Carbon-carbon bond formation via thermal intermolecular hydrogen atom transfer: Two serendipitous heterocyclic examples

Article abstract of DOI:10.1039/a909891g

Formation of an anomalous 3-nitrothiophene product, encountered during the preparation of potential bioreductive anti-cancer agents, is rationalised in terms of a hydrogen atom transfer mechanism which also accounts for the unexpected formation of previously described 5,5′-biimidazoles. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a909891g

View Article Online / Journal Homepage / Table of Contents for this issue
Carbon–carbon bond formation via thermal intermolecular
hydrogen atom transfer: two serendipitous heterocyclic examples
Richard J. Hamlyn, Richard H. Jones and Christopher A. Ramsden*
1
School of Chemistry and Physics, Keele University, Keele, Staffordshire, UK ST5 5BG
Received (in Cambridge, UK) 16th December 1999, Accepted 18th May 2000
Published on the Web 1st June 2000
Formation of an anomalous 3-nitrothiophene product,
encountered during the preparation of potential bio-
reductive anti-cancer agents, is rationalised in terms of a
hydrogen atom transfer mechanism which also accounts
for the unexpected formation of previously described 5,5Ј-
biimidazoles.
(i) the preparation of two analogues (4 and 5) of the derivatives
3, (ii) an anomalous reaction which appears to proceed via a
mechanism of synthetic and biological interest and (iii) a
comparison of this mechanism with other examples from our
previous studies. In particular we propose that these reactions
proceed via a general mechanism [eqn. (1)] that is equivalent
to intermolecular hydrogen atom transfer followed by radical
coupling.
Due to their fundamentally similar bonding, structurally
diverse molecules containing three-centre, four-electron [3c–4e]
bonds undergo similar modes of reaction, including syn-
addition, ligand coupling and single electron transfer (SET).1
For example, in terms of chemical bonding the nitro group
(–NO2) and xenon difluoride (XeF2) are closely related and
show similarities in chemical behaviour. We have recently
defined conditions under which XeF2 reacts by a SET mechan-
A ϩ DH → A Ϫ ϩ DH → HA ϩ D →
ϩ
ؒ
ؒ
᝽
᝽
HA–D → products (1)
Reaction of 5-chloro-1-methyl-4-nitroimidazole 6 (X = CH,
Y = N)6 with 2-benzylimidazoline 7 (2-benzyl-4,5-dihydro-1H-
imidazole) (2 equiv.) in hot propiononitrile gave the SNAr
product 8 (X = CH, Y = N) (69%), mp 185–186 ЊC† (Scheme
1). Cyclisation of this product at 25 ЊC using KOtBu or DBU
gave the N-oxide 4 (13%), mp 195–198 ЊC as bright red needles
together with the N-deoxy derivative (26%), mp 189–192 ЊC as
orange crystals. In a similar sequence, 5-chloro-1,3-dimethyl-4-
nitropyrazole6 6 (X = N, Y = CMe) gave the N-oxide 5 (26%),
mp 224–226 ЊC as yellow needles and the N-deoxy derivative
(29%), mp 198–201 ЊC as yellow needles. When 2-chloro-3-
nitrothiophene7 9 was treated with the imidazoline 7 (Scheme
2) in propiononitrile a dark red colouration occurred and the
Ϫ
2
ism (XeF2→XeF2
) and this mode of reaction to form the
᝽
radical anions RNO2 Ϫ is also well known for nitro compounds.
᝽
One-electron reduction of nitro compounds also has biological
significance and is a key mechanistic step in the selective activ-
ity against anaerobic bacteria of the widely prescribed drug
metronidazole 1 (R1 = CH2CH2OH, R2 = CH3).3
Selective bioreduction in the hypoxic environment of tumour
cells, leading to formation of toxic products at the site of
action, is also an attractive approach to cancer chemotherapy.
In this context a number of heterocyclic N-oxides, which
formally contain conjugated [3c–4e] bonds, have shown promis-
ing properties and these include the di-N-oxide 2 (SR-4233)4
and the imidazo[1,2-a]pyrido[3,2-e]pyrazine 5-oxides 3 (Scheme
1).5 Our general interest in molecules containing [3c–4e] bonds
and their application to biological problems have led us to
investigate heterocyclic modifications of the derivatives 3 which
might result in increased potency and selectivity. We now report
Scheme 2
product, which formed without heating, was isolated as fine
beige needles, mp 140–142 ЊC and shown to have the constitu-
tion C14H13N3O3S (62% yield). This compound was not the
expected SNAr product (cf. 8) and requires the incorporation of
additional oxygen. When the reaction was carried out under an
oxygen free argon atmosphere no reaction was observed and
product formation only occurred when dioxygen was intro-
duced to the reaction mixture. Under anhydrous conditions the
reaction proceeded to completion demonstrating that water
is not a reactant. Initially the structure of this novel product
was elusive due to its unexpected mode of formation and our
failure to obtain crystals suitable for X-ray crystallography.
Subsequently we have shown that this product is the tertiary
alcohol 10 and this structure is fully supported by its
spectroscopic properties. Furthermore, whereas treatment of
compound 10 with one equivalent of MeI gave the expected
N-methylimidazoline (86%), mp 186–189 ЊC, reaction with
Scheme 1
DOI: 10.1039/a909891g
J. Chem. Soc., Perkin Trans. 1, 2000, 1811–1813
This journal is © The Royal Society of Chemistry 2000
1811

View Article Online
Scheme 3
Scheme 4
excess MeI gave 3-nitro-2-thienyl phenyl ketone 12 (62%), mp
140–144 ЊC. We propose that this reaction proceeds via the
N1,N2-dimethylimidazolinium salt 11 and that the carbene 138
acts as a leaving group (Scheme 2) which is rapidly protonated
to form the imidazolinium salt.
The mechanism shown in Scheme 3 accounts for the form-
ation of the alcohol 10. Initial SET between the electron-rich
imidazoline 7 and the electron-deficient nitrothiophene 12 gives
the contact ion pair 14. The basic radical anion then abstracts a
proton from the acidic radical cation to give a radical pair 15.
At this stage the overall reaction is equivalent to hydrogen atom
transfer (1eϪ ϩ Hϩ) to the nitroheterocycle 9. One-electron
reduction potentials of nitroimidazoles (e.g. 1), as measured by
the calculated LUMO energy, are known to be low. Using the
AM1 method we have calculated the LUMO energies of the
nitroimidazole 4 and the nitropyrazole 5 to be Ϫ0.82 and Ϫ0.87
eV respectively, in agreement with previous calculations on
close analogues.9 The calculated LUMO energy of the nitro-
thiophene 9 (Ϫ1.34 eV) is even lower and consistent with the
proposed anomalous behaviour. The radical pair may then
react together with C–C bond formation (15→16) and elimin-
ation of HCl to give the intermediate 17. The molecule 17 can
be expected to readily give the relatively stable conjugated
radical 18 by reaction with dioxygen. The alcohol 10 is then
formed via the peroxide 19.
Indirect evidence for the involvement of the intermediate 17
in the mechanism of formation of compound 10 was obtained
during attempts to determine structure 10 by X-ray crystal-
lography of simple derivatives. From among unsuitable
elongated needles the crystallographer selected a small cuboid
crystal which was found to be the hydrobromide salt of com-
pound 17. Fig. 1 shows a perspective view and atom labelling
of this salt.10 Attempts to improve the yield of this product
were unsuccessful.
Fig. 1 Perspective view and atom labeling of the X-ray structure of
the HBr salt of the imidazoline 17.
pair 20. Proton transfer may then take place from the acidic
2-alkyl group of the radical cation to the basic radical anion to
give a pair of conjugated radicals 21 which combine to link the
imidazole rings (21→22). Tautomerism with elimination of
water (22→23) and further catalytic reduction of the nitroso
derivative 23 then leads to the isolated bi(imidazole)s after
subsequent condensation of the primary amines with diethyl
ethoxymethylenemalonate.
In addition to the thermal examples described above, we
have recently reported what we believe to be a photochemical
example of this mode of reaction of a heterocycle.11 Thus, the
photoaddition of methyl acrylate to 2-phenylindole gives an
adduct isomeric with the Michael adduct (Scheme 5). We have
We have previously reported9 the formation of 5,5Ј-
bi(imidazole) derivatives during catalytic reduction of 2-alkyl-
4(5)-nitroimidazoles 1 (R1 = H, R2 = alkyl) to the correspond-
ing aminoimidazoles. This novel dehydrodimerisation was
limited to 2-alkyl derivatives and the substituent effect was
difficult to rationalise. We now recognise that this reaction is
probably also mediated by sequential SET and proton transfer,
and that the proton is delivered from the essential 2-alkyl sub-
stituent (Scheme 4). We propose that initial electron transfer
occurs between the electron-rich aminoimidazole (HOMO:
ca. Ϫ8.2 eV)9 and the electron-deficient 4-nitroimidazole
precursor (LUMO: ca. Ϫ0.5 eV)9 to give the contact ion
Scheme 5
1812
J. Chem. Soc., Perkin Trans. 1, 2000, 1811–1813

View Article Online
previously rationalised this reaction in terms of a formal
hydrogen atom transfer followed by rapid radical coupling
[eqn. (1)].11
mmϪ1, R values wR2 = 0.0877, R = 0.0369 (1820 reflections with
F > 4σ(F)), reflections: measured 6383, independent 2281;
Rint = 0.0872.
We consider it unlikely that the thermal nitroheterocycle
reactions described above occur via a hydride transfer mechan-
ism (2eϪ ϩ Hϩ) analogous to that for quinone reactions which
involve initial ion-pair formation.12 It may well be that some
quinone reactions under aprotic conditions also take place via
hydrogen atom transfer (1eϪ ϩ Hϩ) (eqn. 1) to give a solvent
caged semiquinone and donor radical: this pathway was first
proposed by Becker.13 Many quinone reaction products can be
satisfactorily accounted for in terms of initial rate determining
formation of a radical pair followed by regiospecific radical
disproportionation (oxidised products)14 or radical recombin-
ation (addition products).13
Acknowledgements
We thank the Cancer Research Campaign for the award of a
Project Grant to support this work, the EPSRC Crystal-
lography Service (Cardiff) for collection of X-ray data and
Professor Gurnos Jones (Keele) for his interest and suggestions.
Notes and references
† All new compounds were characterised by spectroscopy and
elemental analysis.
1 J. I. Musher, Angew. Chem., Int. Ed. Engl., 1969, 8, 54; C. A.
Ramsden, Chem. Soc. Rev., 1994, 23, 111; C. A. Ramsden, J. Hetero-
cycl. Chem., 1999, 36, 1573.
Experimental
The following procedure was used to prepare compound 10.
Calculations were carried out using the AM1 semi-empirical
method15 and energy was minimised with respect to all
geometrical variables.
2 C. A. Ramsden and R. G. Smith, J. Am. Chem. Soc., 1998, 120,
6842; C. A. Ramsden and R. G. Smith, Org. Lett., 1999, 1, 1591.
3 Nitroimidazoles: Chemistry, Pharmacology and Clinical Applications,
eds. G. E. Adams, A. Breccia and B. Cavalleri, Nato Advanced
Study Institutes Series, Series A: Life Sciences, Plenum Press, New
York and London, 1982, vol. 42; J. A. McFadzean, Flagyl: The Story
of a Pharmaceutical Discovery, Parthenon Publishing Group Ltd.,
1986.
1-(4,5-Dihydro-1H-imidazol-2-yl)-1-phenyl-1-(3-nitro-2-
thienyl)methanol 10
A solution of 2-chloro-3-nitrothiophene 9 (1.0 g, 6.0 mmol) in
propiononitrile (30 ml) containing molecular sieves (4 Å) was
stirred and 2-benzyl-4,5-dihydro-1H-imidazole 7 (1.93 g, 12.0
mmol) in propiononitrile (20 ml) was added dropwise. After
maintaining stirring (24 h) at ambient temperature, the solution
was filtered and evaporated under vacuum to give a dark red
residue that was purified by flash chromatography on silica
(9:1, ethyl acetate–MeOH as eluent) to give a single product
that was recrystallised from EtOH and identified as compound
10 (1.13 g, 62%), fine beige needles, mp 140–142 ЊC (Found: C,
55.59; H, 4.33; N, 13.69. C14H13N3O3S requires C, 55.43; H,
4.32; N, 13.85%); νmax(KBr)/cmϪ1 3428, 1613, 1534, 1496, 1446
and 1333; δH (CDCl3) 1.65–1.85 (1H, br s, NH), 3.69 (4H, s,
2 × CH2), 7.12 (1H, d, J 5.7 Hz, thiophene-H), 7.41–7.43 (3H,
m, phenyl-H), 7.59 (1H, d, J 5.7 Hz, thiophene-H), 7.67–7.69
(2H, m, phenyl-H); δC (CD3OD) 50.21 (2 × CH2), 75.48
(C-OH), 125.81 (CH), 126.31 (CH), 127.91 (CH), 129.05 (CH),
130.18 (CH), 141.99 (C), 146.24 (C), 151.81 (C), 172.56 (C);
4 E. M. Zeman, J. M. Brown, M. J. Lemmon, V. K. Hirst and
W. W. Hirst, Int. J. Radiat. Oncol. Biol. Phys., 1986, 12, 1239.
5 M. A. Naylor, M. A. Stephens, J. Nolan, B. Sutton, J. H. Tocher,
E. M. Fielden, G. E. Adams and I. J. Stratford, Anti-Cancer Drug
Des., 1993, 8, 439.
6 J. Sarasin and E. Wegmann, Helv. Chim. Acta, 1924, 7, 713;
5-chloro-1,3-dimethyl-4-nitropyrazole was purchased from May-
bridge Chemical Co. Ltd.
7 C. D. Hurd and K. L. Kreuz, J. Am. Chem. Soc., 1952, 74, 2965;
G. Consiglio, V. Frenna, C. Arnone, E. Mezzina and D. Spinelli,
J. Chem. Soc., Perkin Trans. 2, 1994, 2187.
8 A. J. Arduengo, III, Acc. Chem. Res., 1999, 32, 913; A. J. Arduengo,
III, J. R. Goerlich and W. J. Marshall, J. Am. Chem. Soc., 1995, 117,
11027; A. J. Arduengo, III, H. V. R. Dias, R. L. Harlow and
M. Kline, J. Am. Chem. Soc., 1992, 114, 5530.
9 A. H. M. Al-Shaar, D. W. Gilmour, D. J. Lythgoe, I. McClenaghan
and C. A. Ramsden, J. Chem. Soc., Perkin Trans. 1, 1992, 2779;
A. H. M. Al-Shaar, R. K. Chambers, D. W. Gilmour, D. J. Lythgoe,
I. McClenaghan and C. A. Ramsden, J. Chem. Soc., Perkin Trans. 1,
1992, 2789.
10 G. M. Sheldrick, SHELX-93, University of Göttingen, 1986.
11 R. H. Jones, C. A. Ramsden and H. L. Rose, J. Heterocycl. Chem.,
1996, 33, 1031.
m/z 303.0675 (M ϩ: C14H13N3O3S requires 303.0678), 286
᝽
(M Ϫ OH) and 257 (M Ϫ NO2).
12 B. M. Trost, J. Am. Chem. Soc., 1967, 89, 1847.
13 H.-D. Becker, J. Org. Chem., 1969, 34, 1203.
14 W. Brown and A. B. Turner, J. Chem. Soc. C, 1971, 2057.
15 M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart,
J. Am. Chem. Soc., 1985, 107, 3902.
Crystal data. C14H14BrN3O2S, M 368, monoclinic, a =
9.051(2), b = 13.430(2), c = 12.144(4) Å, β = 99.870(2)Њ, U =
1454.3(6) Å3, T = 150(2) K, space group P21/n, Z = 4, µ = 2.977
J. Chem. Soc., Perkin Trans. 1, 2000, 1811–1813
1813