Amidinyl radicals: New and useful intermediates for the synthesis of imidazolines and imidazoles

Article abstract of DOI:10.1039/b304545e

Amidinyl radicals are readily generated from amidoxime benzoates by treatment with a stannane-diazo initiator or with Ni-AcOH and captured by an internal olefin to give the corresponding imidazoline.

Full text of DOI:10.1039/b304545e

Amidinyl radicals: new and useful intermediates for the synthesis of
imidazolines and imidazoles
Dominique Gennet, Samir Z. Zard* and Haiwen Zhang
Laboratoire de Synthèse Organique associé au CNRS, Ecole Polytechnique, 91128 Palaiseau, France.
E-mail: zard@poly.polytechnique; Fax: +33(0)169333851; Tel: +33 (0)169334872
Received (in Cambridge, UK) 25th April 2003, Accepted 3rd June 2003
First published as an Advance Article on the web 23rd June 2003
Amidinyl radicals are readily generated from amidoxime
benzoates by treatment with a stannane–diazo initiator or
with Ni–AcOH and captured by an internal olefin to give the
corresponding imidazoline.
As part of our work on the generation and capture of nitrogen
centred radicals,1 we had briefly examined the possibility of
accessing 1,2-diamines through the internal capture of ureidyl
radicals such as 3 (Scheme 1). Unfortunately, our initial efforts
were frustrated by a disappointingly low yield in the cyclisation
step, as shown by the inefficient conversion of 1 into
imidazolinone 2.2 This is presumably caused by a slow rotation
Scheme 2 Synthesis of an imidazoline.
around the amide-like bond so that the equilibrium leading to
the radical intermediate with the correct geometry for cyclisa-
tion cannot compete with premature hydrogen abstraction from
the stannane.
This problem is well known in the analogous and much better
documented case of amides.3 Amidinyl radicals 4, in contrast,
would not be expected to suffer from such limitations and
appeared from the outset to be ideally suited for our purposes.
The reactivity of these intermediates, as far as we could tell, had
not hitherto been studied, despite their obvious synthetic
potential.
Oxime benzoates were found in the past to be convenient
precursors for iminyl radicals4 and, in principle, the analogous
amidoxime benzoates could act as substrates for generating the
corresponding amidinyl radicals. The amidoxime benzoates are
easily assembled by reacting an allylic amine5 with an
oximinoyl chloride, itself obtained by chlorination of an
aldoxime, followed by benzoylation. This sequence is illus-
trated in Scheme 2 by the synthesis of 7a, obtained as essentially
one geometrical isomer; but since this had no influence on the
subsequent step no attempt was made to establish which of the
two isomers it was. We were indeed pleased to find that slow
addition of a solution of tri-n-butylstannane and AIBN in
toluene to a refluxing solution of amidoxime benzoate 7a in the
same solvent indeed resulted in the smooth formation of the
corresponding imidazoline 8a (88%).
This reaction was extended to various amidoxime benzoates,
as shown by the results compiled in Table 1. Bicyclic and spiro
imidazolines could be readily prepared as well as derivatives
containing quaternary centres. The synthesis of the glucal
derived imidazoline is especially noteworthy. Furthermore, we
later found that slow addition of the stannane was not necessary:
mere heating of the amidoxime benzoate (0.5 mmol), tri-n-
butyltin hydride, and 0.2 equiv. of 1,1A-azobis(cyclohex-
anecarbonitrile) (ACCN) in refluxing toluene (5 ml) was
sufficient to bring about the desired transformation in good
yield.† By analogy with the case of iminyls,6 the reduction of
the amidinyl radicals by the stannane appears to be sufficiently
slow to allow the cyclisation to proceed unimpeded by
premature reduction even under relatively concentrated condi-
tions. This represents a notable simplification in the experi-
mental procedure.
The use of allyl tri-n-butylstannane in the case of substrate 7b
resulted in the clean formation of allyl imidazoline 9 (Scheme
3). As expected allylation occurs from the least hindered exo
face to give the isomer shown. In addition to being precursors of
1,2-diamines by acid hydrolysis,7 imidazolines can be oxidised
to imidazoles.8 The latter transformation is illustrated by the
synthesis of imidazole 11 by saponification of 8b with
methanolic sodium hydroxide, followed by heating the resulting
imidazoline 10 in toluene under air in the presence of palladium
over charcoal.
Scheme 1 Cyclisation of ureidyl and amidinyl radicals.
Scheme 3 Cyclisative allylation and synthesis of imidazole.
1870
CHEM. COMMUN., 2003, 1870–1871
This journal is © The Royal Society of Chemistry 2003

Table 1 Synthesis of imidazolines 8 from amidoxime benzoates 7
We have briefly examined the possibility of generating and
capturing the amidinyl radical by electron transfer from metallic
nickel to the amidoxime ester. We had found previously that
this method was useful for producing iminyl radicals from
oxime esters.9 Indeed, heating a mixture of amidoxime benzoate
7j, nickel powder, and diphenyl diselenide in toluene and acetic
acid resulted in the formation of selenide 12a but in only poor
yield (29%; Scheme 4). Somewhat better yields were observed
starting from derivatives 7k and 7h. Despite the modest, but as
yet unoptimised yields, this approach has the advantage of
avoiding tin-containing reagents and allowing the synthesis of
selenides.
In summary, these preliminary results demonstrate the
potential of amidinyl radicals for the synthesis of imidazolines
and imidazoles. The precursors are readily available and the
radical cyclisation can be easily incorporated into various
tandem sequences, thus opening a straightforward access to a
variety of complex structures.
Notes and references
†
Typical experimental procedure:
Synthesis of amidoxime benzoates 7: to a solution of aldehyde (20 mmol)
in methanol (20 mL) were added hydroxylamine hydrochloride (22 mmol)
and sodium acetate (24 mmol). After 20 min, the mixture was diluted with
water and extracted with ether. The organic layer was dried and
concentrated. The residue was dissolved in DMF (10 mL) and NCS (20
mmol) was added portion-wise. Once the reaction was complete (TLC), the
solution was diluted with water and extracted with ether. The organic layer
was dried, concentrated, and the crude oximinoyl chloride (CAUTION:
oximinoyl chlorides can be harmful and allergenic) added to a solution of
the allylic amine (25 to 50 mmol) in dry ether (50 mL). After 12 hours, the
mixture was diluted with water and extracted with ether. The organic layer
was dried and concentrated. The residue was dissolved in pyridine (15 mL),
and benzoyl chloride (50 mmol) was added dropwise at room temperature.
After 12 to 24 hours, the solution was diluted with water, extracted with
ether, and the organic layer washed twice with 1 M citric acid, water,
saturated aqueous NaHCO3, water, saturated aqueous CuSO4 and brine,
then dried and concentrated. Filtration of the oily residue over a silica pad
removes the last traces of amine and other polar impurities from the
amidoxime benzoate 7 which can then be recrystallised from ether. The
overall yield in the sequence was generally greater than 50%.
Radical cyclisation: to a solution of amidoxime benzoate 7 (0.5 mmol) in
toluene (5 mL), degassed by refluxing for a few minutes under argon, were
added Bu3SnH (0.6 mmol) and ACCN (30 mg). After 30 minutes at reflux,
the mixture was allowed to cool to room temperature and was concentrated.
The residue was purified by chromatography on silica gel (petroleum ether,
then ethyl acetate–petroleum ether 2 : 8) to give compound 8, which could
generally be recrystallised from ether.
a Yield obtained by simple mixing of reagents (see text).
1 For a review, see: S. Z. Zard, Synlett, 1996, 1148.
2 J. Boivin, A.-C. Callier-Dublanchet, B. Quiclet-Sire, A.-M. Schiano and
S. Z. Zard, Tetrahedron, 1995, 51, 6517.
3 O. M. Musa, J. Horner and M. Newcomb, J. Org. Chem, 1999, 64, 1022
and references cited therein.
4 J. Boivin, A.-M. Schiano and S. Z. Zard, Tetrahedron Lett., 1994, 35,
249.
5 R. Ben Cheikh, R. Chaabouni, A. Laurent, P. Mison and A. Nafti,
Synthesis, 1983, 685; M. Johannsen and K. A. Jorgensen, Chem. Rev.,
1998, 98, 1689.
6 M.-H. Le Tadic-Biadatti, A.-C. Callier-Dublanchet, J. H. Horner, B.
Quiclet-Sire, S. Z. Zard and M. Newcomb, J. Org. Chem, 1997, 62,
559.
7 For some examples, see: J. A. King and F. H. McMillan, J. Am. Chem.
Soc., 1946, 68, 1774; H.-J. Kallmayer and K.-H. Seyfang, Arch. Pharm.,
1980, 313, 603; J. M. Mitchell and N. S. Finney, Tetrahedron Lett., 2000,
41, 8431.
8 M. E. Campos, R. Jimenez, F. Martinez and H. Salgado, Heterocycles,
1995, 40, 841; P. K. Martin, H. R. Matthews, H. Rapoport and G.
Thyagarajan, J. Org. Chem., 1968, 33, 3758; J. L. Hughey, S. Knapp and
H. Schugar, Synthesis, 1980, 489; Y. Amemiya, D. D. Miller and F.-L.
Hsu, Synth. Commun., 1990, 20, 2486; G. M. Dubowchik, L. Padilla, K.
Edinger and R. A. Firestone, J. Org. Chem., 1996, 61, 4676.
9 J. Boivin, A.-M. Schiano, S. Z. Zard and H. Zhang, Tetrahedron Lett.,
1999, 40, 4531.
Scheme 4 Cyclisative selenylation.
CHEM. COMMUN., 2003, 1870–1871
1871