Radical 4-exo cyclizations onto O-alkyloxime acceptors: Towards the synthesis of penicillin-containing antibiotics

Article abstract of DOI:10.1002/hlca.200690202

The 4-exo cyclizations of two types of carbamoyl radicals onto O-alkyloxime acceptor groups were studied as potential routes to 3-amino-substituted azetidinones and hence to penicillins. A general synthetic route to 'benzaldehyde oxime oxalate amides' (=2

Full text of DOI:10.1002/hlca.200690202

Helvetica Chimica Acta – Vol. 89 (2006)
2133
Radical 4-exo Cyclizations onto O-Alkyloxime Acceptors: Towards the
Synthesis of Penicillin-Containing Antibiotics
by Eoin M. Scanlan1) and John C. Walton*
University of St. Andrews, EaStChem, School of Chemistry, St. Andrews, Fife, UK, KY169ST
(phone: +44 1334 463864; fax: +44 1334 463808; e-mail: jcw@st-and.ac.uk)
Dedicated to the memory of Hanns Fischer and his outstanding contribution to scientific excellence
The 4-exo cyclizations of two types of carbamoyl radicals onto O-alkyloxime acceptor groups were
studied as potential routes to 3-amino-substituted azetidinones and hence to penicillins. A general syn-
thetic route to ꢀbenzaldehyde oxime oxalate amidesꢁ (=2-[(benzylideneamino)oxy]-2-oxoacetamides;
see, e.g., 10c) of 2-{[(benzyloxy)imino]methyl}-substituted thiazolidine-4-carboxylic acid methyl esters
9 was developed (Scheme 3). It was shown by EPR spectroscopy that these compounds underwent sen-
sitized photodissociation to the corresponding carbamoyl radicals but that these did not ring close. An
analogous open-chain precursor, benzaldehyde O-(benzylaminoacetaldehyde-O-benzyloxalyl)oxime,
15, lacking the 5-membered thiazolidine ring, was shown by EPR spectroscopy to release the correspond-
ing carbamoyl radical (Scheme 4). The latter underwent 4-exo cyclization onto its C=NOBn bond in non-
H-atom donor solvents. The rate constant for this cyclization was determined by the steady-state EPR
method. Spectroscopic evidence indicated that the reverse ring-opening process was slower than cycliza-
tion.
Introduction. – The cyclobutane ring is an important structural motif found in many
natural products. The cyclizations of pent-4-enyl and pentadienyl radicals in the 4-exo
mode (C4x) to give cyclobutylmethyl and cyclobutenylmethyl radicals, respectively, are
comparatively slow [1], due to the strain in the 4-membered ring, and the reverse ring-
opening processes are about three orders of magnitude faster [2–5]. However, the
equilibrium can be shifted in favor of ring-closed species either by rapid trapping of
the cyclized radical, or by 2,2-dimethyl substitution, and/or by inclusion of electron-
withdrawing substituents at the terminus of the alkene [6][7]. For example, the recently
established general stereoselective syntheses of cyclobutanols and cyclobutanones,
involving samarium(II) reductions of unsaturated aldehydes [8][9], and titanium(III)
reductions of unsaturated epoxides [10][11], owe their success to the rapid trapping
tactic.
The 4-membered azetidinone ring system 1 occurs in several families of powerful b-
lactam antibiotics. Radical-based cyclizations leading to azetidinones seem to differ
from the 4-exo norm, because no less than four distinct radical-based disconnections
have been investigated [7], although only the three shown in Scheme 1 have proved suc-
cessful in synthetic applications.
1
)
Present address: University of Oxford, Chemistry Research Laboratory, Mansfield Road, Oxford,
OX1 3TA, UK.
ꢂ 2006Verlag Helvetica Chimica Acta AG, Zürich

2134
Helvetica Chimica Acta – Vol. 89 (2006)
Scheme 1. Synthetically Useful Radical-Based Disconnections of the Azetidinone Ring
Disconnection a in Scheme 1 implies a C4x closure of an amidoalkyl (a-carbamoyl)
radical 2 onto an enamide acceptor. This approach has been employed in numerous
successful b-lactam syntheses [6]. Disconnection c connotes C4x ring closure of an
acyl radical 4 onto an imine acceptor. Ryu and co-workers have prepared suitable
acyl radicals and made successful syntheses by carbonylations of azaenynes [12][13].
Disconnection b implies C4x closure of a carbamoyl-type radical (aminoacyl radical)
3 onto an unsaturated acceptor group.
Quite a range of b-lactams have been made by methodologies following disconnec-
tion b. Pattenden and co-workers made carbamoyl(salophen)cobalt complexes and
showed that on photolysis carbamoyl radicals were released and underwent C4x cycliza-
tions [14][15] (salophen=2,2’-[1,2-phenylenebis(nitrilomethylidyne]bis[phenol]). We
have recently developed several ꢀcleanꢁ, metal-free routes to carbamoyl radicals. Cyclo-
hexadiene-based reagents have proved useful as precursors of C- and heteroatom-cen-
tered radicals [16][17]. Cyclohexadienecarboxamides released carbamoyl radicals, and
those with allyl and cinnamyl side chains afforded azetidinones in moderate yields
[18–20]. ꢀOxime oxalate amidesꢁ (=2-[(alkylideneamino)oxy]-2-oxoacetamides)
R1C(R2)=NÀOÀC(O)ÀC(O)ÀN(R3)R4 are easily prepared from an oxime, oxalyl
chloride, and an N-alkenylamine, and they also function as novel, clean sources of car-
bamoyl radicals [21]. The best conditions for preparations of lactams involved photol-
yses of dilute ꢀoxime oxalate amideꢁ solutions in toluene, with a 2–3-fold excess of 4-
methoxyacetophenone (=1-(4-methoxyphenyl)ethanone; MAP). In these photosensi-
tized reactions, sufficient energy was transferred to the ꢀoxime oxalate amideꢁ to break
the weak NÀO bond and release the phenyliminyl radical PhCH=NC (R1 =Ph, R2 =H)
together with a carbamoyl radical (R4N(R3)ÀC(O)C). The phenyliminyl radicals simply
abstracted an H-atom from the toluene solvent, and the resulting imine was hydrolyzed
to benzaldehyde during workup. By means of this methodology, several b-lactams were
prepared, including a bicyclic example [22].
An attempt was made to access penicillin derivatives by this route. Thiazolidine-
containing ꢀoxime oxalate amideꢁ 5 was prepared and photolyzed in the presence of
MAP in toluene (Scheme 2). EPR Spectroscopy showed unequivocally that the carba-
moyl radical 6 was generated. GC/MS Evidence suggested the penicillin derived from
radical 7 had formed, but the yield was low and none could be isolated [22].
Cyclizations onto O-alkyloxime functional groups >C=NOR are regiospecific to
the C-atom [23] and can be significantly faster than with analogous alkene acceptors
[24]. Moreover, useful N-functionality remains available for further synthetic elabora-

Helvetica Chimica Acta – Vol. 89 (2006)
2135
Scheme 2. Putative Synthesis of a Penicillin Derivative from Thiazolidine-Containing ꢀOxime Oxalate
Amideꢁ 5
tion. We thought, therefore, that C4x ring closures onto suitable O-alkyloximes to give
(oxoazetidinyl)aminyl radicals, and hence 3-amino-b-lactams, could show advantages
over previous methods. Accordingly, we prepared two types of ꢀoxime oxalate amideꢁ
containing O-benzyloxime side chains and investigated their photodissociations. Part
of this research has been published in a preliminary communication [25].
Results and Discussion. – A synthetic route to thiazolidines incorporating O-alkyl-
oxime side chains at the 2-position was first devised. Glyoxal oxime 8 was prepared by
condensation of O-benzylhydroxylamine hydrochloride with glyoxal (=ethanedial)
according to the method of Zimmermann and co-workers [26]. The {[(benzyloxy)imi-
no]methyl}thiazolidines 9 were prepared by addition of cysteine (R1 =H) or penicill-
amine (R1 =Me) esters (or their HCl salts) to a solution of oxime 8 in EtOH, followed
by stirring at room temperature for 24 h (Scheme 3). Thiazolidines 9a–c were all
obtained as inseparable 1:1 mixtures of isomers, and 9c was unstable and had to be
used immediately. The thiazolidine 9c was converted to the corresponding ꢀoxime oxa-
late amideꢁ 10c through reaction with benzaldehyde O-(chlorooxalyl)oxime (=2-[(ben-
zylideneamino)oxy]-2-oxoacetyl chloride). Precursor 10c was obtained as a colorless
1
oil, and its H-NMR spectrum showed it to be a mixture of isomers.
By analogy with other ꢀoxime oxalate amidesꢁ, it was expected that photosensitized
dissociation of 10c would occur at the weak NÀO bond followed by rapid CO2 loss to
release a phenyliminyl radical 11 and carbamoyl radical 12c (Scheme 3). Potentially,
radical 12c could undergo C4x ring closure to afford ꢀpenicillano-aminylꢁ radical 13c
that should abstract an H-atom from the toluene solvent and hence yield the penicillin
derivative 14c, with N-functionality conveniently placed at the 3-position.
The photodecomposition of 10c was first examined by EPR spectroscopy. A solu-
tion containing 10c (0.05 g, 0.1 mmol) and MAP (1 equiv.) in (tert-butyl)benzene (0.2
ml) was placed in a quartz tube and deaerated by bubbling N2. The tube was transferred
to the resonant cavity of a 9-GHz EPR spectrometer and photolyzed with unfiltered
light from a 500-W Hg arc. The spectrum shown in Fig. 1,a was obtained at 220 K.
The two sets of N-triplets at low and high field are clearly due to the phenyliminyl rad-
ical 11, and the EPR parameters (g=2.0034, a(1H)=7.97 mT, a(N)=0.98 mT) were
essentially identical to those in the literature [27]. The g-factor and N hyperfine split-
ting (hfs) of the second radical (Fig. 1,a, centre) were very similar to those of other car-
bamoyl radicals [21][22] which indicated that the spectrum was due to radical 12c. The
spectrum also displays a small long-range d hfs from one or other of the nonequivalent
g-H-atoms. A density-functional-theory (DFT) computation (UB3LYP/6-31G*) [28]

2136
Helvetica Chimica Acta – Vol. 89 (2006)
Scheme 3. Preparation of an O-Alkyloxime-Substituted ꢀOxime Oxalate Amideꢁ and Its Proposed
Photochemical Transformation
a) BnONH2 ·HCl, NaOH, r.t., 16h. b) Et3N, r.t., 24 h. c) Pyridine, CH2Cl2, 3 h. d) UV, 4-methoxyace-
G
tophenone (MAP), PhMe.
on model radical 12d (Fig. 2) gave hfs in good agreement with the experimental data.
The computed hfs for Hg4 (see Scheme 3) was considerably larger than the computed
hfs for Hg2 (Table 1). This explains the origin of the d hfs and accordingly, we have
assigned the experimental hfs to Hg4. These spectroscopic results indicated that
ꢀoxime ester amideꢁ 10c underwent ready homolysis of its weak NÀO bonds and that
loss of CO2 from the initial acyloxyl radicals was rapid. At higher temperatures, the
EPR spectra weakened, as expected, eventually leaving no sign of carbamoyl 12c
(Fig. 1,b). However, no trace of the ring-closed penicillano-aminyl radical 13c was
observable.
t
Table 1. EPR Parameters for Thiazolidinylcarbamoyl Radicals in BuPh solution
Radical
Temp./K or method
g-Factor
a(N)/mT
a(Hg2)/mT
a(Hg4)/mT
12c
12a
12d
220
220
DFTa)
2.0014
2.0017
2.22
2.22
2.20
0.10
0.15
À0.16
À0.06
a) UB3LYP/6-311++G(d,p)//UB3LYP/6-31G*
Attempts to carry out preparative-scale photolyses in toluene as a suitable H-donor
were hampered by the comparatively rapid degradation of the precursor 10c. 1H-NMR
Analyses showed that solutions of 10c rapidly developed by-products when kept at or

Helvetica Chimica Acta – Vol. 89 (2006)
2137
Fig. 1. EPR Spectra obtained on photolysis of tBuPh solutions of 10c containing MAP at a) 220 K
and b) 290 K
Fig 2. DFT-Computed structure of model radical 12d (UB3LYP/6-31G*)
above ca. 270 K. A photosensitized reaction with 10c and 2 equiv. of MAP in toluene
was carried out at room temperature, but the NMR spectrum showed no trace of
14c, and a GC/MS of the whole product mixture showed a complex spread of individ-
ually minor components. Disappointingly, therefore, no evidence of C4x ring closure of
thiazolidinylcarbamoyl radicals onto the oxime ether acceptor could be obtained. The
lack of cyclization could be due to precursor degradation, or to the extra strain imposed
on the C4x process by the presence of the adjacent 5-membered ring, or to an interaction
of the carbamoyl radical centre with the neighboring ester group. The structure of
model thiazolidine-containing carbamoyl 12d, computed at the UB3LYP/6-31G*
level (Fig. 2), showed an all-trans arrangement of the O-alkyloxime chain with the car-
bamoyl C=O well placed for ring closure. However, the ester group approaches the car-

2138
Helvetica Chimica Acta – Vol. 89 (2006)
bamoyl rather closely and adopts a conformation in which the two C=O groups are
nearly parallel; supporting the idea of an interaction.
To probe for C4x ring closure onto a precursor without an adjacent 5-membered
ring, the open-chain ꢀoxime oxalate amideꢁ 15 was prepared as described previously
[25]. The EPR spectrum obtained on photolysis of 15 and MAP in (tert-butyl)benzene
at 280 K is shown in Fig. 3,b (see also Scheme 4). The signals of the phenyliminyl rad-
ical are clearly visible in the wings of the spectrum. Another species characterized by
t
Fig. 3. Isotropic EPR spectra obtained during UV photolysis of a BuPh solution of 15 and MAP: a)
experimental spectrum at 220 K, b) experimental spectrum at 280 K, and c) computer simulation
Scheme 4. Ring Closure (C4x) onto an O-Alkyloxime Acceptor

Helvetica Chimica Acta – Vol. 89 (2006)
2139
four m is also present at 280 K and was satisfactorily simulated (Fig. 3,b) with the fol-
lowing parameters: g=2.0049, a(N)=1.37, a(1Hb)=1.37, a(2Hg)=0.25, a(2H)=0.1
mT at 280 K). The g-factor is typical of an alkoxyaminyl radical, and the other hfs
are as expected for (oxoazetidinyl)aminyl radical 17. This assignment was supported
by a DFT computation (UB3LYP/EPR-iii//UB3LYP/6-31G*) which gave hfs in reason-
able agreement viz.: a(N)=1.16, a(1Hb)=1.85, a(2Hg)=0.15, a(2H)=À0.03 mT. A
third radical (marked A in Fig. 3) was also present and dominated spectra taken at
lower temperatures (Fig. 3,a). The EPR parameters, i.e., g=2.0018, a(N)=2.32 mT,
showed this to be the carbamoyl radical 16 (the other peaks from 16 are overlapped
by the spectrum of 17 in Fig. 3,b). The spectroscopic evidence strongly supports the
conclusion that C4x ring closure of the open-chain carbamoyl radical 16 does indeed
rapidly take place onto the O-benzyloxime acceptor; even at temperatures well
below room temperature.
Competition from the reverse ring fission process (kf) needs to be considered in
assessing the kinetics. When ring opening is included, the kinetic expression of Eqn.
1 is easily derived, where 2kt is the diffusion-controlled rate constant for termination
reactions of 16 and 17 [29][30]. The concentrations of 16 and 17 were determined by
double integration of their EPR spectra and comparison with the spectrum of a
known concentration of DPPH run under identical conditions [4]. Experiments were
carried out with MAP concentrations ranging from 0.1 to 5 equiv. so as to vary the
ratio [16]/[17]. It was found that, to within experimental error, the results were not sen-
sitive to this ratio, and hence the second term on the right of Eqn. 1 is small, meaning
that kf is significantly less than kc. Under these conditions, kc values were derived in the
temperature range 205–280 K from the simplified form of Eqn. 1 giving Arrhenius
parameters of log(Ac/sÀ1)=9.6and Ea =30 kJ molÀ1.
[17]+[17]2/[16]=kc/2kt Àkf/2kt
A
(1)
The rate constants for ring closure, and the reverse where available, are compared
in Table 2 with data for some related species. The rate constant for C4x ring closure of
carbamoyl 16 onto the C=NOBn acceptor is four orders of magnitude greater than the
rate constant for the archetype pent-4-enyl (20) C4x cyclization. Furthermore, it is also
faster than cyclization of pent-4-enyl radical 21, even though this is favored by dimethyl
substitution and produces a resonance-stabilized product radical. This is in good agree-
ment with the deduction from preparative work that azetidinone rings form more easily
than cyclobutyl-type rings in radical reactions. Ring opening is of the same order of
magnitude for both types of ring, and the main difference is the faster cyclization giving
azetidinones.
As anticipated, the 4-exo cyclization of 16 has a smaller rate constant than the 5-exo
cyclizations of the hex-5-enyl radical 23. Interestingly, carbamoyl 4-exo cyclizations

2140
Helvetica Chimica Acta – Vol. 89 (2006)
Table 2. Rate Constants for C4x and C5x Ring Closures at 300 K
Radical
kc/sÀ1
kf/sÀ1
Ref.
20
21
16
22
23
24
1
103
[1–4]
[31]
this work
[25]
[32]
[33]
2·104a)
3·104
5·104
2·105
ca. 103
4·107b
)
a) At 323 K. b) At 353 K.
were found to be less reversible than 4-exo cyclizations of amidoalkyls for which there
is much evidence that azetidinone formation is reversible [34]. The spectroscopic evi-
dence clearly indicated that 4-exo closure of the open-chain carbamoyl radical 16 was
faster than 4-exo ring closure of carbamoyls 6 or 12c with adjacent thiazolidine rings.
The 4-exo cyclization onto a >C=NOR bond was expected to be faster than onto a
C=C bond. The error limits are large on the kinetic data for carbamoyl radicals 16
and 22 so that, although Table 2 shows a marginally greater kc for the latter, the two
are indistinguishable within the error limits. The two specific examples are stereoelec-
tronically dissimilar so that a general conclusion about the kinetics of 4-exo ring clo-
sures onto the two types of acceptors cannot be made.
Attempts were made to devise a suitable methodology for the preparative use of O-
alkyloxime containing ꢀoxime oxalate amidesꢁ. A solution of the ꢀoxime oxalate amideꢁ
15 in toluene containing MAP was photolyzed by light from a 400-W UV lamp for 5 h at
room temperature. The solvent was removed and the resulting yellow oil was analyzed
by 1H-NMR spectroscopy. The HÀC(3) of the azetidinone ring usually appear between
d 3.0 and 4.0 ppm. This region of the spectrum was clear suggesting that the b-lactam 18
was not formed. The peaks corresponding to the benzaldehyde oxime H-atom had dis-
appeared and were replaced by a s at d 10.2 corresponding to benzaldehyde (derived
from radical 11). This suggested that the photodissociation had actually occurred in
accordance with Scheme 4 but that the cyclized product had not formed. The remaining
peaks were closely related to those of the amine precursor to 15. In addition, the pres-
ence of a new series of peaks at d 8.2 suggested that the major product was the form-
amide 19 resulting from direct H-abstraction by the carbamoyl radical. Formation of
19 was unexpected since the EPR study had demonstrated that cyclization occurred
readily at low temperatures in the non-H-donating solvent (tert-butyl)benzene. It is
probable that cyclization of 16 could not compete with the H-atom abstraction from
the toluene solvent. The reaction was repeated at a higher temperature, but once
again, only the formamide was observed. The photolysis was repeated in the non-H-
donating solvent trifluorotoluene (=(trifluoromethyl)benzene), and also in THF and
in trichlorobromomethane, but the expected azetidinone NMR peaks were not
observed, and a complex mixture resulted in each case.

Helvetica Chimica Acta – Vol. 89 (2006)
2141
Experimental Part
General. Column Chromatography (CC): ICN silica gel (63–200, 60 Š); FC=flash CC. IR spectra:
1
in nujol or neat; Perkin-Elmer RX-I-FT-IR spectrometer; in cmÀ1. H (300 MHz)- and 13C (75 MHz)-
NMR Spectra: CDCl3 soln.; d in ppm rel. to SiMe4 (=0 ppm) as an internal standard, J in Hz; Varian-
Gemini-2000 and Bruker Avance-300 spectrometers. EI-MS and high-resolution (HR) MS: 70 eV ioniza-
tion, or chemical ionization (CI) with isobutane as target gas; VG-Autospec spectrometer; in m/z (rel.
%).
EPR Spectra. EPR spectra were obtained with a Bruker EMX-10/12 spectrometer operating at 9.5
GHz with 100 kHz modulation. Solns. of freshly purified ꢀoxime oxalate amideꢁ (ca. 0.1–0.2M) and 4-
methoxyacetophenone (MAP; usually 1 mol-equiv.) in (tert-butyl)benzene were placed in 4-mm (o.d.)
quartz tubes and deaerated by bubbling N2 gas for 20 min. Samples were irradiated in the resonant cavity
by unfiltered light from a 500-W super-pressure Hg arc. In all cases where spectra were obtained, hfs were
assigned with the aid of computer simulations by the Bruker SimFonia software package. For concentra-
tion measurements, signals were double-integrated with the Bruker WinEPR software, and radical con-
centrations were calculated by reference to a known concentration of DPPH (2,2-diphenyl-1-(2,4,6-trini-
trophenyl)hydrazyl), as described previously. In kinetic experiments, to minimize sample-depletion
effects, ꢀsingle-shotꢁ runs were carried out where samples were irradiated for only one spectrum (<5
min) before a fresh sample was introduced.
Glyoxal Mono-O-benzyloxime (=Ethanedial Mono-O-(phenylmethyl)oxime; 8). O-Benzylhydroxyl-
amine hydrochloride (800 mg, 5 mmol) was dissolved in H2O (50 ml) neutralized with NaOH (3 mol-
equiv.). To this was added glyoxal (7.5 g of a 40% soln. in H2O). The mixture was left at r.t. for 16h
after which time it was extracted with CH2Cl2 (4ꢃ25 ml). The org. phase was dried (MgSO4), the solvent
evaporated, and the crude yellow oil purified by bulb-to-bulb distillation at 60–628 at 0.2 Torr: 8 (515 mg,
63%) ([26]: 63%). Colorless oil. 1H-NMR: 5.33 (s, CH2); 7.34–7.80 (m, 5 arom. H); 7.56( d, J=8,
CH=N); 9.59, (d, J=8, CH=O).
2-{[(Benzyloxy)imino]methyl}-1,3-thiazolidine-4-carboxylic Acid Benzyl Ester (9a). To a soln. of the
benzyl ester of L-cysteine (2.11 g, 10 mmol) in EtOH (50 ml) at 08 was added 8 (1.63 g, 10 mmol). The
mixture was stirred at r.t. for 24 h. After this time, the solvent was removed and the crude colorless
oil subjected to CC (AcOEt/hexane). The product was further purified by recrystallization from
AcOEt/hexane: 9a as a 1:1 mixture of isomers (3.06g, 86%). Colorless needles. M.p. 78–80 8. IR
1
(NaCl): 1741 (C=O). H-NMR: 2.94 (dd, J=7, 10, 1/2 H, CH2(5)); 3.02 (dd, J=7, 9, 1/2 H, CH2(5));
3.29 (dd, J=7, 10, 1/2 H, CH2(5)); 3.37 (dd, J=7, 10, 1/2 H, CH2(5)); 3.95 (dd, J=7, 9, 1/2 H, HÀ
C(4)); 4.11 (t, J=6, 1/2 H, HÀC(4)); 5.05, 5.09 (2s, 2 H, OCH2); 5.12 (d, J=6, 1/2 H, HC=N); 5.20,
5.22 (2s, 2 H, OCH2); 5.30 (d, J=6, 1/2 H, HC=N); 7.30–7.37 (m, 10 arom. H); 7.33 (1/2 H, NCHS,
under arom. H); 7.48 (d, J=6, 1/2 H, NCHS). 13C-NMR (CDCl3): 37.9, 38.8 (C(5)); 64.9, 65.7, 66.1,
66.5 (C(4), C(2)); 76.6, 76.9 (OCH2); 128.3, 128.4, 128.5, 128.7, 128.8, 128.9, 129.1, 135.5 (arom. C);
147.4, 149.1, 149.4, 152.3 (C=O, C=N). Anal. calc. for C19H20N2O2: C 64.0, H 5.66, N 7.86; found: C
64.1, H 5.45, N 7.79.
2-{[(Benzyloxy)imino]methyl}-1,3-thiazolidine-4-carboxylic Acid Methyl Ester (9b). To a stirred
soln. of L-cysteine methyl ester hydrochloride (3.42 g, 20 mmol) in EtOH (25 ml) at À108 was added a
soln. of Et3
A
precipitate of Et3N·HCl started to appear). To this was added a soln. of 8 (3.26g, 20 mmol) in EtOH
(5 ml). The mixture was stirred between À10 and 58 for 1 h and then at r.t. for 24 h. After this time,
the mixture was filtered and the solvent evaporated. The crude product was purified by CC (AcOEt/hex-
ane): pure 9b as a 1:1 mixture of isomers (4.98 g, 89%). Colorless oil. IR (NaCl): 1753 (C=O), 1635 (C=
N). 1H-NMR: 2.95 (dd, J=7, 10, 1/2 H, CH2(5)); 3.03 (dd, J=6, 10, 1/2 H, CH2(5)); 3.24 (dd, J=7, 10 1/2
H, CH2(5)); 3.28 (dd, J=6, 10, 1 H, CH2(5)); 3.78, 3.80 (s, 3 H, MeO); 3.92 (dd, J=6, 9, 1/2 H, HÀC(4));
4.10 (t, J=7, 1/2 H, HÀC(4)); 5.06, 5.10 (2s, 2 H, OCH2); 5.13 (d, J=6, 1/2 H, HC=N); 5.31 (d, J=7, 1/2
H, HC=N); 7.30–7.40 (m, 5 arom. H); 7.38 (1/2 H, NCHS, under arom.); 7.49 (d, J=6, 1/2 H, NCHS).
13C-NMR: 37.8, 38.7 (C(5)); 53.1 (MeO); 64.8, 65.2, 66.0, 66.4 (C(4), C(2)); 76.6, 76.9 (OCH2); 128.5,
128.7, 128.8, 128.9 (arom. C); 147.4, 149.1, 149.2, 152.3 (C=O, C=N). CI-MS: 281 (45, [M+H]+), 108
(46), 91 (100). HR-MS: 281.0960 (C
A
N2
G
N

2142
Helvetica Chimica Acta – Vol. 89 (2006)
2-{[(Benzyloxy)imino]methyl}-5,5-dimethyl-1,3-thiazolidine-4-carboxylic Acid Methyl Ester (9c). As
described for 9b, with the methyl ester hydrochloride of L-penicillamine (=3-mercapto-L-valine; 3.03 g,
18.6mmol), EtOH (70 ml), Et 3N (1.90 g) in EtOH (5 ml), and 8 (3.03 g, 18.6mmol) in EtOH (5 ml): pure
G
9c (1.43 g, 25%). Colorless oil. 1H-NMR: 1.21 (s, 3 H, Me); 1.63 (s, 3 H, Me); 3.72 (s, 1 H, NCH); 3.79 (s, 3
H, MeO); 5.10 (s, 2 H, OCH2); 5.21 (d, J=6, 1 H, HC=N); 7.33–7.37 (m, 5 arom. H); 7.46( d, J=6, 1 H,
NCHS); no further characterization was carried out because 9c was unstable.
Benzaldehyde O-(Chlorooxalyl)oxime (=Oxo{[(phenylmethylene)amino]oxy}acetyl Chloride) [35].
A soln. of benzaldehyde oxime (1.21 g, 10 mmol) in Et
(À408) soln. of oxalyl chloride (=ethanedioyl dichloride; 1.90 g, 15 mmol) in Et
for 1 h at À208, the solvent was evaporated at À108/13 Torr to leave a colorless temperature-sensitive
ACHTREUNG
AHCTREUNG
powder which was dried under vacuum for 2 h to remove residual Et2O: title oxime (2.01 g, 96%)
E
([35]: 97%). IR (NaCl): 1791 (C=O). 1H-NMR: 7.2–7.49 (m, 5 arom. H); 8.55 (s, CH). 13C-NMR:
128.2, 128.8, 129.2, 132.9 (arom. C); 159.4, 160.4 (C=N, C=O). MS: 211 (20, M+).
3-{2-[(Benzylideneamino)oxy]-2-oxoacetyl}-2-{[(Benzyloxy)imino]methyl}-5,5-dimethyl-1,3-thiazo-
lidine-4-carboxylic Acid Methyl Ester (10c). To a stirred soln. of benzaldehyde O-(chlorooxalyl)oxime
(1.0 g, 5 mmol) in CH2Cl2 (20 ml) at 08 was added a soln. of Et3N (400 mg, 5 mmol) in CH2Cl2 (5 ml) fol-
G
lowed by a soln. of 9c (1.43 g, 5 mmol) in CH2Cl2 (5 ml). The mixture was stirred at 08 for 10 min and then
at r.t. for 3 h. After this time, a small quantity of pentane was added to promote formation of the pyridine
hydrochloride salt. The mixture was filtered, the solvent evaporated, and the residue purified by FC: 10c
(1.97 g, 92%). Colorless oil. 1H-NMR: 1.47 (s, 3 H, Me); 1.58 (s, 3 H, Me); 3.78 (s, 3/2 H, MeO); 3.81 (s, 3/
2 H, MeO); 4.73 (s, 1/2 H, HÀC(4)); 4.82 (s, 1/2 H, HÀC(4)); 4.97 (s, 1 H, OCH2); 5.13 (s, 1 H, OCH2);
5.95 (d, J=7, 1/2 H, HC=N); 6.00 (d, J=7, 1/2 H, HC=N); 7.17–7.73 (m, 10 arom. H); 8.26( s, 1 H, HC=
NO); additional peaks were observed on standing at À208. Further characterization was not carried out
due to product degradation.
2-Oxo-N-{2-[(phenylmethoxy)imino]ethyl}-N-(phenylmethyl)-2-{[(phenylmethylene)amino]oxy}acet-
amide (15) was prepared as described previously [25].
EPR Study of 15. The ESR samples were prepared by taking 70 mg of the starting ꢀoxime oxalate
amideꢁ 15, 1 mol-equiv. of MAP, and dissolving them in (tert-butyl)benzene (3.5 ml). The soln. was stirred
until homogeneous, and this allowed the preparation of seven identical samples, each of 500 ml. The
degassed sample was placed in the EPR resonant cavity, and the cavity was cooled to 220 K. Once ther-
mal equilibrium was reached, the sample was photolyzed with light from a 500-W UV lamp, and the EPR
spectrum was recorded. The cavity temp. was then increased by 10 K to 230 K, and a fresh sample was
placed in the resonant cavity. This was also photolyzed and the spectrum recorded. The procedure was
repeated, increasing temp. in increments of 10 K until 280 K.
Preparative-Scale Photolysis of 15. A soln. of 15 (800 mg, 1.9 mmol) in toluene (400 ml) and in the
presence of MAP (840 mg, 5.7 mmol) was photolyzed by light from a 400-W UV lamp for 5 h at r.t.
1
After this time, the solvent was evaporated, and a sample was analyzed by H-NMR. The photolysis
was repeated at 808.
REFERENCES
[1] A. L. J. Beckwith, G. Moad, J. Chem. Soc., Perkin Trans. 2 1980, 1083.
[2] K. U. Ingold, B. Maillard, J. C. Walton, J. Chem. Soc., Perkin Trans. 2 1981, 970.
[3] J. R. Bews, C. Glidewell, J. C. Walton, J. Chem. Soc., Perkin Trans. 2 1982, 1447.
[4] B. Maillard, J. C. Walton, J. Chem. Soc., Perkin Trans. 2 1985, 443.
[5] J. C. Walton, J. Chem. Soc., Perkin Trans. 2 1989, 173.
[6] A. Srikrishna, in ꢀRadicals in Organic Synthesisꢁ, Eds. P. Renaud and M. Sibi, Wiley, Weinheim,
2001, Vol. 2, p. 151.
[7] J. C. Walton, Top. Curr. Chem. 2006, 264, 163.
[8] D. Johnston, C. M. McCusker, D. J. Procter, Tetrahedron Lett. 1999, 40, 4913.
[9] T. K. Hutton, K. Muir, D. J. Procter, Org. Lett. 2002, 4, 2345.
[10] A. Fernandez-Mateos, E. M. de la Nava, G. P. Coca, A. R. Silva, R. R. Gonzalez, Org. Lett. 1999, 1,
607.

Helvetica Chimica Acta – Vol. 89 (2006)
2143
[11] A. Gansauer, T. Lauterbach, S. Zarayan, Angew. Chem., Int. Ed. 2003, 42, 5556.
[12] M. Tojino, N. Otsuka, T. Fukuyama, H. Matsubaru, C. H. Schiesser, H. Kuriyama, H. Miyazato, S.
Minakata, I. Ryu, Org. Biomol. Chem. 2003, 4262.
[13] I. Ryu, H. Miyazata, H. Kuriyama, K. Matsu, M. Tojino, T. Fukuyama, S. Minakata, M. Komatsu, J.
Am. Chem. Soc. 2003, 125, 5632.
[14] G. Pattenden, S. J. Reynolds, Tetrahedron Lett. 1989, 30, 3229.
[15] G. Pattenden, G. B. Gill, S. J. Reynolds, J. Chem. Soc., Perkin Trans. 1 1994, 369.
[16] G. Binmore, J. C. Walton, L. Cardelini, J. Chem. Soc., Chem. Commun. 1995, 27.
[17] J. C. Walton, A. Studer, Acc. Chem. Res. 2005, 38, 794.
[18] L. V. Jackson, J. C. Walton, Chem. Commun. 2000, 2327.
[19] A. F. Bella, L. V. Jackson, J. C. Walton, J. Chem. Soc., Perkin Trans. 2 2002, 1839.
[20] A. F. Bella, L. V. Jackson, J. C. Walton, Org. Biomol. Chem. 2004, 2, 421.
[21] E. M. Scanlan, J. C. Walton, Chem. Commun. 2002, 2086.
[22] E. M. Scanlan, A. M. Z. Slawin, J. C. Walton, Org. Biomol. Chem. 2004, 2, 716.
[23] M. J. Tomaszewski, J. Warkentin, N. H. Wierstiuk, Aust. J. Chem. 1995, 48, 291.
[24] A. G. Fallis, I. M. Brinza, Tetrahedron 1997, 53, 17543.
[25] G. DiLabio, E. M. Scanlan, J. C. Walton, Org. Lett. 2005, 7, 155.
[26] B. Zimmermann, H. Lerche, T. Severin, Chem. Ber. 1986, 119, 2848.
[27] R. F. Hudson, K. A. F. Record, J. Chem. Soc., Chem. Commun. 1976, 539.
[28] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Mont-
gomery Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B.
Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene,
X. Li, J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann,
O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A.
Voth, P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain,
O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A.
G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi,
R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P.
M. W. Gill, B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, J. A. Pople, ꢀGaussian 03, Revision
A.1ꢁ, Gaussian, Inc., Pittsburgh PA, 2003.
[29] H. Fischer, H. Paul, Acc. Chem. Res. 1987, 20, 200.
[30] H. Schuh, H. Fischer, Helv. Chim. Acta 1978, 61, 2130.
[31] S.-U. Park, T. R. Varick, M. Newcomb, Tetrahedron Lett. 1990, 31, 2975.
[32] A. L. J. Beckwith, C. H. Schiesser, Tetrahedron Lett. 1985, 26, 373.
[33] S. Kim, G. H. Joe, J. Y. Do, J. Am. Chem. Soc. 1993, 115, 3328.
[34] J. S. Bryans, N. E. A. Chessum, N. Huther, A. F. Parsons, F. Ghelfi, Tetrahedron 2003, 59, 6221.
[35] J. C. Jochims, S. Hehl, S. Herzberger, Synthesis 1990, 1128.
Received February 20, 2006