Passerini multicomponent reaction of protected a-aminoaldehydes as a tool for
combinatorial synthesis of enzyme inhibitors
Luca Banfi,* Giuseppe Guanti* and Renata Riva
Università di Genova, Dipartimento di Chimica e Chimica Industriale, and C.N.R., CSCCCA, via Dodecaneso 31,
I-16146 Genova, Italy. E-mail: banriv@chimica.unige.it, guanti@chimica.unige.it
Received (in Liverpool, UK) 10th March 2000, Accepted 20th April 2000
Three-component Passerini condensation of N-Boc-a-ami-
noaldehydes with various isocyanides and carboxylic acids
leads, after Boc-deprotection/transacylation, to complex
peptide-like structures containing an a-hydroxy-b-aminoa-
cid unit or, after oxidation, an a-oxo-b-aminoacid unit.
course of the synthesis; in the other case the yields were low and
no further manipulation of the condensation adduct was
attempted.5b
The -N-Boc protected a-aminoaldehydes employed in this
L
work are all known4c,6 and have been prepared either through
LiAlH4 reduction of Weinreb hydroxamates,6b or by Swern
oxidation of the corresponding N-Boc aminoalcohols,6d,e in turn
prepared by BH3·Me2S reduction of the N-Boc aminoacids.6a
The latter are all commercially available compounds except for
O-methyl-N-Boc-serine, which was prepared as previously
described.7
Isocyanide employing multicomponent reactions have emerged
as very powerful tools for the combinatorial synthesis of various
pharmacologically important derivatives.1 Of the two classical
reactions belonging to this family, the Ugi condensation has
been more widely studied and used in the generation of
chemical libraries. On the other hand, the Passerini reaction,
although older, has been employed less in combinatorial
chemistry.2 The reasons for this lower success are associated
with the fact that a four-component condensation (Ugi)
introduces a higher degree of diversity than a three-component
one (Passerini). Moreover, the two amide bonds that link the
components in Ugi adducts are more suitable for the synthesis
of peptidomimetics, than the combination of one ester and one
amide bond produced in the Passerini reaction. Finally,
intramolecular variants declass the Passerini process to a two-
component reaction, making it less suited for combinatorial
chemistry.
In this work we show that, when protected a-aminoaldehydes
are employed in Passerini condensation, a simple rearrangement
of the reaction products allows an easy combinatorial entry to
peptide-like structures, making this old methodology more
valuable, particularly in the field of peptidomimetics and
enzyme inhibitors.
The general strategy is depicted in Scheme 1 and involves
condensation of N-Boc protected a-aminoaldehydes 1 with
various isocyanides and various carboxylic acids, followed by
one-pot Boc deprotection and acyl migration. This two step
protocol gives rise to complex peptide-like substances 3
possessing a central a-hydroxy-b-aminoacid unit. This type of
monomer has been widely used in the synthesis of enzyme
inhibitors.3 Moreover, a simple oxidation will produce oligo-
peptides 4 containing an a-oxo-b-aminoacid unit, which is an
even more attractive structure, thanks to its similarity with
protease transition state.4
Table 1 reports the results obtained in this two-step protocol
for various combination of six N-Boc-a-aminoaldehydes, six
isocyanides and ten carboxylic acids (including also protected
aminoacids). In particular, isocyanides and carboxylic acids
were chosen in order to check the effect of various kinds of side
chains with different steric and electronic requirements. All
isocyanides and carboxylic acids were commercially available,
apart from benzyl 3-isocyanopropionate, which was prepared in
two steps (1, BnOH, DCC, DMAP, 70%; 2, diphosgene, N-
methylmorpholine, CH2Cl2, 215 °C, 67%) from known N-
formyl-b-alanine.8
Although we did not perform all the 360 possible combina-
tions, we think that the 20 examples shown represent a good
‘statistical sample’, which gives an idea of the generality of the
presented methodology. In all cases, even when the reaction
involved bulky substrates (e.g. entry 18), the yields of the
Passerini condensation were satisfactory. Surprisingly, notwith-
standing the wide variety of substrates employed, the ster-
eoselectivity was at nearly the same level in all cases, being ca.
2+1. In view of combinatorial applications, the modest
diastereoselection is not necessarily a drawback. Moreover, if
the final targets are a-oxoamides 4, diastereoselection at this
level is unimportant.
When enantiomerically pure carboxylic acids were employed
(entries 10–12, 14 and 15), only two diastereoisomers were
detected by NMR spectroscopy, indicating that the starting a-
aminoaldehydes do not undergo significant racemization under
the reaction conditions. For entries 14 and 15 the diaster-
eoisomeric mixtures were also checked by HPLC, which
showed only 2–5% of isomers deriving from aldehyde racemi-
zation. This percentage was shown to depend on the method of
preparation of the a-aminoaldehyde, the reduction–oxidation
protocol turning out better from this point of view.
Only two examples of Passerini reactions involving protected
a-aminoaldehydes have been reported previously.5 In one case,
however,5a the carboxylic component was not retained during
In most cases, treatment of the Passerini adducts 2 with
CF3CO2H led to both Boc cleavage and in situ transacylation to
give directly the oligopeptides 3. However under these
conditions, the acyl migration was not always complete and in
some cases (entries 16–18) it did not take place at all. Thus the
procedure of choice involves brief treatment with Et3N in
CH2Cl2 after CF3CO2H removal. For adducts deriving from
Boc-protected aminoacids as the acid component (entries 12, 14
and 15) this procedure cleaved both Boc groups producing the
expected transacylated compounds, whose identity was further
confirmed by acylation with various reagents of the free amino
group.
The overall yields of the sequence leading to 3 from 1 were
in all cases > 50% and often > 70%.
Scheme 1
DOI: 10.1039/b002027n
Chem. Commun., 2000, 985–986
This journal is © The Royal Society of Chemistry 2000
985