Reaction of aromatic nitroso compounds with chemical models of 'thiamine active aldehyde'

Article abstract of DOI:10.1016/j.tet.2008.06.008

Aromatic nitroso compounds in the presence of base and 2-(α-hydroxyalkyl)-3,4-dimethylthiazolium trifluoromethanesulfonate and related salts furnish in variable yields O- and N-acyl-aryl hydroxylamines and 3,4-dimethylthiazolium trifluoromethanesulfonate. A primary kinetic isotope effect of 4.9, obtained for the appropriate 2α-deuterated thiazolium salt, points to the C_2α-H bond cleavage as the rate determining step. Radical species detected by ESR were unambiguously identified as phenylhydronitroxide, but attempted trapping of the corresponding C-heterocyclic radicals by TEMPO was not successful, and substrates incorporating a potential cyclopropyl radical clock gave products with the cyclopropyl ring intact. Theoretical calculations revealed a large activation energy for such reaction, which thus cannot per se exclude the intervention of such radical species. Evidence for the likely operation of two concurrent mechanisms, a radical and a preponderant ionic pathway, involving the conjugate base of the thiazolium salt, as the chemical model for 'active thiamine', and ArNO is presented for the formation of the products of the reaction.

Full text of DOI:10.1016/j.tet.2008.06.008

Tetrahedron 64 (2008) 7759–7770
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Reaction of aromatic nitroso compounds with chemical models
of ‘thiamine active aldehyde’
a
a
a
a
´
´
Luısa M. Ferreira , M. Manuel B. Marques , Paulo M.C. Gloria , Humberto T. Chaves ,
a
a
a
b
´
Joa˜o-Pedro P. Franco , Isabel Mourato , Jose-Rafael T. Antunes , Henry S. Rzepa ,
Ana M. Lobo a , Sundaresan Prabhakar
,
a,
*
*
a Chemistry Department, REQUIMTE/CQFB, Faculty of Sciences and Technology, New University of Lisbon and SINTOR-UNINOVA, 2829-516 Monte de Caparica, Portugal
b Chemistry Department, Imperial College London, South Kensington Campus, SW7 2AY London, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Aromatic nitroso compounds in the presence of base and 2-(a-hydroxyalkyl)-3,4-dimethylthiazolium
Received 13 May 2008
Accepted 4 June 2008
Available online 7 June 2008
trifluoromethanesulfonate and related salts furnish in variable yields O- and N-acyl-aryl hydroxylamines
and 3,4-dimethylthiazolium trifluoromethanesulfonate. A primary kinetic isotope effect of 4.9, obtained
for the appropriate 2a-deuterated thiazolium salt, points to the C2a–H bond cleavage as the rate de-
termining step. Radical species detected by ESR were unambiguously identified as phenylhydronitroxide,
but attempted trapping of the corresponding C–heterocyclic radicals by TEMPO was not successful, and
substrates incorporating a potential cyclopropyl radical clock gave products with the cyclopropyl ring
intact. Theoretical calculations revealed a large activation energy for such reaction, which thus cannot
per se exclude the intervention of such radical species. Evidence for the likely operation of two con-
current mechanisms, a radical and a preponderant ionic pathway, involving the conjugate base of the
thiazolium salt, as the chemical model for ‘active thiamine’, and ArNO is presented for the formation of
the products of the reaction.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
The ability of thiamine-dependent enzymes to convert aromatic
nitroso compounds into hydroxamic acids has been investigated by
Thiamine diphosphate 1 is a coenzyme of importance in mam-
malian carbohydrate metabolism as it is involved in vital non-oxi-
dative and oxidative processes.1 The simplest reaction catalysed by
this coenzyme in aqueous systems is pyruvate decarboxylation2 in
yeast (Scheme 1). According to this pathway there are three co-
valent intermediates between the coenzyme/conjugate base and
Corbett et al.,5a,b and more recently Ying et al.5c also used nitroso-
benzene as the electrophile in the reaction with the homoenolate
equivalent to the enamine 3. Earlier work from our laboratory5d also
showed (Scheme 2) that the thiazolium salts 6a, modelled on the
thiamine diphosphate coenzyme, reacted with nitrosobenzene (7a)
in the presence of Et3N to give a mixture of a model procarcinogen
O-acetylphenylhydroxylamine6 (8a), phenylacetohydroxamic acid
(8b), thiazolium salt 9 and azoxybenzene 10.
the substrate/product: the direct adduct 2, the 2-
equilibrium with the corresponding enamine (‘active aldehyde’)
and the 2-( -hydroxy)-ethyl derivatives 4. Final decomposition of 4
a carbanion 3 in
a
yields the ylide derivative of 1 and acetaldehyde. Alternatively, as in
the enzyme pyruvate oxidase found in lacto bacteria,3 4 may suffer
oxidation leading to the reactive 2-acetylthiazolium species 5,
which reacts with inorganic phosphate producing acetyl phosphate
and regenerating the ylide. Much work has been done in studying
properties of species such as 4 and its structurally simplified
analogues.4
2. Results and discussion
Various possibilities that may be considered for their formation
are summarised as follows (Scheme 3): that PhNHOAc and
PhN(OH)Ac could originate by a hydride ion transfer from 6a0 to
PhNO to produce initially the phenylhydroxylamine anion 11a/11b
and 2-acetylthiazolium salt 12. Alternatively the latter two sub-
stances could be produced from 13b and PhNO by a process in-
volving electron transfer and protonation. Subsequent reaction
between the ambident nucleophile 11a/11b and ketone 12 would
furnish products 8a and 8b resulting from oxygen and nitrogen
attack, respectively. Either an isomerisation of 8a or a condensation
* Corresponding authors. Fax: þ351 212948550.
E-mail addresses: aml@fct.unl.pt (A.M. Lobo), sprabhakar@fct.unl.pt (S.
Prabhakar).
URL: http://www.dq.fct.unl.pt/qoa/lpg.html
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.06.008

7760
L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
Scheme 1.
Scheme 2.
of enamine 13b and PhNO was mentioned as a possible route to 8b.
Azoxybenzene (10) was thought to arise from PhNHOAc and PhNO.
Given the biological importance of such reaction, and the
presence of nitroso aromatics as xenobiotics in the environment,
the study of the reaction was extended to other nitroso aromatics
and related thiazolium salts (Table 1). These substances also gave
largely similar results except that in certain cases anilines (14) and/
or azobenzenes (15) were obtained (entries 3, 5, 6 and 8; Table 1).
Scheme 3.
DABCO, the first two were found to be more satisfactory than the
last. NaH was also used as a base for the reaction.
2.1. Solvent/base
Although the reaction was found to occur in solvents of different
polarities (dimethoxyethane, dioxane, dichloromethane, chloro-
form, dichloroethane, acetonitrile) the eventual choice of CH2Cl2
was dictated by solubility considerations and acceptable rates of
the reactions. Of the three organic bases Et3N, Hunig’s base and
2.2. Preparation of starting materials: thiazolium carbinols 6
All thiazolyl ketones, except 18d, were secured from the ap-
propriate ester 16 and 2-lithium-4-methylthiazole 17 (1: 1.1 molar
ratio), as described earlier7a (Scheme 4).

L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
7761
Table 1
Reaction of aromatic nitroso with thiazolium salts and base
Entry
R–C6H4N]O
Thiazolium salta
O-Acyl hydroxylaminesb (%)
ArNHOAc 8a
Hydroxamic acids (%)
ArN(O)]NAr (%)
ArN]NAr (%)
ArNH2 (%)
ArN(OH)Ac 8b
1
2
3
4
5
6
7a, R¼H
6a
6a
6a
6a
6a
6a
45
14
13c
5
11
d
14
21
d
15
19
d
22
59
48
30
43
d
d
d
23
d
d
d
d
d
d
d
6
7b, R¼4-Br
7c, R¼4-NO2
7d, R¼4-Cl
7e, R¼3-Me
7f, 2,4,6-tri-tert-butyl
nitrosobenzene
28
ArNHOCOPh 8c
ArN(OH)COPh 8d
7
8
9
10
7b
7c
7d
7e
6b
6b
6b
6b
21
46d
24
25
d
17
23
46
29
31
43
d
12
d
d
d
d
d
d
28
11
7a
6c
50
ca. 1
19
d
d
12
7a
6d
23
2
d
d
d
a
Counter ion for all salts, CF3SOꢀ3 .
b
O-Acyl hydroxylamines are, in general, labile substances and hence they were characterised and quantified, unless stated otherwise, as stable N-acyloxy-4-nitro-
benzamides (see Section 4).
c
Characterised as the corresponding N-benzamide.
Characterised as the corresponding N-acetamide.
d
The above procedure, however, did not give acceptable results
for ethyl 20,20-diphenylcyclopropyl-carboxylate 16d.7b The requisite
ketone 18d was found to be admixed with the bis-addition product
19f and their separation proved to be difficult. Therefore it was
obtained in three steps involving: (i) 2 equiv of the lithium salt per
mole of 16d, (ii) N-monoalkylation of the resulting 19f to 19g and
(iii) DABCO induced fragmentation of 19g to 18d in 50% overall yield
(Scheme 5).
Reduction of ketones 18 with NaBH4 furnished the corre-
sponding secondary alcohols 19. These were alkylated with
CF3SO3Me to yield respective the N-quaternised carbinol
6
(Scheme 6).
Scheme 4.
Scheme 5.

7762
L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
carbanion 13a in resonance with enamine 13b, and any one of them
could be implicated in the formation of 8a and 8b (cf. Scheme 3).
The pKa of the C2a–H in thiamine is estimated to be between 17 and
19,8,9 a value close to that of a typical secondary alcohol.
Based on the chemistry reported for the active aldehyde with
electrophiles, both ionic1a and radical pathways can be envisioned
for the reaction 6a/8aþ8b (Scheme 2). We first consider in some
detail the latter process since PhNO is a good electron acceptor
and thiazolium salts in basic medium have been shown to
undergo easy oxidation to form relatively stable heterocyclic
radicals.10
Scheme 6.
2.4. Detection and identification of phenylhydronitroxide
radical (PhNH–Oꢁ)
2.3. Carbanion chemistry of some derivatives of
2
a
-hydroxyethyl thiazolium salts
1. When the reaction of 6a and 7a was studied by ESR it was
found that in the absence of base a mixture of 6a and 7a, in
equimolar proportions in CH2Cl2, was indefinitely stable and
did not generate any radical species detectable by ESR
spectroscopy.
2. Addition of an equivalent of Et3N to the above solution resulted
in the appearance of strong ESR signals (bandwidth 44 G), as
shown in the spectrum of Figure 1a.
When the OH group of 6a was protected either as an acetyl 20a
or as a silyl derivative 20b, their reactions with ArNO, in the pres-
ence of a base, afforded the nitrones 21a,b (pathway a) and the
hydroxamic acid 8b (Ar¼Ph) (pathway b), respectively. These re-
actions undoubtedly proceed via the derived carbanion or enamine
as shown below (Scheme 7).
However, in the absence of such protection, the deprotonation
of 6a could generate the corresponding alkoxide 6a0 and/or the
3. An identical spectrum was recorded with N,N-diisopropyl-
amine, or when the solvent was changed to CD2Cl2, indicating
Scheme 7.
Figure 1. (a) Observed ESR spectrum for the reaction involving equimolar quantities of 3,4-dimethyl-2-(a a-hydroxyethyl)-thiazolium triflate, Et3N and PhNO; (b) simulated
-[2H]-
spectrum of PhND–Oꢁ.

L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
7763
that radicals were not derived from either the solvent or the
base. The signals generated were attributed exclusively to
PhNH–Oꢁ11 (Fig. 1a).
entries 11 and 12) with PhNO were examined under standard
conditions. In the event19 none of butyric acid derivatives 26d and
26f could be detected when run against authentic samples on TLC
(Scheme 10). Only products 27c and 27e were formed and these
characterised as their p-nitrobenzoyl derivatives 27d and 27f,
respectively.
Substrates incorporating a potential cyclopropyl radical probe
gave products with the cyclopropyl ring intact, but theoretical
calculations reveal this to be due to the endothermicity of the ring
opening. This probe therefore cannot per se be used as a test to
exclude the intervention of such radical species.
Calculations using Gaussian 03 (revision d.02) were done on
the diphenylcyclopropyl substituted radical 60460046%. At the
optimised geometry, the calculated spin densities suggest that al-
though representations 60 and 6% are both more realistic than 600,
the unpaired electron is nevertheless also delocalised around the
thiazolium ring. This would account for the reluctance of the radical
to ring open the cyclopropyl ring (see Fig. 2), a process calculated as
being endothermic in free energy by 15.9 kcal molꢀ1 and hence
very unfavoured.
4. Compound 6e, the 2a-deuteriated analogous to 6a, obtained by
NaBD4 reduction of 12 followed by N-methylation of the
resulting alcohol, on reaction with 7a, originated spectrum
Fig. 1a and not that of PhND–Oꢁ (Fig. 1b) showing that a C2a–H
homolytic bond cleavage is not involved in the production of
the above radical.
5. The spectrum generated from 2-chloronitrosobenzene (which
also furnished the corresponding N-acetoxyaniline) was found
to be identical with that of 2-chlorophenylhydronitroxide
recorded in the literature.12
6. The simplified spectrum resulting from pentadeuteriated
nitrosobenzene13 and its simulated spectrum were consistent
with the structure of pentadeuteriated hydronitroxide.
7. ESR signals due to phenylhydronitroxide radical have been
often observed during the electrochemical reduction of
nitrosobenzene even in supposedly ‘dry solvents’ and this was
attributed to the protonation of the initially formed anilinoxyl
radical by the adventitious presence of water molecules.14 To
eliminate this possibility the reaction was conducted in dry
CH2Cl2 with NaH15 (which assured anhydrous conditions of
the medium) as the base. Since the same radical was observed
by ESR it was concluded that phenylhydronitroxide is not
formed from anilinoxyl and subsequent protonation
(Scheme 8).
On these grounds ketyl 24 cannot be definitely excluded as an
intermediate in the reaction leading to PhNH–Oꢁ.
In an attempt to trap the heterocyclic radical, derived from the
thiazolium moiety, the reaction was performed in the presence of
an equivalent of TEMPO relative to nitrosobenzene. A solid residue
obtained on removal of the solvent in vacuum was leached several
times with dry ether to remove unreacted TEMPO and other ether-
soluble substances such as azoxybenzene. The resulting polar res-
idue was purified by PTLC to furnish a low yield of nitrone 21a and
acetate 20a, both identical with their respective authentic samples.
In the absence of PhNO only 20a was isolated (33%). A reasonable
mechanism (Scheme 11) for the reaction would consist of an oxi-
dation of 6a to 12 followed by acetylation of the former to 20a. The
formation of 21a from 20a and PhNO is already referred to in
Scheme 6.
With TEMPO in excess (5 equiv), relative to nitrosobenzene,
practically no azoxybenzene, PhN(OH)Ac or PhNHOAc were
formed. Virtually all PhNO remained unreacted.20 Clearly, in re-
actions involving the thiazolium salt 6a in basic medium, TEMPO
is acting as the oxidant.21 If this were to be so, then TEMPOL or its
anion should be formed. The failure to isolate it under normal
conditions could be due to its facile air reoxidation. Therefore, the
experiment was repeated in degassed CH2Cl2 under dry argon
(Scheme 12). When a significant quantity of TEMPOL was detec-
ted on the TLC,22 acetic anhydride was added to the reaction
mixture and N-acetoxytetramethylpiperidine (28), mp 63 ꢂC (lit.23
63.5–65 ꢂC) was formed and isolated by PTLC in 43% yield. Its IR
and 1H NMR spectra were identical with an authentic sample
obtained from TEMPOL,24 Ac2O and Et3N.
Scheme 8.
These observations were consistent with a hydrogen atom
abstraction from the OH group of enamine 13b by PhNO. The
complementary heterocyclic radical (not discernable by virtue of
the strong signals of PhNH–O16), engendered simultaneously, could
have one of the following canonical structures, namely the enoxyl
22, the a-keto radical 23 or the ketyl 24 (Scheme 9).
Ring opening of a ketyl of type 25 (cf. Scheme 9) is known.17
Therefore, the reaction of the cyclopropyl salts18 6c and 6d (Table 1,
Scheme 9.

7764
L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
Scheme 10.
These results could be best interpreted as involving a hydrogen
atom abstraction25 by TEMPO from enamine to furnish TEMPOL
and the enoxyl 22. Further reaction of TEMPO, which is present in
excess, with the
a
-keto radical26 23 would give the tetrahedral
intermediate 29 from which 12 could be formed along with TEMPO
anion (Scheme 13). We ascribe the failure of TEMPO to undergo
acetylation by 12 to unacceptable steric congestion in the tetra-
hedral intermediate produced in an ionic reaction. In fact no acetyl-
ation was observed between these two substances (TEMPOLþ12)
in the presence of base. Therefore acetylation of alcohol 6a occurs
by default to give 20a.
During the mechanistic studies, it was assumed that azoxy-
benzene formation could be due, inter alia, to a condensation be-
tween PhNHOAc and PhNO. Indeed a mixture of these two
substances in the presence of Et3N rapidly produced a high yield of
azoxybenzene. However, azoxybenzene would also be expected for
a reaction between PhNO and PhNHOH if the latter were to be
formed by a hydride transfer from enamine 13b to PhNO (Scheme
14). In fact, on mixing PhNHOH, PhNO and Et3N in CH2Cl2, a good
yield of azoxybenzene was obtained. On monitoring this reaction
by ESR spectroscopy signals due to PhNH–Oꢁ were also observed.27
Its concentration, as determined with reference to an external
sample of TEMPO of known concentration, was 3.3ꢃ10ꢀ4 M. The
maximum concentration of phenylhydronitroxide, achieved after
20 min, for a 0.1 M solution each of PhNO and thiazolium salt 6a
containing Et3N (1 equiv) in CH2Cl2 was determined to be
Figure 2. Spin density (ROB3LYP/cc-pVDZ) calculation on radical 60460046%. Num-
bers in brackets refer to spin densities.

L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
7765
Scheme 11.
Scheme 12.
Scheme 13.
9ꢃ10ꢀ5 M. The nitroso group could accept an H atom from enol 13b
to produce directly the radical species (Scheme 14, via a). Alter-
natively in a preponderant ionic reaction, a hydride ion transfer
could occur from 13b to produce PhNHOꢀ. The latter, in its sub-
sequent reaction with PhNO leading to azoxybenzene, could also
originate phenylhydronitroxide (Scheme 14, via b) or react directly
with the 2-acetylthiazolium (12) to give PhNHOAc (8a).28 Whatever
the genesis of the radicals, their eventual conversion into PhNHOH
and thence to PhNHOAc (Scheme 14), could follow the same
chemistry as outlined for the formation of TEMPOL (Scheme 13).
Finally, the mechanism referred to above should display a pri-
mary kinetic isotopic effect if the deprotonation, i.e., enamine 13b
formation, were to be the rate determining step (Scheme 14).
Therefore the reactions of PhNO with 6a and the deuterium la-
non-deuterated 6a and the deuterated 6e substrates. The value of
kHo /kDo of 4.9, obtained for the primary kinetic isotope effect, is in
close agreement with that found for the deprotonation of similar
substances.29
3. Conclusions
While the involvement of an ionic mechanism for the formation
of a ‘thiamine active aldehyde’ analogue in the rate determining
step and its subsequent attack to the electrophilic nitroso group
appears the most likely pathway for the base catalysed reaction
under study; two different hydrogen abstracting reactions are
proposed to account for the generation of a small concentration of
PhNH–Oꢁ radicals. While substrates incorporating a potential
cyclopropyl radical clock gave products with the cyclopropyl ring
intact, a result supported by the calculated high endotermicity of
the reaction, an attempted trapping of the corresponding C–
heterocyclic radicals by TEMPO was not successful. Thus evidence
for the operation of a preponderant ionic pathway, involving the
belled 3,4-dimethyl-2-(a a-hydroxyethyl)-thiazolium triflate
-[2H]-
(6e), under pseudo-first order conditions (excess base and alcohol),
were conducted. The rate constants as determined by the disap-
pearance of nitrosobenzene (UV) were found to be koH
(8.8ꢄ0.4)ꢃ10ꢀ3 sꢀ1 and koD (1.8ꢄ0.1)ꢃ10ꢀ3 sꢀ1, respectively, for the

7766
L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
Scheme 14.
conjugate base of the thiazolium salt and ArNO, while favoured
for the formation of the isomeric O- and N-acylated aromatic
hydroxylamines, cannot rule out the intervention of radical
intermediates.
up implies that organic extracts were washed with water and dried
over anhydrous sodium sulfate or magnesium sulfate, filtered and
solvent removed from the filtrate under reduced pressure. Ana-
lytical thin-layer chromatography and preparative TLC (PTLC) were
performed on E. Merck Kieselgel 60, F254 silica gel (0.2 mm thick)
and 0.5, 1 or 2 mm thick plates (20ꢃ20 cm), respectively. Column
chromatography was performed on E. Merck Kieselgel 60 (240–
4. Experimental
4.1. General
400 mm) silica gel.
Melting points were recorded on a Reichert-Thermovar hot
stage apparatus and are uncorrected. Ordinary mass spectra were
recorded on a Fisons Trio or AEI MS-9 spectrometers. High reso-
lution mass spectra were recorded on an AutoSpeQ spectrometer.
1H and 13C NMR spectra recorded in CDCl3 on a Bruker ARX 400
spectrometer (400 MHz for 1H and 100.63 MHz for 13C). Chemical
shifts reported are relative to tetramethylsilane as the internal
4.2. Starting materials
4.2.1. General procedure for the preparation of nitrosoarenes
These substances were prepared from the corresponding nitro
compounds by reduction first with zinc dust in boiling aq ethanol
followed by oxidation of the resulting arylhydroxylamines with aq
FeCl3 as described.31 Nitrobenzene-d5 was processed as above
32
reference (dH 0.00) for 1H NMR spectra and to CDCl3
13C NMR spectra. Chemical shifts are expressed in parts per million,
downfield from TMS (
¼0) or residual dichloromethane (dH¼5.32,
(
dC 77.00) for
to give nitrosobenzene-d5 (25%), mp 65–67 ꢂC (ethanol) (lit.33 65–
67 ꢂC).
d
dC¼53.1) as internal standards. IR spectra were run on Perkin–
Elmer 683 and Spectrum 1000 instrument with absorption fre-
quencies expressed in reciprocal centimetres. Electronic spin
resonance spectra (ESR) were recorded at rt on a Bruker EMX
spectrometer. Typical conditions were as follows: microwave
power, 2 mW; modulation amplitude, 0.1 G at 100 KHz; time con-
stant, 20.48 ms; sweep time 83.886 ms. The ESR parameters for
the various radical species were determined by signal simulations
using the software ‘SimFonia’ from Bruker. Spin concentration was
calculated with reference to pure TEMPO (sublimed) of known
concentration as external standard under the following condition:
microwave power, 2 mW; modulation amplitude, 0.4 G at 100 KHz;
time constant, 20.48 ms; sweep time 83.886 ms; sweep width
200 G and receiver gain at 2ꢃ103. Microanalyses were performed
on a Carlo Erba 1106R microanalyser. All reagents and solvents were
purified and dried by standard methods30 before use. Usual work-
4.2.2. General procedure for the preparation of 2-acyl-4-
methylthiazoles
The 2-acyl-4-methylthiazoles mentioned in the present work
were prepared as described7a earlier except for 18d, which was
obtained as follows: 16d/19f/19g/18d.
4.2.2.1. (20,20-Diphenylcyclopropyl)-bis-(4-methyl-2-thiazoyl)-carbinol
(19f). Ethyl-20,20-diphenylcyclopropane carboxylate7b (16d) (2.8 g,
10.53 mmol) was treated with 2-lithio-4-methylthiazole [from
4-methylthiazole, 1.93 mL, 21.06 mmol and n-BuLi (1.6 M in
n-hexane, 2 equiv)] to give 19f as a pale yellow solid (73%), mp 95–
96 ꢂC (CH2Cl2–n-hexane); nmax(KBr)/cmꢀ1 3426; dH (CDCl3) 1.23
(1H, dd, J 4.7 and 9.1 Hz), 2.11 (1H, t, J 5.4 Hz), 2.39 (3H, s), 2.45 (3H,
s), 3.10 (1H, dd, J 6.3 and 9.1 Hz), 3.77 (1H, exchangeable with D2O),
6.75 (1H, s), 6.91 (1H, s), 6.98 (2H, d, J 4.7 Hz), 7.08–7.13 (4H, m),
7.20–7.25 (2H, m), 7.45 (2H, d, J 7.4 Hz); dC 14.2, 17.1, 17.3, 35.5, 36.3,

L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
7767
114.2, 115.1, 126.1, 126.3, 127.9, 128.2, 128.7, 130.0, 140.2, 146.8, 151.3,
152.7, 173.7; HRMS(EI): m/z C24H22N2OS2 requires: 418.11735,
found: 418.11731.
4.2.4.1. 2-(a-Hydroxyethyl)-3,4-dimethylthiazolium triflate (6a).
Compound 19a yielded 6a (91.5%), mp 83–85 ꢂC (Et2O–butanone);
nmax(KBr)/cmꢀ1 3350; dH (DMSO-d6) 1. 60 (3H, d, J 7.3 Hz), 2.46 (3H,
s), 3.83 (3H, s), 4.90 (1H, m), 5.34 (1H, m), 7.57 (1H, s); dc 13.7, 22.2,
37.0, 64.7, 117.9, 122.5, 147.2. Anal. Calcd for C8H12F3NO4S2: C, 31.27;
H, 3.94; N, 4.56; S, 20.87%. Found: C, 31.03; H, 3.82; N, 4.52; S, 20.99.
4.2.2.2. Compound 19g. Compound 19f (500 mg, 1.2 mmol) in dry
ether (10 mL) was treated with CF3SO3Me (135.35 mL, 1.2 mmol)
and the resulting mixture stirred overnight at rt. The solid that had
separated was collected by filtration and crystallised (ethanol–
Et2O) to give thiazolium salt 19g (626 mg) as a colourless solid
(90%), mp 155–156 ꢂC (CH2Cl2–n-hexane); nmax(KBr)/cmꢀ1 3426; dH
(DMSO-d6) 1.47 (1H, dd, J 5.0 and 8.6 Hz), 2.30 (3H, s), 2.60 (3H, s),
3.08–3.11 (2H, m), 4.12 (3H, s), 7.04 (5H, s), 7.16–7.36 (5H, m), 7.88
(1H, s), 8.41 (1H, s), 10.04 (1H, s, exchangeable with D2O); HRMS
(FAB) m/z C25H25N2OS2 requires: 433.14083, found: 433.13968.
4.2.4.2. 2-(a-Hydroxy-a-deuterioethyl)-3,4-dimethylthiazolium tri-
flate (6e). Compound 19e yielded 6e (85%), mp 84–86 ꢂC (buta-
none–Et2O); nmax (KBr)/cmꢀ1 3300; dH (CD2Cl2) 7.44 (1H, s), 5.70
(1H, s), 3.93 (3H, s), 2.51 (3H, s), 2.51 (3H, s), 1.61 (3H, s); HRMS
(FAB) m/z C7H11DNOS requires: 159.0702, found: 159.0697.
4.2.4.3. 2-(a-Hydroxybenzyl)-3,4-dimethylthiazolium triflate (6b).
Compound 19b yielded 6b (92.4%), mp 88–89 ꢂC (Et2O–butanone);
nmax(film)/cmꢀ1 3300; dH (CDCl3) 7.52 (1H, s), 6.38 (2H, m), 3.86
(3H, s), 2.50 (3H, s); dc 13.7, 37.3, 70.6,118.2,118.4,127.4,129.1,129.6,
136.8, 147.5. Anal. Calcd for C13H14F3NO4S2: C, 42.27; H, 3.82; N,
3.79. Found: C, 42.58; H, 3.80; N, 3.63.
4.2.2.3. 2-(20,20-Diphenylcyclopropanoyl)-4-methylthiazole (18d).
Compound 19g (500 mg, 1.2 mmol) dissolved in a minimum
quantity of ethanol was treated with DABCO (134 mg, 1.2 mmol).
Usual work-up (after 10 min) followed by crystallisation of the
resulting residue provided 18d (294 mg, 77%) as a colourless solid,
mp 122–124 ꢂC (CH2Cl2–n-hexane); nmax(KBr)/cmꢀ1 1672; dH
(CDCl3) 1.78 (1H, dd, J 4.3 and 7.8 Hz), 2.51 (1H, d, J 5.2 Hz), 2.60 (3H,
s), 4.12 (1H, t, J 7.0 Hz), 7.15–7.32 (8H, m), 7.45 (2H, d, J 7.5 Hz); dc
17.3, 21.6, 32.5, 44.9, 121.0, 126.6, 128.0, 128.5, 129.8, 130.1, 139.4,
145.1, 155.3, 189.0; HRMS m/z C20H17NOS requires: 319.10308,
found: 319.10270.
4.2.4.4. 2-(a-Hydroxycyclopropylmethyl)-3,4-dimethylthiazolium tri-
flate (6c). Compound 19c yielded 6c, as a pale yellow oil (95%),
nmax(film)/cmꢀ1 3300; dH (CDCl3) 0.73 (4H, m), 1.32 (1H, m), 2.57
(3H, s), 4.07 (3H, s), 4.94 (1H, t, J 6.3 Hz), 5.57 (1H, d, J 6.30 Hz), 7.52
(1H, s); dC (CDCl3þDMSO-d6) 2.7, 3.3, 13.9, 69.1, 118.1, 122.4, 147.2;
HRMS (FAB) m/z C9H14NOS requires: 184.07960, found: 184.0780.
4.2.3. General procedure for the preparation of 2-(
a-hydroxyethyl)-
4.2.4.5. 2-(a
-Hydroxy-20,20-diphenylcyclopropylmethyl)-3,4-dimethyl-
4-methylthiazole and 2-( -hydroxybenzyl)-4-methylthiazole from
the corresponding keto compounds
a
thiazolium triflate (6d). Compound 19d yielded 6e (92%), mp 150–
151 ꢂC (Et2O–butanone); nmax(KBr)/cmꢀ1 3350; dH (DMSO-d6) 1.47
(1H, dd, J 5.3 and 8.6 Hz), 1.63 (1H, t, J 5.5 Hz), 2.20 (1H, dd, J 9.1 and
15.4 Hz), 2.50 (3H, s), 3.85 (3H, s), 4.19 (1H, t, J 8.4 Hz), 6.89 (1H, d, J
6.9 Hz, exchangeable with D2O), 7.19 (1H, t, J 7.0 Hz), 7.26–7.38 (7H,
m), 7.55 (2H, d, J 7.4 Hz), 7.93 (1H, s), 8.97 (1H, exchangeable with
D2O); dc 30.4, 31.0, 50.3, 126.3, 127.8, 128.5, 129.2, 137.6, 137.9, 143.9,
167.6; HRMS (FAB) m/z C21H22NOS requires: 336.14221, found:
336.14360.
The appropriate ketone (1 mmol) in ethanol (5 mL) was reduced
with a suspension of NaBH4 (12.5 mg; 0.33 mmol) in ethanol
(5 mL). On completion of the reaction (TLC control, SiO2, MeOH–
CH2Cl2; 2%), solvent was removed under reduced pressure, the
residue mixed with water and extracted with Et2O (3ꢃ5 mL). The
product worked-up in the usual manner was purified by column
chromatography.
4.2.3.1. 2-(
a
-Hydroxyethyl)-4-methylthiazole (19a). 2-Acetyl-4-
4.2.4.6. 2-(a-Acetoxyethyl)-3,4-dimethylthiazolium triflate (20a). 2-
methylthiazole7a (18a) gave 19a2b as a colourless oil (85%);
nmax(film)/cmꢀ1 3300; dH (CD3CN) 1.48 (3H, d, J 6 Hz), 2.37 (3H,
s), 3.70 (1H, br), 5.05 (1H, q, J 6 Hz), 6.98 (1H, s).
(a
-Acetoxyethyl)-4-methylthiazole7a gave 20a (92%), mp 103–
104 ꢂC (butanone–Et2O); nmax(KBr)/cmꢀ1 1740; dH (DMSO-d6) 1.65
(1H, d, J 6.6 Hz), 2.14 (3H, s), 2.51 (3H, s), 3.98 (3H, s), 6.41 (1H, q, J
6.6 Hz), 7.01 (1H, s). Anal. Calcd for C10H14F3NO5S2: C, 34.38; H,
4.04; N, 4.01%. Found: C, 34.40; H, 4.06; N, 3.97.
4.2.3.2. 2-(a-Hydroxy-a-deuterioethyl)-4-methylthiazole (19e). 2-
Acetyl-4-methylthiazole (18a) gave 19e by reduction with NaBD4 in
68%, as an oil; nmax(film)/cmꢀ1 3300; dH (CD2Cl2) 1.54 (3H, s), 2.36
(3H, s), 3.35 (1H, s), 6.82 (1H, s); m/z (FAB) 144 (MHþ).
4.2.4.7. 2-[
flate (20b). 2-[
thiazole (20a) was first prepared by stirring at rt a mixture of
2-( -hydroxyethyl)-4-methylthiazole (3 mmol) with TMSCl
(1 equiv) and hexamethyldisilasane (1 equiv) in pyridine. It was
then filtered, the filtrate concentrated under reduced pressure to
yield the silyl derivative as a colourless liquid (88%); dH (DMSO-d6)
0.12 (9H, s), 1.44 (3H, d, J 6 Hz), 2.31 (3H, s), 5.07 (1H, q, J 6.0 Hz),
7.10 (1H, s). This compound, without further purification, was
quaternised as described above for alcohol 19 to give 20b (87%), mp
98–100 ꢂC (butanone–Et2O); dH (CD3CN) 0.20 (9H, s), 2.40 (3H, s),
3.80 (3H, s), 5.50 (1H, q, J 6.0 Hz), 7.70 (1H, s); HRMS (FAB) m/z
a
-(Trimethylsilyloxy)-ethyl]-3,4-dimethylthiazolium tri-
a
-(Trimethylsilyloxy)-ethyl]-3,4-dimethylmethyl-
4.2.3.3. 2-(
a
-Hydroxybenzyl)-4-methylthiazole (19b). 2-Benzoyl-4-
a
methylthiazole7a (18b) gave 19b in 44% yield, mp 94–95 ꢂC (buta-
none–Et2O) (lit.34 96 ꢂC); dH (CDCl3) 2.40 (3H, s), 4.90 (1H, br), 6.05
(1H, s), 6.85 (1H, s), 7.45 (5H, m).
4.2.3.4. 4-Methyl-2-thiazolyl cyclopropylcarbinol (19c). 2-Cyclo-
propanoyl-4-methylthiazole7a (18c) gave 19c in 77% yield, as an oil,
nmax(film)/cmꢀ1 3250; dH (CDCl3) 0.71–0.50 (4H, m), 1.30 (1H, m),
2.42 (3H, s), 3.77 (1H, d, J 2.4 Hz), 4.32 (1H, dd, J 2.4 and 8.1 Hz), 6.83
(1H, s). The alcohol was used as such for the subsequent quater-
nisation reaction.
C10H20NOSSi requires: 230.1035, found: 230.1058.
4.3. General procedure for reactions between 2-(
hydroxyalkyl or -hydroxybenzyl)-3,4-dimethylthiazolium
triflate (6) and arylnitroso compounds (7)
a-
4.2.4. General procedure for 2-substituted thiazolium triflates
The appropriate thiazole-alcohol 19 (1.26 mmol) in dry ethyl
ether (10 mL) was treated with methyl trifluoromethylsulfonate
a
(135.5
m
L, 1.26 mmol) and the mixture allowed to stand at rt (24 h).
The thiazolium salt (0.5 mmol) and the arylnitroso compound
(0.5 mmol) in dry CH2Cl2 (10 mL/mmol) at rt were treated with
the appropriate base (1.1 equiv). On completion of the reaction
The solid that separated was collected and crystallised to give the
corresponding triflates.

7768
L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
(0.5–2 h, TLC control; CH2Cl2 or CH2Cl2–MeOH 98:2) the products
were isolated by diluting the mixture with ethyl ether and rapid
washing with an ice-cold aq NaOH (0.5 N) to remove hydroxamic
acid. Usual work-up of the ethyl ether solution led to a residue,
which was thoroughly dried in vacuo. It was redissolved in dry THF,
cooled to ꢀ30 ꢂC before adding NaH (3 equiv). The derived Na salt
thus obtained was treated with the appropriate acid chloride and
the resulting mixture allowed to reach rt while being stirred. The
derived N-acyloxy-N-acylaniline, and other products such as
azoxybenzene, aniline (as N-acyl derivative) and azobenzene were
isolated by PTLC. Acidification of the initial alkaline extract with aq
HCl (1 N) furnished hydroxamic acids.
4,40-dibromoazoxybenzene (46%), mp 170–171 ꢂC (ethanol) (lit.37
172 ꢂC).
4.3.8. Reaction between 7c and 6b (Table 1, entry 8)
Furnished O-benzoyl-N-(4-nitrophenyl)hydroxylamine (46%),
mp 118–120 ꢂC (CH2Cl2–n-hexane); characterised as the N-acetyl
derivative, mp 105–106 ꢂC (CH2Cl2–n-hexane); nmax(KBr)/cmꢀ1
1760, 1695; dH (CDCl3) 2.29 (3H, s), 7.50–7.65 (3H, m), 7.76 (2H, d, J
9 Hz), 8.13–8.29 (4H, m). Anal. Calcd for C15H12N205: C, 60.00; H,
4.03; N, 9.33. Found: C, 59.93; H, 4.40; N, 9.07; 4,40-dinitro-
azoxybenzene (29%), mp 188–190 ꢂC (ethanol) (lit.37 192 ꢂC) and
4,40-dinitroazobenzene (12%), mp 223–225 ꢂC (lit.39 222–223 ꢂC).
4.3.1. Reaction between 7a and 6a (Table 1, entry 1)
4.3.9. Reaction between 7d and 6b (Table 1, entry 9)
Furnished
O-acetyl-N-(40-nitrobenzoyl)-N-phenylhydroxyl-
Furnished O-benzoyl-N-(4-chlorophenyl)-N-(40-nitrobenzoyl)-
hydroxamic acid (24%), mp 120–121 ꢂC (CH2Cl2–n-hexane);
amine (45%), mp 146–148 ꢂC (CH2Cl2–n-hexane) (lit.35 146–148 ꢂC),
N-phenylacetohydroxamic acid (14%), mp 65–67 ꢂC (CH2Cl2–n-
hexane) (lit.36 65–66.5 ꢂC) and azoxybenzene (22%), mp 33–35 ꢂC
(ethanol) (lit.37 35 ꢂC).
nmax(KBr)/cmꢀ1 1766, 1681, 1599, 1524, 1488, 1347, 1010 cmꢀ1
; dH
(CDCl3) 7.35 (d, J 8.7 Hz, 2H), 7.41 (d, J 8.4 Hz, 2H), 7.50 (t, J 7.6 Hz,
2H), 7.65 (t, J 7.5 Hz, 1H), 7.78 (d, J 8.7 Hz, 2H), 8.02 (d, J 7.5 Hz,
2H), 8.18 (d, J 8.8 Hz, 2H). Anal. Calcd for C20H13ClN2O5: C, 60.54; H,
3.30; N, 7.06. Found: C, 60.65; H, 3.43; N, 7.55; N-4-chloro-
phenylbenzohydroxamic acid (17%), mp 155–156 ꢂC (CH2Cl2–n-
hexane) (lit.36 156–157 ꢂC) and 4,40-dichloroazoxybenzene (31%),
mp 153–154 ꢂC (ethanol) (lit.37 155–156 ꢂC).
4.3.2. Reaction between 7b and 6a (Table 1, entry 2)
Furnished
O-acetyl-N-(40-nitrobenzoyl)-N-(4-bromophenyl)-
hydroxylamine (14%), mp 164–166 ꢂC (CH2Cl2–n-hexane) (lit.35
162–163 ꢂC), N-4-bromophenylacetohydroxamic acid, (21%), mp
129–130 ꢂC (CH2Cl2–n-hexane) (lit.38 129–130 ꢂC) and 4,40-dibro-
moazoxybenzene (59%), mp 170–171 ꢂC (ethanol) (lit.37 172 ꢂC).
4.3.10. Reaction between 7e and 6b (Table 1, entry 10)
Furnished N-(3-methylphenyl)-benzohydroxamic acid (23%),
mp 67–69 ꢂC (n-pentane) (lit.40 67 ꢂC); O-benzoyl-N-(3-methyl-
phenyl)-40-nitrobenzohydroxamic acid (28%), mp 134–135 ꢂC
(CH2Cl2–n-hexane); HRMS(EI) m/z C21H16N2O5 requires: 376.1059,
found m/z: 376.1069; and 3,30-dimethylazoxybenzene (43%), mp
35 ꢂC (lit.42 55 ꢂC).
4.3.3. Reaction between 7c and 6a (Table 1, entry 3)
Furnished O-acetyl-N-(4-nitrophenyl)-N-phenylhydroxylamine
characterised as the N-benzoyl derivative (13%), mp 112–113 ꢂC
(CH2Cl2–n-hexane), nmax(KBr)/cmꢀ11795, 1680; dH (CDCl3) 2.15 (3H,
s), 7.38–7.55 (7H, m), 8.18 (2H, d, J 8.30 Hz). Anal. Calcd for
C
15H12N2O5: C, 60.00; H, 4.03; N, 9.33. Found: C, 60.29; H, 4.08; N,
9.35%; 4,40-dinitroazoxybenzene (48%); mp 191–192 ꢂC (ethanol)
(lit.37 192 ꢂC) and 4,40-dinitroazobenzene (23%), mp 223–224 ꢂC
(ethanol) (lit.39 222–223 ꢂC).
4.3.11. Reaction between 7a and 6c (Table 1, entry 11)
Furnished N-cyclopropanecarbonyloxy-N-phenyl-4-nitrobenz-
amide (50%), mp 133–135 ꢂC (CH2Cl2–n-hexane) (lit.7a 133–135 ꢂC);
azoxybenzene (19%) and traces of a substance (ca. 1%), presumably
the hydroxamic acid, giving a positive test with FeCl3 spray on TLC.
4.3.4. Reaction between 7d and 6a (Table 1, entry 4)
Furnished
O-acetyl-N-(40-nitrobenzoyl)-N-(4-chlorophenyl)-
hydroxylamine (5%), mp150–153 ꢂC (lit.7a 152–155 ꢂC), N-(4-
chlorophenyl) acetohydroxamic acid (15%), mp 111–112 ꢂC (lit.36
113 ꢂC) and 4,40-dichloroazoxybenzene (30%), mp 153–154 ꢂC
(ethanol) (lit.37 155–156 ꢂC).
4.3.12. Reaction between 7a and 6d (Table 1, entry 12)
Furnished N-(2,2-diphenylcyclopropanecarbonyloxy)-N-phenyl-
40-nitrobenzamide (27d) (23%), mp 133–135 ꢂC (CH2Cl2–n-hexane),
O-(40-nitrobenzoyl)-N-(2,2-diphenylcyclopropancarbonyloxy)-N-
phenylhydroxylamine (27f) (2%), mp 55–56 ꢂC (CH2Cl2–n-hexane)
both identical in all respects (TLC, mp, 1H NMR, IR spectra) with the
respective authentic samples prepared as described below.
4.3.5. Reaction between 7e and 6a (Table 1, entry 5)
Furnished O-acetyl-N-(40-nitrobenzoyl)-N-(3-methylphenyl)-
hydroxylamine7a (11%), N-(3-methylphenyl) acetohydroxamic acid
(19%), mp 67–69 ꢂC (n-pentane) (lit.40 67 ꢂC), N-40-nitrobenzoyl-
3-methylaniline (6%), mp 146–148 ꢂC (lit.41 155 ꢂC) and 3,30-di-
methylazoxybenzene, oil (43%) (lit.42 33–35 ꢂC).
4.4. Preparation of authentic samples 26d–26g and 27c–27f
4.4.1. N-(20,20-Diphenylcyclopropanecarbonyloxy)-N-phenyl-N-4-
nitrobenzamide (27d)
4.3.6. Reaction between 2,4,6-tri-tert-butyl nitrosobenzene (7f)
and 6a (Table 1, entry 6)
Furnished 2,4,6-tri-tert-butylaniline (28%) as the only isolable
product, mp 145–147 ꢂC (lit.43 144.5–145.5 ꢂC), nmax(CH2Cl2)/cmꢀ1
2955, 2928, 2860; dH (CDCl3) 1.30 (9H, s), 1.50 (18H, s), 7.27 (2H, s);
HRMS (EI) m/z C18H31N requires: 261.2456, found: 261.2426.
2,2-Diphenylcyclopropane carboxylic acid7b (100 mg, 0.42 mmol),
DEPC45 (70
mL, 0.46 mmol) and Et3N (59 mL) in CH2Cl2 (6 mL) were
stirred (10 min). Phenylhydroxylamine (48 mg, 42 mmol) in CH2Cl2
(6 mL) was added. After 2 min, the mixture containing 27c was
cooled to ꢀ78 ꢂC, NaH (35 mg; 1.47 mmol) added and the mixture
stirred at that temperature (30 min). 4-Nitrobenzoylchloride
(117 mg, 0.84 mmol) in CH2Cl2 (2 mL) was then added dropwise. On
completion of the reaction (TLC control) HOAc (0.5 mL) in CH2Cl2
(2 mL) was added and the product worked-up in the usual manner
to give a solid. PTLC purification furnished 27d, mp 144–145 ꢂC
(EtOAc–n-hexane); nmax(KBr)/cmꢀ1 1779, 1681; dH (CDCl3) 1.81 (1H,
dd, J 5.0 and 8.0 Hz), 2.24 (1H, t, J 5.3 Hz), 2.68 (1H, t, J 6.3 Hz), 7.09–
7.37 (15H, m), 7.63 (2H, d, J 8.4 Hz), 8.10 (2H, d, J 8.4 Hz); dC 21.6,
26.0, 42.0, 123.3, 126.9, 127.2, 127.5, 127.6, 128.4, 128.6, 129.2, 129.7,
4.3.7. Reaction between 7b and 6b (Table 1, entry 7)
Furnished O-benzoyl-N-(40-nitrobenzoyl)-N-(4-bromophenyl)-
hydroxylamine (21%), mp 142–144 ꢂC (CH2Cl2–n-hexane),
nmax(KBr)/cmꢀ1 1770, 1680; dH (CDCl3) 6.95–7.21 (3H, m), 7.34 (2H,
d, J 8 Hz), 7.53 (2H, d, J 9.0 Hz), 7.75 (2H, d, J 8.0 Hz), 8.14–8.29 (4H,
m). Anal. Calcd for C20H13N205: C, 54.44; H, 2.97; N, 6.35. Found: C,
53.91; H, 4.53; N 6.46%; N-4-bromophenylbenzohydroxamic acid
(25%), mp 160–161 ꢂC (CH2Cl2–n-hexane) (lit.44 161–163 ꢂC) and

L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
7769
138.6, 139.5, 143.7, 164.60, 168.0; HRMS (EI) m/z: C29H22N2O5
requires: 478.152872, found: 478.15364.
31.8, 30.7, 49.9, 123.7, 126.3, 127.8, 128.5, 129.6, 131.2, 132.6, 138.8,
143.9, 151.0, 162.3; HRMS (EI) m/z: C29H24N2O5 requires: 480.16862,
found: 480.16778.
4.4.2. N-Hydroxy-N-phenyl-2,2-diphenylcyclopropane-
carboxamide (27e)
2,2-Diphenylcyclopropane carboxylic acid (220 mg, 0.92 mmol)
in benzene (2.2 mL) containing DMF (2.2 mL) was mixed with SOCl2
4.5.5. N-(40,40-Diphenylbutyroyloxy)-N-phenyl-N-4-nitro-
benzamide (26e)
Prepared from 26d (101 mg; 0.42 mmol) as described for 27c,
in 15% yield, as a colourless solid (30 mg, 15%), mp 110–112 ꢂC
(n-hexane–EtOAc); nmax(KBr)/cmꢀ1 1785, 1676; dH (CDCl3) 2.39 (4H,
m), 7.17–7.37 (15H); dC (CD3CN) 30.6, 30.7, 50.8, 124.2, 126.1, 127.5,
127.6, 128.8, 129.8, 130.5, 130.6, 140.7, 141.1, 145.3, 150.3, 172.0;
HRMS (EI) m/z: C29H24N2O5 requires: 480.1685, found: 480.1676.
(100 mL, 1.37 mmol). The resulting acid chloride 27b formed in situ
was treated with NaHCO3 (78 mg, 0.92 mmol) and PhNHOH
(101 mg, 0.92 mmol) in ethyl ether containing a few drops of
water. The solid, which was obtained, after the usual work-up, on
crystallisation gave hydroxamic acid 27e (46%), mp 133–135 ꢂC (n-
hexane–EtOAc); nmax(KBr)/cmꢀ1 3177, 1626; dH (CDCl3) 1.56 (1H, dd,
J 4.8 and 8.0 Hz), 2.39 (1H, t, J 5.0 Hz), 2.59–2.61 (1H, m), 7.0–7.53
(15H, m), 8.69 (1H, s, exchangeable with D2O); dC 19.9, 26.9, 40.2,
123.3, 126.5, 127.0, 127.6, 128.3, 128.4, 129.4, 139.6, 144.6, 157.0;
HRMS (EI) m/z: C22H19NO2 requires: 329.14157, found: 329.14130.
4.6. Nitrone triflate 21a from 20a and 7a
2-(10-Acetoxyethyl)-3,4-dimethylthiazolium triflate (20a) (70 mg,
0.2 mmol), PhNO (21 mg, 0.2 mmol) in CH2Cl2 (10 mL) was treated
with Et3N (0.2 mmol) at rt. The precipitate that separated was
collected by decantation and recrystallised (butanone–Et2O) to give
21a (95%), mp 197–198 ꢂC; lmax (CH3CN) 349 nm (3max 18,600);
nmax(KBr)/cmꢀ1 1580, 1160; dH (DMSO-d6) 8.06 (1H, s), 7.68–7.63
(5H, m), 4.23 (3H, s), 2.65, 2.64 (6H, 2s). Anal. Calcd for
4.4.3. N-(40-Nitrobenzoyloxy)-N-phenyl-2,2-diphenylcyclo-
propanecarboxamide (27f)
Compound 27e (100 mg, 0.30 mmol), DMAP (44.3 mg,
0.36 mmol) on treatment with 4-nitrobenzoylchloride (67.4 mg,
0.36 mmol) in THF (2 mL) gave 27f as a colourless solid (125 mg,
86%), mp 55–56 ꢂC (n-hexane-EtOAc); nmax(KBr)/cmꢀ1 1771, 1692;
dH (CDCl3) 1.61 (1H, s), 2.38 (1H, t, J 5.2 Hz), 2.52 (1H, m), 7.00–7.52
(15H, m), 8.23 (4H, dd, J 8.8 and 2.4 Hz); dC 20.1, 28.1, 41.2, 123.6,
126.6, 127.1, 127.6, 128.3, 128.4, 129.8, 131.2, 132.6, 139.3, 144.6,
151.0, 162.5; HRMS (EI) m/z: C29H22N2O5 requires: 478.15287,
found: 478.15203.
C14H15F3N2O4S2: C, 42.42; H, 3.81; N, 7.07%. Found: C, 42.24; H,
3.72; N, 7.02%.
4.7. Nitrone triflate 21b from 20a and 7c
Similarly 20a and 4-NO2C6H4NO furnished nitrone 21b (48%) as
a colourless solid, mp 124–125 ꢂC; nmax(KBr)/cmꢀ11570; dH (DMSO-
d6) 2.64 (3H, s), 4.23 (3H, s), 7.98 (2H, d, J 7.8 Hz), 8.13 (1H, s), 8.49
(2H, d, J 7.8 Hz). Anal. Calcd for C14H14F3N3O6S2: C, 38.10; H, 3.20; N,
9.52%. Found: C, 38.01; H, 3.17; N, 9.49%.
4.5. Preparation of butyric acid derivatives 26a and 26b
4.5.1. Ethyl 4,4-diphenylbutyrate (26a)
Ethyl 2,2-diphenylcyclopropane carboxylate7b (16d) (0.6 g,
2.52 mmol) in ethanol (30 mL) containing concd sulfuric acid
(0.3 mL) was hydrogenated in the presence of Pd/C (0.6 g) at rt
(24 h) at atmospheric pressure. Filtration over Celite followed by
evaporation of the solvent furnished ethyl 4,4-diphenylbutyrate
(26a) as a pale yellow oil (0.575 mg, 96%); dH (CDCl3) 1.24 (3H, t, J
7.1 Hz), 2.28 (2H, t, J 3.9 Hz), 2.36–2.42 (2H, m), 3.93 (1H, t, J 4 Hz),
4.10 (2H, q, J 7.2 Hz), 7.08–7.31 (10H, m); dC 14.2, 26.4, 32.8,
50.5, 60.3, 126.3, 127.7, 128.5, 144.1, 173.4; HRMS (EI) m/z: C18H20O2
requires: 268.14633, found: 268.14568.
4.8. N-Phenylacetohydroxamic acid from silylether
triflate 20b
Compound 20b (145 mg, 0.38 mmol) and PhNO (40 mg,
0.38 mmol) in CH2Cl2 (10 mL) was treated with Et3N (0.53 mL,
0.38 mmol) at rt. The organic phase was extracted with ice-cold
aq NaOH (0.5 N) and acidified (1 N HCl) to give N-phenylaceto-
hydroxamic acid (47 mg, 82%) identical with an authentic sample.
4.9. Kinetic experiment
4.5.2. 4,4-Diphenylbutyric acid (26b)
The kinetic disappearance of nitrosobenzene in a CH2Cl2 (Uva-
Saponification of 26a (70 mg, 0.26 mmol) by the general pro-
cedure7b furnished the corresponding acid 26b (64%), mp 89–90 ꢂC
(n-hexane–Et2O); nmax(KBr)/cmꢀ1 3056; dH (CDCl3) 2.31–2.43 (4H,
m), 3.94 (1H, t, J 7.6 Hz), 7.17–7.13 (10H, m), 11.14 (1H, s, ex-
changeable with D2O); dC 30.1, 30.2, 50.2, 126.5, 127.9, 128.6, 144.0,
151.0, 179.5; HRMS (EI) m/z: C16H16O2 requires: 240.11503, found:
240.11453.
solÒ) solution, containing 3,4-dimethyl-2-(
a
-deuterio-
-hydroxyethyl)-thiazolium
triflates and Et3N, was followed by UV using the chromophore at
a-hydroxy-
ethyl)-thiazolium or 3,4-dimethyl-2-(
a
l
750 nm under pseudo-first order conditions, in a thermostated
flask at 20ꢄ1 ꢂC. Data were acquired at predetermined intervals
using the m-Quant software. The values of kobs were obtained from
the expression kobs¼ln(AoꢀAf)ꢀln(AꢀAf)t, where kobs represents
the first order rate constant (sꢀ1), Ao, Af and A represent the
absorbance at 750 nm at the beginning, at the end and at time t,
respectively. A good linearity was found for the first 70% of
conversion.
4.5.3. N-Hydroxy-N-phenyl-4,4-diphenylbutyramide (26f)
Prepared via acid chloride 26c and PhNHOH, mp 122–124 ꢂC
(n-hexane–EtOAc); nmax(KBr)/cmꢀ1 3170, 1626; dH (CDCl3) 2.28–
2.34 (2H, m), 3.90 (1H), 7.16–7.35 (15H), 8.97 (1H, exchangeable
with D2O); dC 30.4, 31.0, 50.3, 126.3, 127.8, 128.5, 129.2, 137.6, 137.9,
143.9, 167.6; HRMS (EI) m/z: C22H21NO2 requires: 332.16505, found:
332.1650.
Acknowledgements
We thank the Foundation for Science and Technology for partial
financial support and for research fellowships (to P.M.C.G., H.T.C.,
J.P.P., I.M., J.R.A.). It is a pleasure to record our gratitude to Ms. Luz
Fernandes and Ms. Carla Rodrigues of REQUIMTE Analytical
Services Laboratories for mass spectra and elemental analyses,
respectively.
4.5.4. N-(4-Nitrobenzoyloxy)-N-phenyl-40,40-diphenyl-
butyramide (26g)
Prepared in 86% yield, as described for 27c, mp 41–42 ꢂC
(EtOAc–n-hexane); nmax(KBr)/cmꢀ1 1770, 1688; dH (CDCl3) 2.26–
2.28 (2H, m), 2.44 (2H, dd, J 7.4 and 15 Hz), 3.98 (1H, t, J 7.8 Hz); dC

7770
L.M. Ferreira et al. / Tetrahedron 64 (2008) 7759–7770
20. PhNHOH underwent a rapid oxidation to PhNO with an excess of TEMPO with
References and notes
or without addition of Et3N.
21. TEMPO in conjunction with Cuþ1 is known to oxidise benzylic, primary and
secondary alcohols. The failure of benzhydrol or benzyl alcohol to undergo any
oxidation under the conditions mentioned (5 equiv, TEMPO) showed that the
2-acetylthiazolium salt was not formed due to the adventitious presence of
such transition metals in the solvent used: Semmelhack, M. F.; Schmid, C. R.;
Cortes, D. A.; Chou, C. S. J. Am. Chem. Soc. 1984, 106, 3374–3376; Dijksman, A.;
Arends, I.; Sheldon, R. A. Org. Biomol. Chem. 2003, 1, 3232–3237.
22. TEMPOL acetate was not present in the reaction mixture before Ac2O addition.
23. Anderson, J. E.; Corrie, J. E. T. J. Chem. Soc., Perkin Trans. 2 1992, 1027–1031.
24. Paleos, C. M.; Dais, P. J. Chem. Soc., Chem. Commun. 1977, 345–346.
25. TEMPO is reported to oxidise 9,10-dihydroanthracene and 1,4-cyclohexadiene
to the corresponding aromatic compounds: Ciriano, M. V.; Korth, H. G.;
Scheppingen, W. B.; Mulder, P. J. Am. Chem. Soc. 1999, 121, 6375–6381.
26. In this context it is interesting to note that in reactions of pyruvate oxidore-
ductases the radical derived from a one-electron transfer to thiamine di-
phosphate may be a universal intermediate because these enzymes typically
carry one-electron oxidations mediated by iron-sulfur centres (see Ref. 1d, p 11
and references therein).
27. (a) For the detection and identification of phenylhydronitroxide from PhNO and
PhNHOH under neutral conditions, see: Russell, G. A.; Geels, E. J.; Smentows,
F. J.; Chang, K. Y.; Reynolds, J.; Kaupp, G. J. Am. Chem. Soc. 1967, 89, 3821–
3828; (b) Mulvey, D.; Waters, W. A. J. Chem. Soc., Perkin Trans. 2 1977, 1868–
1876; (c) Bowman, D. F.; Brokenshire, J. L.; Gillan, T.; Ingold, K. U. J. Am. Chem.
Soc. 1971, 93, 6551–6555; (d) For a detailed discussion on azoxybenzene
formation from PhNHOH and PhNO, see: Pizzolatti, M. G.; Yunes, R. A. J. Chem.
Soc., Perkin Trans. 2 1990, 759–764.
28. It should be noted that a close similarity exists between the reactions of
ArNHOH and 2-acetylthiazolium salts on one hand, and that of ArNO and
hydroxyethyl thiazolium salts on the other, in so far as the products are con-
cerned. However, the fact that hydroxamic acid was never obtained with the
former reagents strongly indicates that its formation does not involve iso-
merisation of ArNHOAc.
29. A kinetic isotope effect of 4–6 is typical of a primary isotope effect for proton
transfers from carbon acids to hydroxide ion, see: Casamassina, T. E.; Huskey,
W. P. J. Am. Chem. Soc. 1993, 115, 14–20; Barletta, G. L.; Zou, Y.; Huskey, W. P.;
Jordan, F. J. Am. Chem. Soc. 1997, 119, 2356–2362.
30. Perrin, D. D.; Armarego, W. L. F.; Perrin, D. R. Purification of Laboratory Chem-
icals, 2nd ed.; Pergamon: Oxford, 1980.
31. Lutz, R. E.; Lytton, M. R. J. Org. Chem. 1938, 2, 68–75.
32. Nitrobenzene-d5 was purchased from Aldrich Chemical Co. Ltd, Gillingham,
Dorset, UK.
33. Becker, Y.; Bronstein, S.; Eisenstadt, A.; Shvo, Y. J. Org. Chem. 1976, 41, 2496–
2498.
1. (a) Kluger, R. Chem. Rev. 1990, 90, 1151–1169; (b) Kluger, R. Chem. Rev. 1987, 87,
863–876; (c) Thiamine Pyrophosphate Biochemistry; Schellenberger, A., Schowe,
R. L., Eds.; CRC: Boca Raton, FL, 1988; Vol. 1; (d) ThiaminedCatalytic Mechanisms
in Normal and Disease States; Jordan, F., Patel, M. S., Eds.; Marcel Dekker: New
York, NY, 2004.
2. (a) Breslow, R. J. Am. Chem. Soc. 1958, 80, 3719–3726; (b) Breslow, R.; McNelis, E.
J. Am. Chem. Soc. 1959, 81, 3080–3082; (c) Daigo, K.; Reed, L. J. J. Am. Chem. Soc.
1962, 84, 659–662; (d) White, F. G.; Ingraham, L. L. J. Am. Chem. Soc. 1962, 84,
3109–3111; (e) Enders, D.; Niemeier, O.; Henseler, A. Chem. Rev. 2007, 107,
5606–5655 and references therein.
3. (a) Sedewitz, B.; Schleifer, K. H.; Gotz, F. J. Bacteriol. 1984, 160, 273–278; (b)
Gotz, F.; Sedewitz, B. In Biochemistry and Physiology of Thiamine Diphosphate
Enzymes; Bisswanger, H., Ulrich, J., Eds.; VCH: Weinheim, 1991; pp 286–
293.
4. (a) Hilvert, D.; Breslow, R. Bioorg. Chem. 1984, 12, 206–220; (b) Shinkai, S.;
Yamashita, T.; Kusano, Y.; Manabe, O. Tetrahedron Lett. 1980, 21, 2543–
2546; (c) Shinkai, S.; Yamashita, T.; Kusano, Y.; Manabe, O. J. Org. Chem.
1980, 45, 4947–4952; (d) Yano, Y.; Tsukagoshi, Y. J. Chem. Res., Synop. 1984,
406–407.
5. (a) Corbett, M. D.; Corbett, B. R.; Doerge, D. R. J. Chem. Soc., Perkin Trans. 1 1982,
345–350 and references therein; (b) Corbett, M. D.; Doerge, D. R.; Corbett, B. R.
J. Chem. Soc., Perkin Trans. 1 1983, 765–769; (c) Seayad, J.; Patra, P. K.; Zhang, Y.;
Ying, J. Y. Org. Lett. 2008, 10, 953–956 and references therein; (d) Ferreira, L. M.;
Chaves, H. T.; Lobo, A. M.; Prabhakar, S.; Rzepa, H. S. J. Chem. Soc., Chem.
Commun. 1993, 133–134 and references therein.
6. Both O-acyl arylhydroxylamines and N-arylhydroxamic acids are implicated in
carcinogenesis, see: Flammang, T. J.; Westra, J. G.; Kadlubar, F. F.; Beland, F. A.
Carcinogenesis 1985, 6, 251–258; Yoshioka, T.; Uematsu, T. Biochem. J. 1993, 290,
783–790.
7. (a) Ferreira, L. M.; Lobo, A. M.; Prabhakar, S.; Teixeira, A. C. Tetrahedron 1999, 55,
3541–3552; (b) Compound 16d prepared by adoption of the procedure de-
scribed for (–)-menthyl 2,20-diphenylcyclopropanecarboxylate: Impastato, F. J.;
Barash, L.; Walborsky, H. M. J. Am. Chem. Soc. 1959, 81, 1514–1515.
8. Jordan, F. Nat. Prod. Rep. 2003, 20, 184–201.
9. (a) Washabaugh, M. W.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 674–683; (b)
Washabaugh, M. W.; Jencks, W. P. J. Am. Chem. Soc. 1989, 111, 683–692.
10. (a) Barletta, G. A.; Chung, C.; Rios, C. B.; Jordan, F.; Schlegel, J. M. J. Am. Chem.
Soc. 1990, 112, 8144–8149; (b) Nakanishi, I.; Itoh, S.; Fukuzumi, S. Chem.dEur. J.
1999, 5, 2810–2818.
11. For detection of phenylhydronitroxide, formed probably by an H atom ab-
straction from NADHP and ascorbic acid by PhNO, see: Fujii, H.; Koscielniak, J.;
Kakinuma, K.; Berliner, L. J. Free Radic. Res. 1994, 21, 235–243.
12. Gronchi, G.; Tordo, P. Res. Chem. Intermed. 1993, 19, 733–753.
13. Prepared from nitrobenzene-d5 (see Section 4).
34. Erlenmeyer, H.; Baumann, H.; Sorkin, E. Helv. Chim. Acta 1948, 31, 1978.
35. Lobo, A. M.; Marques, M. M.; Prabhakar, S.; Rzepa, H. S. J. Org. Chem. 1987, 52,
2925–2927.
36. Baumgarten, H. E.; Staklis, A.; Miller, E. M. J. Org. Chem. 1965, 30, 1203–1206.
37. Gore, P. H.; Wheeler, O. H. J. Am. Chem. Soc. 1956, 78, 2160–2163.
38. Novak, M.; Pelecanou, M.; Roy, A. K.; Andronico, A. F.; Plourde, F. M.;
Olefirowicz, T. M.; Curtin, T. J. J. Am. Chem. Soc. 1984, 106, 5623–5631.
39. Heilborn, I.; Bunbury, H. M. Dictionary of Organic Compounds, 4th ed.; Eyre &
Spottswood: London, 1965.
40. Gupta, V. K.; Tandon, S. G. J. Indian Chem. Soc. 1971, 48, 753–756.
41. Grammaticakis, P. Bull. Soc. Chim. Fr. 1960, 1956–1968.
42. McKillop, A.; Raphael, R. A.; Taylor, E. D. J. Org. Chem. 1970, 35, 1670–1672.
43. Bartlett, P. D.; Roha, M.; Stiles, R. M. J. Am. Chem. Soc. 1954, 76, 2349–2353.
44. Ayyangar, N. R.; Brahme, K. C.; Kalkote, U. R.; Srinivasan, K. V. Synthesis 1984,
938–941.
14. Asirvath, M.; Hawley, M. D. J. Electroanal. Chem. 1974, 57, 179–190.
15.
A reaction carried out in CH2Cl2, with NaH as the base, gave PhNHOAc,
PhN(OH)Ac, azoxybenzene and aniline in 20, 18, 8 and 25% yields, respectively.
16. Simultaneous detection of PhNH–Oꢁ and the putative enoxyl proved difficult
since the width and strength of signals of the former (44 G) would effectively
mask the spectrum due to enoxyl (33 G). The latter spectrum can be generated
from 6a, O2 and Et3N in CH2Cl2.
17. Tanko, J. M.; Phillips, J. P. J. Am. Chem. Soc. 1999, 121, 6078–6079.
18. The ring opening tendency of diphenylcyclopropyl methyl radical is reported to
be an order of magnitude greater than that of the unsubstituted one: New-
comb, M.; Johnson, C. C.; Manek, M. B.; Varick, T. R. J. Am. Chem. Soc. 1992, 114,
10915–10921.
19. It is tacitly assumed that the rearranged radical once formed would pick up a H
atom or an electron from any reducing species existing in the reaction mixture
rather than re-close.
45. Shioiri, T.; Yokoyama, Y.; Kasai, Y.; Yamada, S. Tetrahedron 1976, 32, 2211–2217.