysed by TLC (silica gel, n-butanol:acetic acid:water=3:1:1, ninhydrin). Fractions containing 4 which
were homogeneous by TLC were combined and concentrated under reduced pressure to give a white
crystalline solid. Recrystallisation from H2O:EtOH gave colourless rectangular plates, m.p. 225–226°C
1
(lit.1 m.p. 225–226°C). m/z (%) [M+1]+ 116 (47), 99 (76), 98 (100). H NMR (D2O) δ 3.16 (1H, dd,

R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
2543
JAB=13.1 Hz, JAX=7.1 Hz, CH2N), 3.06 (1H, dd, JAB=13.1 Hz, JBX=7.7 Hz, CH2N), 1.72–1.64 (1H, m,
H1), 1.40–1.27 (1H, m, H2), 1.03–0.96 (1H, m, H3), 0.82–0.76 (1H, m, H3). 13C NMR (D2O) δ 181.0
(CO), 39.9 (CH2N), 21.7 (C2), 16.1 (C1), 11.6 (C3).
4.3. (ꢀ)-trans-2-(Aminomethyl)-1-carboxycyclopropane 5
Crude (ꢀ)-TAMP 5 prepared by the method of Allan et al.7 was purified by the procedure described
for the purification of 4. The white crystalline solid obtained from ion exchange and gel filtration
chromatographic purification was recrystallised from H2O:EtOH to give colourless needles, m.p. 245°C
(subl. from 240°C), [lit.1 m.p. 270–275°C]. m/z (%) [M+1]+ 116 (24), 99 (42), 98 (100). 1H NMR (D2O)
δ 2.97–2.88 (1H, m, CH2N), 2.83–2.74 (1H, m, CH2N), 1.48–1.36 (2H, m, H1 and H2), 1.05–0.99 (1H,
m, H3), 0.75 (1H, ddd, J=7.9, 6.6, 4.7 Hz, H3). 13C NMR (D2O) δ 182.3 (CO), 43.7 (CH2N), 22.8 (C2),
18.2 (C1), 13.0 (C3).
4.4. cis-2-(Acetamidomethyl)-1-ethoxycarbonylcyclopropane 6
Compound 6 was prepared by hydrogenation of cis-2-cyano-1-ethoxycarbonylcyclopropane catalysed
by platinum oxide in the presence of acetic anhydride following the method described by Allan et al.7
The mixture was concentrated to dryness under reduced pressure to give 6 as a light brown oil. 1H NMR
(CDCl3) δ 5.9–5.7 (1H, br s, NH), 4.15 (2H, q, J=7.2 Hz, OCH2), 3.68–3.58 (1H, m, CH2N), 3.42–3.32
(1H, m, CH2N), 1.98 (3H, s, COCH3), 1.75–1.49 (2H, m, H1 and H2), 1.29 (3H, t, J=7.2 Hz, CH3),
1.16–1.08 (1H, m, H3), 1.06–1.00 (1H, m, H3).
4.5. cis-2-(Acetamidomethyl)-1-carboxycyclopropane 7
A solution of compound 6 (15.6 g, 0.084 mol) in acetic acid (80 ml) was stirred with 6 M HCl (160 ml)
at room temperature overnight. The mixture was concentrated under reduced pressure then dried under
high vacuum over P2O5 to give cis-acid 7 as a very viscous pale brown oil (12.5 g, 95%). m/z (%) 198 (5),
[M+29]+ 186 (15), [M+1]+ 158 (41), 140 (26), 98 (100). 1H NMR (D2O) δ 3.36 (1H, dd, JAB=14.2 Hz,
JAX=6.3 Hz, CH2N), 3.15 (1H, dd, JAB=14.2 Hz, JBX=8.8 Hz, CH2N), 2.09 (3H, s, CH3CO), 1.76–1.69
(1H, m, H1), 1.60–1.40 (1H, m, H2), 1.15–1.07 (1H, m, H3), 0.97–0.91 (1H, m, H3). 13C NMR (D2O) δ
179.4, 177.8 (2×CO), 39.8 (CH2N), 22.4 (CH3), 21.2 (C2), 18.2 (C1), 13.3 (C3).
4.6. cis-2-(Acetamidomethyl)-1-((R)-2-(3,3-dimethylbutyro-1,4-lactonyl)oxycarbonyl)cyclopropane
8 and 9
To a solution of 7 (10.2 g, 65 mmol) in acetonitrile (250 ml) was added R-(−)-pantolactone (10.2 g,
78 mmol) and dicyclohexylcarbodiimide (21.8 g, 112 mmol). The mixture became homogeneous briefly
when it was brought to reflux under stirring. A white precipitate of dicyclohexyl urea was observed,
and refluxing and stirring was continued overnight. The mixture was cooled to room temperature and
filtered. The filtrate was concentrated to dryness under reduced pressure and dissolved in CH2Cl2 (100
ml). The solution was refiltered if more precipitate appeared before being subjected to purification
using vacuum chromatography13 on a column (40 mm×40 mm i.d.) packed with silica gel H60 under
reduced pressure. The mixture was eluted with 10% acetonitrile in benzene to remove most of the
unreacted pantolactone and dicyclohexylcarbodiimide, and some side products. The partly purified
diastereoisomers were separated on a longer (70 mm×40 mm i.d.) column, eluting in turn with CHCl3

2544
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
(2×100 ml), 1% MeOH in CHCl3 (2×100 ml), 2% MeOH in CHCl3 (4×100 ml), then collecting 20
ml fractions with 2% MeOH in CHCl3 until the diastereoisomers were completely eluted. Fractions
were analysed by TLC using two solvent systems, CHCl3:MeOH=95:5 and benzene:acetonitrile=4:1.
Generally, TLC plates were developed twice in the same solvent system before visualisation with 0.5%
KMnO4 in 2.5% aqueous K2CO3. Repeated column separations resulted in complete separation of the
whole mixture of diastereoisomers.
4.7. (−)-(1R,2S)-2-(Acetamidomethyl)-1-((R)-2-(3,3-dimethylbutyro-1,4-lactonyl)oxycarbonyl)cyclo-
propane 8
The less polar diastereoisomer of CAMP 8 was isolated as a colourless viscous oil (2.0 g, 11.4%),
25
[α]D −165.8 (c 1, CHCl3). IR (neat) 3610 sh, 3310 br, 3205 sh, 3379, 3289, 3083, 2969, 2935, 2878,
1778, 1737, 1669, 1549, 1467, 1445, 1404, 1389, 1375, 1338, 128, 1233 cm−1, IR (0.1 M, CCl4) 3405,
2990, 1780, 1750, 1685, 1145 cm−1, IR (0.1 M, CHCl3) 3400, 2995, 1785, 1740, 1665, 1150 cm−1
.
HRMS (CI) calcd for C13H20O5N: 270.1341. Found: 270.1341. m/z (%) [M+29]+ 298 (24), [M+1]+ 270
1
0
(100), 140 (19), 130 (7), 115 (5). H NMR (CDCl3) δ 6.95–6.85 (1H, br s, NH), 5.44 (1H, s, H2 ),
4.11 (2H, br s, OCH2), 4.32–4.23 (1H, m, CH2N), 2.68–2.59 (1H, m, CH2N), 1.988 (3H, s, CH3CO),
2.20–1.75 (2H, m, H1 and H2), 1.261 (3H, s, CH3), 1.227 (3H, s, CH3), 1.20–1.15 (2H, m, H3). 1H NMR
0
(CDCl3+D2O) δ 5.44 (1H, s, H2 ), 4.12 (2H, br s, OCH2), 4.29 (1H, dd, JAM=14.5 Hz, JAX=3.3 Hz,
CH2N), 2.64 (1H, dd, JAM=14.7 Hz, JMX=9.8 Hz, CH2N), 1.986 (3H, s, CH3CO), 2.20–1.80 (2H, m, H1
and H2), 1.261 (3H, s, CH3), 1.228 (3H, s, CH3), 1.20–1.15 (2H, m, H3). 13C NMR (CDCl3) δ 173.6,
0
0
0
171.9, 170.8 (3×CO), 76.7 (C4 ), 75.4 (C2 ), 39.7 (CH2N), 37.3 (C3 ), 23.1 (CH3CO), 22.9 (C2), 22.7
(CH3), 20.1 (CH3), 17.3 (C1), 12.6 (C3).
4.8. (+)-(1S,2R)-2-(Acetamidomethyl)-1-((R)-2-(3,3-dimethylbutyro-1,4-lactonyl)oxycarbonyl)cyclo-
propane 9
The more polar diastereoisomer of CAMP 9 was isolated as a colourless viscous oil (2.1 g, 12%),
25
[α]D +157.0 (c 1, CHCl3). IR (neat) 3568, 3391, 3297, 3210 sh, 3088, 2968, 2932, 2878, 1788, 1738,
1656, 1546, 1467, 1445, 1403, 1390, 1373, 1336, 1285, 1232 cm−1, IR (0.1 M, CHCl3) 3400, 3000,
1790, 1735, 1665, 1160 cm−1. HRMS (CI) calcd for C13H20O5N: 270.1341. Found: 270.1341. m/z (%)
1
[M+29]+ 298 (25), [M+1]+ 270 (100), 140 (27). H NMR (CDCl3) δ 6.7–6.6 (1H, br s, NH), 5.47 (1H,
0
s, H2 ), 4.18 (0.1H, d, JAB=7.2 Hz, OCH2), 4.14 (0.1H, d, JAB=7.2 Hz, OCH2), 4.12 (0.9H, d, JAB=9.1
Hz, OCH2), 4.09 (0.9H, d, JAB=9.1 Hz, OCH2), 3.95 (0.87H, ddd, J=14.4, 6.1, 5.7 Hz, CH2N), 3.62
(0.13H, ddd, J=14.3, 6.0, 6.0 Hz, CH2N), 3.37 (0.13H, ddd, J=14.3, 8.3, 6.0 Hz, CH2N), 2.95 (0.87H,
ddd, J=14.4, 9.7, 6.3 Hz, CH2N), 1.982 (3H, s, CH3CO), 2.10–1.89 (1H, br m, H2), 1.88–1.79 (1H, m,
1
H1), 1.31–1.23 (1H, m, H3), 1.234 (3H, s, CH3), 1.22–1.02 (1H, m, H3), 1.134 (3H, s, CH3). H NMR
0
(CDCl3+D2O) δ 5.47 (1H, s, H2 ), 4.18 (0.1H, d, JAB=7.1 Hz, OCH2), 4.14 (0.1H, d, JAB=7.1 Hz, OCH2),
4.12 (0.9H, d, JAB=9.1 Hz, OCH2), 4.09 (0.9H, d, JAB=9.1 Hz, OCH2), 3.95 (0.85H, dd, JAM=14.4 Hz,
JAX=5.5 Hz, CH2N), 3.62 (0.15H, dd, JAB=14.3 Hz, JAX=5.7 Hz, CH2N), 3.37 (0.15H, dd, JAB=14.3 Hz,
JBX=8.3 Hz, CH2N), 2.95 (0.85H, dd, JAM=14.4 Hz, JMX=9.6 Hz, CH2N), 2.04–1.88 (0.9H, m, H2), 1.982
(3H, s, CH3CO), 1.88–1.78 (0.9H, m, H1), 1.75–1.69 (0.1H, m, H1), 1.66–1.58 (0.1H, m, H2), 1.31–1.23
(0.9H, m, H3), 1.231 (3H, s, CH3), 1.22–1.02 (0.9H, m, H3), 1.134 (3H, s, CH3), 1.08–0.82 (0.2H, m,
H3). 13C NMR (CDCl3) δ 173.3, 171.7, 170.5 (3×CO), 76.6 (C4 ), 75.1 (C2 ), 40.0 (C3 ), 38.6 (CH2N),
0
0
0
23.1 (CH3CO), 22.9 (CH3), 22.4 (C2), 20.0 (CH3), 17.8 (C1), 14.2 (C3).

R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
2545
4.9. (−)-(1R,2S)-2-(Aminomethyl)-1-carboxycyclopropane hydrochloride 10
To a solution of the less polar (−)-diastereoisomer 8 (240 mg, 8.9 mmol) in acetic acid (15 ml) was
added 6 M HCl (45 ml) and the mixture was heated under reflux with stirring overnight. Evaporation
under reduced pressure gave a white solid. Purification (ion exchange chromatography, Dowex 50, 1
M pyridine) gave the free amino acid 10 as white powder. Its TLC (n-butanol:acetic acid:water=3:1:1,
1
ninhydrin), Rf 0.4 and H NMR (D2O) spectrum were identical to (ꢀ)-CAMP 4. The solid was
redissolved in 6 M HCl and concentrated to dryness under reduced pressure to give the hydrochloride salt
as a white crystalline solid. Recrystallisation (H2O:EtOH) gave (1R,2S)-CAMP·HCl 10 (67 mg, 48%)
as colourless rectangular plates, m.p. 244–246°C, [lit.10,11 m.p. 239–241°C, 240–241°C], [α]D −38.5
25
(c 1, 1 M HCl), [lit.10,11 [α]D −38.5 (c 0.99, 1 M HCl), [α]D −38.1 (c 1, 1 M HCl)]. m/z (%) [M+1]+
30
116 (6), 99 (52), 98 (100). Anal. calcd for C5H9O2N·HCl: C, 39.62; H, 6.65; N, 9.24. Found: C, 39.66;
H, 6.45; N, 9.14.
4.10. (+)-(1S,2R)-2-(Aminomethyl)-1-carboxycyclopropane hydrochloride 11
The more polar (+)-diastereoisomer of CAMP 9 (270 mg, 1 mmol) was converted to (1S,2R)-
CAMP·HCl 11 (120 mg, 80%), using the same procedure for the preparation of 10. TLC of the free
1
amino acid (n-butanol:acetic acid:water=3:1:1, ninhydrin) Rf 0.4 and H NMR (D2O) spectrum were
identical to (ꢀ)-CAMP 4. The hydrochloride salt 11 was recrystallised from H2O:EtOH as colourless
rectangular plates, m.p. 241–242°C, [lit.10 m.p. 242–243°C], [α]D25 +39.0 (c 1, 1 M HCl), [lit.10 [α]D
26
+37.3 (c 0.99, 1 M HCl)]. m/z (%) [M+1]+ 116 (9), 99 (29), 98 (100). Anal. calcd for C5H9O2N·HCl: C,
39.62; H, 6.65; N, 9.24. Found: C, 39.88; H, 6.25; N, 9.04.
4.11. trans-2-(Acetamidomethyl)-1-ethoxycarbonylcyclopropane 12
Compound 12 was prepared by hydrogenation of ethyl trans-2-cyano-1-ethoxycarbonylcyclopropane
over platinum oxide in the presence of acetic anhydride using the method described by Allan et al.7 The
1
mixture was concentrated to dryness under reduced pressure to give 12 as a pale brown oil. H NMR
(CDCl3) δ 4.13 (2H, q, J=7.1Hz, OCH2), 3.34–3.23 (1H, m, CH2N), 3.18–3.09 (1H, m, CH2N), 2.00
(3H, s, CH3CO), 1.70–1.56 (1H, m, H2), 1.57–1.51 (1H, m, H1), 1.26 (3H, t, J=7.1Hz, CH3), 1.24–1.17
(1H, m, H3), 0.86 (1H, ddd, J=8.5, 6.2, 4.4 Hz, H3). 13C NMR (CDCl3) δ 173.5, 170.2 (2×CO), 60.6
(OCH2), 42.1 (CH2N), 23.1 (C2), 21.6 (COCH3), 19.0 (C1), 14.2 (CH3), 13.7 (C3).
4.12. trans-2-(Acetamidomethyl)-1-carboxycyclopropane 13
The trans-ester 12 was converted to the trans-acid 13 using the procedure described for the preparation
of the cis-acid 7. The trans-acid 13 obtained initially as a very viscous colourless oil (90%) which could
be crystallised from either acetone or acetonitrile. Crystallisation from acetone gave 13 as colourless
needles, m.p. 157–159°C (subl. from 150°C). m/z (%) [M+29]+ 186 (14), [M+1]+ 158 (21), 140 (100).
1H NMR (D2O) δ 3.18 (1H, dd, JAB=14.2 Hz, JAX=6.0 Hz, CH2N), 2.98 (1H, dd, JAB=14.2 Hz, JBX=7.1
Hz, CH2N), 1.89 (3H, s, CH3CO), 1.59–1.42 (1H, m, H2), 1.52–1.46 (1H, m, H1), 1.14–1.08 (1H, m,
H3), 0.89 (1H, ddd, J=8.5, 6.3, 4.8 Hz, H3). 13C NMR (D2O) δ 179.3, 174.8 (2×CO), 42.4 (CH2N), 22.6
(CH3), 22.5 (C2), 19.3 (C1), 14.5 (C3). Anal. calcd for C7H11O3N: C, 53.49; H, 7.05; N 8.91. Found: C,
53.09; H, 7.09; N, 9.18.

2546
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
4.13. trans-2-(Acetamidomethyl)-1-((R)-2-(3,3-dimethylbutyro-1,4-lactonyl)oxycarbonyl)cyclopropane
14 and 15
trans-Acetamido acid 13 (4.7 g, 30 mmol) was converted to diastereoisomers 14 and 15 using the pro-
cedure described for the preparation of diastereoisomers 8 and 9. Compounds 14 and 15 were separated
using the same method as described for 8 and 9 with a slight increase in MeOH concentration to 3%.
The separation between the TAMP-diastereoisomers was less than that of the CAMP-diastereoisomers
and repeated column separations were required to achieve complete separation of the whole mixture of
the diastereoisomers.
4.14. (−)-(1R,2R)-2-(Acetamidomethyl)-1-((R)-2-(3,3-dimethylbutyro-1,4-lactonyl)oxycarbonyl)cyclo-
propane 14
25
The less polar diastereoisomer 14 was isolated as a colourless viscous oil (2 g, 23%), [α]D −53.5 (c
1, CHCl3). IR (neat) 3558, 3375, 3292, 3205 sh, 3085, 2969, 2933, 2878, 1792, 1743, 1655, 1549, 1466,
1452, 1413, 1372, 1316, 1296, 1232, 1204 cm−1, IR (0.1 M, CHCl3) 3460, 2990, 1795, 1745, 1670,
1160 cm−1. HRMS (CI) calcd for C13H20O5N: 270.1341. Found: 270.1341. m/z (%) [M+29]+ 298 (33),
+
1
0
[M+1] 270 (100), 140 (81). H NMR (CDCl3) δ 5.9–5.8 (1H, br s, NH), 5.36 (1H, s, H2 ), 4.07 (1H,
d, JAB=10.9 Hz, OCH2) 4.04 (1H, d, JAB=10.9 Hz, OCH2), 3.52–3.44 (1H, m, CH2N), 3.07–2.98 (1H,
m, CH2N), 2.014 (3H, s, CH3CO), 1.76–1.66 (2H, m, H1 and H2), 1.36–1.30 (1H, m, H3), 1.207 (3H,
1
0
s, CH3), 1.134 (3H, s, CH3), 1.04–0.98 (1H, m, H3). H NMR (CDCl3+D2O) δ 5.37 (1H, s, H2 ), 4.08
(1H, d, JAB=9.1 Hz, OCH2), 4.04 (1H, d, JAB=9.1 Hz, OCH2), 3.49 (1H, dd, JAB=14.1 Hz, JAX=5.2 Hz,
CH2N), 3.00 (1H, dd, JAB=14.1 Hz, JBX=7.8 Hz, CH2N), 2.01 (3H, s, CH3CO), 1.76–1.66 (2H, m, H1
and H2), 1.36–1.30 (1H, m, H3), 1.208 (3H, s, CH3), 1.137 (3H, s, CH3), 1.04–0.98 (1H, m, H3). 13C
0
0
0
NMR (CDCl3) δ 172.6, 170.2 (2×CO), 76.3 (C4 ), 75.2 (C2 ), 42.0 (CH2N), 40.2 (C3 ), 23.2 (CH3CO),
23.0 (CH3), 22.9 (C2), 19.9 (CH3), 18.7 (C1), 14.7 (C3).
4.15. (+)-(1S,2S)-2-(Acetamidomethyl)-1-((R)-2-(3,3-dimethylbutyro-1,4-lactonyl)oxycarbonyl)cyclo-
propane 15
The more polar diastereoisomer 15 was isolated as a colourless viscous oil (1.5 g, 17%), [α]D25 +57.4
(c 1, CHCl3). IR (neat) 3568, 3391, 3297, 3205 sh, 3088, 2968, 2932, 2878, 1788, 1738, 1657, 1548,
1466, 1453, 1413, 1370, 1314, 1295, 1264, 1203 cm−1, IR (0.1 M, CHCl3) 3460, 2990, 1795, 1740,
1670, 1160 cm−1. HRMS (CI) calcd for C13H20O5N: 270.1341. Found: 270.1341. m/z (%) [M+29]+ 298
+
1
0
(40), [M+1] 270 (97), 140 (100). H NMR (CDCl3) δ 5.82–5.63 (1H, br s, NH), 5.34 (1H, s, H2 ), 4.06
(1H, d, JAB=9.1 Hz, OCH2), 4.03 (1H, d, JAB=9.1 Hz, OCH2), 3.35–3.17 (2H, m, CH2N), 2.007 (3H, s,
CH3CO), 1.76–1.66 (2H, m, H1 and H2), 1.32 (1H, ddd, J=8.7, 8.7, 4.7 Hz, H3), 1.212 (3H, s, CH3),
1
0
1.117 (3H, s, CH3), 1.00 (1H, ddd, J=8.5, 6.4, 4.6 Hz, H3). H NMR (CDCl3+D2O) δ 5.34 (1H, s, H2 ),
4.06 (1H, d, JAB=9.0 Hz, OCH2), 4.03 (1H, d, JAB=9.0 Hz, OCH2), 3.29 (1H, dd, JAB=14.3 Hz, JAX=6.6
Hz, CH2N), 3.27 (1H, dd, JAB=14.3 Hz, JBX=6.8 Hz, CH2N), 2.003 (3H, s, CH3CO), 1.80–1.66 (2H, m,
H1 and H2), 1.31 (1H, ddd, J=8.8, 8.7, 4.7 Hz, H3), 1.212 (3H, s, CH3), 1.117 (3H, s, CH3), 1.00 (1H,
ddd, J=8.4, 6.4, 4.7 Hz, H3). 13C NMR (CDCl3) δ 172.8, 172.7, 170.6 (3×CO), 76.2 (C4 ), 75.1 (C2 ),
0
0
0
41.9 (CH2N), 40.2 (C3 ), 23.2 (CH3CO), 23.1 (CH3), 22.7 (C2), 19.9 (CH3), 18.7 (C1), 14.6 (C3).

R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
2547
4.16. (−)-(1R,2R)-2-(Aminomethyl)-1-carboxycyclopropane 16
The less polar (−)-diastereoisomer 14 (99 mg, 0.37 mmol) was converted to (1R,2R)-TAMP 16 (41
mg, 96%) using the same procedure for the preparation of 10. Its TLC (n-butanol:acetic acid:water=3:1:1,
ninhydrin), Rf 0.4 and 1H NMR (D2O) spectrum were identical to (ꢀ)-TAMP 5. (1R,2R)-TAMP 16 was
25
recrystallised from H2O:EtOH as square colourless plates, m.p. 270–273°C (subl. from 250°C), [α]D
−81.8 (c 1, 1 M HCl), [lit.10 (1R,2R)-TAMP·HCl, [α]D −65.5 (c 0.95, 1 M HCl)]. m/z (%) [M+1]+
26
116 (75), 100 (26), 99 (64), 98 (100). Anal calcd for C5H9O2N: C, 52.16; H, 7.88; N, 12.17. Found: C,
51.90; H, 8.07; N, 12.16.
4.17. (+)-(1S,2S)-2-(Aminomethyl)-1-carboxycyclopropane 17
The more polar (+)-diastereoisomer 15 (120 mg, 0.45 mmol) was converted to (1S,2S)-TAMP 17 (38
mg, 74%) using the same procedure for the preparation of 10. Its TLC (n-butanol:acetic acid:water=3:1:1,
ninhydrin), Rf 0.4 and 1H NMR (D2O) spectrum were identical to (ꢀ)-TAMP 5. Recrystallisation from
H2O:EtOH gave (1S,2S)-TAMP 17 as square colourless plates, m.p. 264–265°C (subl. from 230°C);
25
[α]D +81.6 (c 1, 1 M HCl). m/z (%) [M+1]+ 116 (19), 100 (40), 99 (52), 98 (100). Anal. calcd for
C5H9O2N: C, 52.16; H, 7.88; N, 12.17. Found: C, 51.82; H, 7.69; N, 12.20.
Acknowledgements
We thank Dr. Bradley M. Collins, Department of Chemistry for recording FT-IR reflection infrared
spectra, Dr. Hue W. Tran, Department of Pharmacology for valuble technical assistance and the NSW
Centre for Parallel Computing for computing time on the Power Challenge. Support for this work by the
Australian National Health and Medical Research Council grant is gratefully acknowledged.
References
1. Johnston G. A. R. Pharmacol. Ther. 1996, 69, 173–198.
2. Kerr, D. I. B.; Ong, J. Pharmacol. Ther. 1995, 67, 187–146.
3. Johnston G. A. R. Trends Pharmacol. Sci. 1996, 17, 319–323.
4. Bormann, J.; Feigenspan, A. Trends Neurosci. 1996, 17, 319–323.
5. Murata, Y.; Woodward, R. M.; Miledi, R.; Overman, L. E. Bioorg. Med. Chem. Lett. 1996, 6, 2071–2076.
6. Ragozzino, D.; Woodward, R. M.; Murata, F.; Eusebi, F.; Overman, L. E.; Miledi, R. Mol. Pharmacol. 1996, 50,
1024–1030.
7. Allan, R. D.; Curtis, D. R.; Headley, P. M.; Johnston, G. A. R.; Lodge, D.; Twitchin, B. J. Neurochem. 1980, 34, 652–654.
8. Kusama, T.; Spivak, C. E.; Whiting, P.; Dawson, V. L.; Schaeffer, J. C.; Uhl, G. R. Br. J. Pharmacol. 1993, 109, 200–206.
9. Kusama, T.; Wang, T.-L.; Guggino, W. B.; Cutting, G. R.; Uhl, G. R. Eur. J. Pharmacol. 1993, 245, 83–84.
10. Morikawa, T.; Sasaki, H.; Hanai, R.; Shibuya, A.; Taguchi, T. J. Org. Chem. 1994, 59, 97–103.
11. Galeazzi, R.; Mobbili, G.; Orena, M. Tetrahedron: Asymmetry 1997, 8, 133–137.
12. Duke, C. C.; Wells, R. J. Aust. J. Chem. 1987, 40, 1641–1654.
13. Ravi, B. N.; Wells, R. J. Aust. J. Chem. 1982, 35, 129–144.
14. Hehre, W. J. Spartan 5.0; Wavefunction Inc., Irvine, CA, 1997.
15. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill, P. M. W.; Johnson, B. J.; Robb, M. A.; Cheeseman, J. R.; Kieth, T.;
Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J.
B.; Ciolowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.;
Andres, J. L.; Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart, J.
P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A. Gaussian 94, Revision D.3; Gaussian Inc., Pittsburgh, PA, 1995.

2548
R. K. Duke et al. / Tetrahedron: Asymmetry 9 (1998) 2533–2548
16. Foresman, J. B.; Frisch A. Exploring Chemistry with Electronic Structure Methods, 2nd Edn; Gaussian Inc., Pittsburgh,
PA, 1996; p. 127.
17. Wolinski, K.; Hilton, J. F.; Pulay, P. J. Am. Chem. Soc. 1990, 112, 8251–8260.
18. Ditchfield, R. Mol. Phys. 1974, 27, 789–807.
19. Cheeseman, J. R.; Trucks, G. W.; Kieth, T. A.; Frisch, M. J. J. Chem. Phys. 1996, 104, 5497–5509.