3674
A. D. Abell et al. / Tetrahedron Letters 43 (2002) 3673–3675
hydrolysis of 1,3 was coupled with a range of amino
acid esters, using standard 1-(3-dimethylaminopropyl)-
3-ethylcarbodiimide hydrochloride (EDCI) peptide cou-
pling methodology,4 to give 3a–e in good yields.
Compound 3c was also prepared in 89% yield using
standard DCC–peptide coupling methodology. Next,
separate dichloromethane solutions of each of 3a–e,
containing diisopropylethyl amine (DIPEA) and a cata-
lytic amount of 4-dimethylaminopyridine (DMAP),
were treated with hydrocinnamoyl chloride. The red
pyrolizin-3-ones 4 were isolated after stirring at rt for
24 h (Scheme 1). The structures of these derivatives
were fully characterised by one- and two-dimensional
NMR spectroscopy, mass spectrometry, ultraviolet
spectroscopy and in the case of 4a, X-ray crystallogra-
phy. Independent conformation of the structures of the
red tags 4a, 4c and 4e was provided by treating each
with sodium methoxide in methanol to give the colour-
less pyrrole acrylic esters 5a, 5c and 5e in yields of 85,
80 and 82%, respectively (Scheme 1).† This reaction also
provides a means to remove the red coloration of the
tag.
Scheme 2. (a) BOP, DIPEA, either
L-LeuGlyOMe·HCl (45%)
or -PheGlyOMe·HCl (49%); (b) DMAP, DIPEA, CH2Cl2
L
then Ph(CH2)2COCl (7a, 33%), (7b, 34%).
Next, we demonstrated that the method could be
applied to dipeptides.
-phenylalanylglycine methyl ester were separately
L
-Leucylglycine methyl ester and
L
reacted with
2
to give the corresponding N-
pyrroloylpeptides 6a and 6b, respectively (Scheme 2).
Reaction of each of these with hydrocinnamoyl chlo-
ride, as described for 3 above, gave the dipeptide-based
pyrolizin-3-ones 7 which were fully characterised.
Scheme 3. (a) 20% piperidine in DCM; (b) 2, EDCI, HOBt,
DIPEA; (c) DMAP, DIPEA, CH2Cl2 then Ph(CH2)2COCl;
(d) 20% TFA/DCM; (e) CH2N2; (f) glycine methyl ester
hydrochloride, EDCI, BOP, DIPEA.
Finally, we investigated whether the method could be
used to develop a red tagging system for resin bound
amino acids (Scheme 3).5 To this end, FMOC-glycine
Wang resin 8 was deprotected by reaction with pipe-
ridine and the resulting free amine 9 was coupled with
the pyrrole acid 2 to give the N-pyrroloyl protected
resin bound amino acid 10. Reaction of this sample
with hydrocinnamoyl chloride, as per the formation of
3 in Scheme 1, gave the red resin beads, 11 (see Fig. 1
for a photograph). The resin bound intermediate 10
was characterised by cleaving it from the resin on
treatment with TFA. Methylation of the resulting car-
boxylic acid 12, with diazomethane, gave 13, which was
shown to be identical to a sample independently synthe-
sised from 2 (see step f in Scheme 3).
The method reported here for the preparation of
pyrrolizin-3-ones is much milder than existing literature
methods.6 The reaction is assumed to occur via N-
acylation of 3 with subsequent intramolecular Knoeve-
† It is interesting to note that in the case of the leucine and phenylala-
nine examples 5c and 5e, only the (Z)-isomer was observed by 1H
NMR. However, in the case of the glycine example 5a, a mixture of
the (Z)- and (E)-isomers was obtained in a ratio of 3:1 (by 1H
NMR). Subsequent recrystallization of this mixture gave a pure
sample of the (E)-isomer which was fully characterized. The assign-
ment of a (Z)-configuration for the alkene 5c (and by analogy for
5a and 5e) was based on an observed strong positive NOE enhance-
ment between the acrylic proton and the methylene protons of the
benzyl group.
Figure 1. Resin bound 11.
nagel-type condensation of the resulting N-acylated
formylpyrrole intermediates. The nature of the acid
chloride used in these reactions would appear to be
important. An equivalent reaction of 3c, using acetyl
chloride in place of hydrocinnamoyl chloride, gave 14
rather than the corresponding pyrrolizin-3-one.