Synthesis of N-acyl-N,alpha,alpha-trialkyl and N-acyl-alpha,alpha-dialkyl glycines by selective cleavage of Ugi-Passerini adducts. Qualitative assessment of the effect of substituents on the path and yield of reaction.

Article abstract of DOI:10.1039/b307111c

Several symmetric N-acyl-N,alpha,alpha-trialkyl glycine amides were synthesised by the Ugi-Passerini four-component reaction and subjected to selective cleavage with trifluoroacetic acid. In almost all cases it was possible to obtain the corresponding N-acyl-N,alpha,alpha-trialkyl and N-acyl-alpha,alpha-dialkyl glycines in fair to good yields directly from the reaction adducts. With some of the bulkier compounds our results show that the selectivity of cleavage is concentration dependent with respect to the acid, which suggests kinetically controlled processes. The isolation of a stable oxazolone as the product of some of the reactions seems to confirm that amide cleavage involves in all cases formation of an oxazolone-type derivative.

Full text of DOI:10.1039/b307111c

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis of N-acyl-N,ꢀ,ꢀ-trialkyl and N-acyl-ꢀ,ꢀ-dialkyl
glycines by selective cleavage of Ugi–Passerini adducts.
Qualitative assessment of the effect of substituents on the path
and yield of reaction†
Wei-Qun Jiang, Susana P. G. Costa and Hernâni L. S. Maia*
Department of Chemistry, University of Minho, Gualtar, P-4710-057 Braga, Portugal.
E-mail: hmaia@quimica.uminho.pt
Received 20th June 2003, Accepted 2nd September 2003
First published as an Advance Article on the web 25th September 2003
Several symmetric N-acyl-N,α,α-trialkyl glycine amides were synthesised by the Ugi–Passerini four-component
reaction and subjected to selective cleavage with trifluoroacetic acid. In almost all cases it was possible to obtain the
corresponding N-acyl-N,α,α-trialkyl and N-acyl-α,α-dialkyl glycines in fair to good yields directly from the reaction
adducts. With some of the bulkier compounds our results show that the selectivity of cleavage is concentration
dependent with respect to the acid, which suggests kinetically controlled processes. The isolation of a stable
oxazolone as the product of some of the reactions seems to confirm that amide cleavage involves in all cases
formation of an oxazolone-type derivative.
decided (i) to explore the possibility of carrying out the
above two cleavages selectively in order to isolate the intermedi-
ate N-(4-methoxybenzyl)-α,α-dialkyl glycines and (ii) to evalu-
ate the effect of the different substituents on the efficiency of
the syntheses. Now, we report the preparation of a series of the
above compounds with which we were able to establish cleavage
patterns and conditions required for selectivity. The isolation
Introduction
During the last decades several authors have been concerned
with the investigation of the conformational preferences
imparted to peptides by the inclusion of one or more residues
of α,α-dialkyl glycines in the peptide chain.1 Such preferences
are related to the hindrance of rotation within the α,α-dialkyl
glycine residues due to steric crowding, which makes these resi-
of an oxazolone derivative confirmed that amide cleavage pro-
dues good candidates for incorporation in peptides in order to
ceeds through a cyclic intermediate and allowed evaluation of
grant them special conformational features. Thus, with the aim
the difficulties one can expect when dealing with the bulkier
of developing antagonists or preventing or retarding recog-
compounds.
nition by enzymes, various important applications of the above
amino acids have been devised in connection with the modifi-
Results and discussion
cation of natural peptides or in molecules mimicking them,
usually when restriction of backbone flexibility is required.2
Owing to steric crowding, the synthesis of these uncommon
amino acids and reactions with them and their derivatives are
slow, thus almost always allowing competitive undesired reac-
tions leading to low yields; usually, this results from difficulties
in isolating and purifying the desired products and discourages
the use of this class of compounds.3,4 As a consequence, most
of the work found in the literature with α,α-dialkyl glycines
deals with the simplest of them, i.e. α,α-dimethyl glycine, or
with those compounds where the “two side chains” are tied
up together in a ring and thus kept sufficiently far from the
reaction centres to avoid too strong an interaction with the
peptide backbone; this is the case with 1-aminocyclopentyl-
and 1-aminocyclohexyl-carboxylic acid and their structural
derivatives.5 By taking advantage of the fact that N-acyl-
N,α,α-trialkyl glycine amides undergo selective amide cleavage
with trifluoroacetic acid (TFA)6 and also of the possibility of
cleaving the 4-methoxybenzyl group from amides on boiling
with TFA,7 recently we have been able to solve some difficulties
related to the use of the Ugi–Passerini four-component reac-
tion8 to synthesise α,α-dialkyl glycine derivatives.9 This allowed
us to propose our approach as a general, efficient, and possibly
the best, strategy available for the synthesis of peptides
incorporating residues of these amino acids. However, prior to
engaging in peptide synthesis with these bulky amino acids, we
Synthesis of Ugi–Passerini adducts 1
A series of eight compounds was developed by combination of
two different groups for each of the substituents R1, R3 and R4
(Scheme 1); R2 was kept as 4-methoxybenzyl in order to make
its cleavage possible. For R1 the smaller and the bulkier of those
groups reported in our previous paper were chosen,9 i.e. methyl
and benzyl. In order to ensure a good crystallinity of the com-
pounds and thus facilitate their isolation, R3 was phenyl and
benzyl, which would be generated by the use of benzoic and
phenylacetic acid, respectively, as the acid component in Ugi–
Passerini reactions. Finally, for R4 we chose methoxybenzyl and
cyclohexyl, which are groups of different bulkiness that would
be generated by isonitriles with which we were already familiar
and which had shown good behaviour. The required Ugi–
Passerini adducts were thus obtained in yields varying within
the range 45–83% (Table 1) in reactions taking two to three
weeks to completion at room temperature; no attempts were
made to carry out the reactions at a higher temperature in order
to avoid isonitrile polymerisation that would then increase the
difficulty in isolating the reaction product.
† Electronic supplementary information (ESI) available: experimental
details. See http://www.rsc.org/suppdata/ob/b3/b307111c/
Scheme 1
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 0 4 – 3 8 1 0
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
3804

View Article Online
Scheme 2
Table 1 Synthesis of Ugi–Passerini adducts 1
N,α,α-trialkyl glycine 3f and the dibenzyl oxazolone 5f were
obtained in almost equal amounts, i.e. 47 and 44%, respectively,
the latter being obtained by extracting the above sodium
hydroxide solution with dichloromethane (DCM). With adduct
1h the above two compounds were also obtained but the
oxazolone was the major product (76%). This behaviour reveals
the occurrence of two competitive reactions according to paths
A and B of Scheme 2. While the reaction with 1f required 7.5
hours to completion, that with 1h needed as much as 36 hours;
hence, when the reaction mixture was treated with base, in the
latter case already a much larger amount of oxazolonium 2f
had decomposed into 4f than it had in the former case. The
positive charge at the nitrogen atom of cyclic intermediates 2
must facilitate the nucleophilic attack at the carbonyl group
when they are treated with aqueous base and thus give the
linear products 3. This is not the case with compounds 4, which
are readily converted into the stable oxazolone 5f by deproton-
ation with the base. The phenyl group at position 2 of the ring
may stabilise the cyclic derivatives 4 and 5 by conjugation
(Scheme 2) but help in destabilising intermediates 2 in which the
bulky benzyl groups at the α-carbon atom already tend to assist
elimination of the substituent at the nitrogen atom. Such assist-
ance becomes obvious if one compares the behaviour of 1h
with its analogue 1d, which has methyl instead of benzyl groups
at the α-carbon atom; although cleavage of the latter requires as
much as 24 hours for completion, in this case no oxazolone
could be isolated and the expected open chain compound 3b
where the N-alkyl group is still present was obtained with
a yield of 81% (path A of Scheme 2). To our knowledge, this
was the first time an oxazolone was isolated by acidic cleav-
age of an amino acid amide bond, its structure being fully
supported by IR and NMR spectroscopy, and also by mass
spectrometry. It is noteworthy that this cyclic compound is
stable in aqueous 6 M NaOH and that to hydrolyse it requires
Product
R1
R2
R3
R4
Yield (%)
1a
1b
1c
1d
1e
1f
Me
Me
Me
Me
CH2Ph
CH2Ph
CH2Ph
CH2Ph
Pmba
Pmb
Pmb
Pmb
Pmb
Pmb
Pmb
Pmb
CH2Ph
Ph
CH2Ph
Ph
CH2Ph
Ph
CH2Ph
Ph
Pmb
Pmb
C6H11
C6H11
Pmb
Pmb
C6H11
C6H11
48
77
76
83
45
59
80
48
1g
1h
a Pmb = 4-Methoxybenzyl.
Cleavage of 1 with diluted TFA: synthesis of N-acyl-N-
(4-methoxybenzyl)-ꢀ,ꢀ-dialkyl glycines 3
Adducts 1a–1h were reacted at room temperature with 2% TFA
in acetonitrile until the presence of starting material could no
more be observed by thin layer chromatography. The reactions
with α,α-dibenzyl glycine derivatives 1e–1h required longer
times to completion (9–36 hours) than those with the α,α-di-
methyl analogues 1a–1d (1.5–24 hours), the same applying to
the cyclohexylamides 1c,1d,1g and 1h (4.5–36 hours) as com-
pared to the corresponding 4-methoxybenzylamides 1a,1b,1e
and 1f (1.5–9 hours). The differences correlate well with the
bulkiness of the substituents at R1 and R4, respectively. The
residue obtained by evaporation of the reaction solvent was
taken up in aqueous sodium hydroxide and the pH adjusted to 1
to convert the sodium salt into the required acid. Work-up and
purification by column chromatography in silica-gel afforded
the N-acyl-N-alkylamino acids 3a,3b and 3e (from 1a and 1c,1b
and 1d, and 1e and 1g, respectively) in yields falling within the
range 60–88% (Scheme 2, Table 2). In the case of 1f the expected
Table 2 Cleavage of Ugi–Passerini adducts with diluted acid
Reagenta
R1
R2
R3
R4
Reaction time/hb
Product
Yield (%)
1a
1b
1c
1d
1e
1f
Me
Me
Me
Me
CH2Ph
CH2Ph
Pmb
Pmb
Pmb
Pmb
Pmb
Pmb
CH2Ph
Ph
CH2Ph
Ph
CH2Ph
Ph
Pmb
Pmb
C6H11
C6H11
Pmb
Pmb
1
5
2
11
7
8
3a
3b
3a
3b
3e
3f
88
84
75
81
60
47
44
70
15
76
5f
1g
1h
CH2Ph
CH2Ph
Pmb
Pmb
CH2Ph
Ph
C6H11
C6H11
12
12
3e
3f
5f
a Reactions with TFA at a concentration of 2% and work-up with NaOH followed by acidification. b Measured when at least 95% of the reagent had
been consumed as estimated by TLC.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 0 4 – 3 8 1 0
3805

View Article Online
Table 3 Cleavage of Ugi–Passerini adducts with neat TFA
Reagent
R1
R2
R3
R4
Producta
Yield (%)
1a
1b
1c
1d
1e
1f
Me
Me
Me
Me
CH2Ph
CH2Ph
Pmb
Pmb
Pmb
Pmb
Pmb
Pmb
CH2Ph
Ph
CH2Ph
Ph
CH2Ph
Ph
Pmb
Pmb
C6H11
C6H11
Pmb
Pmb
6a
6b
6a
6b
6e
5f
6f
6e
5f
6f
7h
6f
5f
66
59
72
70
95
79
9
80
40
12
44
94
72
1g
CH2Ph
CH2Ph
Pmb
Pmb
CH2Ph
Ph
C6H11
C6H11
1hb
5f
7h
CH2Ph
CH2Ph
—
H
Ph
Ph
—
C6H11
a All reactions but that with 5f were carried out with neat TFA followed by work-up with NaOH and acidification; the reaction with 5f was carried
out by boiling the substrate with 6 M HCl for 4 hours. b The results are for 10 minute reactions; in a 5 minute reaction the yields of 5f, 6f and 7h were
42, 0 and 52%, respectively.
Scheme 3
boiling with semi-concentrated hydrochloric acid for 4 hours
(Table 3). This seems to differ from what was found by Good-
man and co-workers6 when cleaving α,α-dimethyl glycine deriv-
atives. In fact, although having postulated formation of an
intermediate oxazolonium, these authors did not report having
isolated either such an intermediate or its deprotonated deriv-
ative, suggesting that it would be readily converted into the
corresponding acid by trace water. In the case of the α,α-di-
methyl compounds 1a–1d we were also unable to isolate an
oxazolone, but our results with compounds 1f and 1h confirm
Goodman’s proposal and show that in these two compounds,
the cyclic intermediate does not undergo ring opening even
when treated with strong base.
3 depict these two competitive processes, respectively. With 1h
also only a small amount of 6f (12%) was obtained, but two
major products were isolated both exhibiting no alkyl group
at the nitrogen atom, viz. oxazolone 5f (40%) and the amino
acid amide 7h (45%); these two major products suggest the
competitive cleavages depicted by paths A and B of Scheme 3.
On boiling 7h with neat TFA until no more starting material
could be observed by TLC, which required 2 hours, and
purifying the product solely by chromatography on silica gel,
oxazolone 5f was obtained with a yield of 72% instead of the
expected oxazolonium trifluoroacetate. In order to confirm that
this salt was decomposed in the chromatographic column, a
genuine sample of oxazolone was treated with excess TFA and
chromatographed through a silica gel column; the first fractions
collected from the column contained the deprotonated
compound, which was proved by a mixed melting point of the
original material with the residue obtained from evaporation of
the solvent. Goodman’s results seem to indicate that scission of
the amide bond via an oxazolonium intermediate would require
an alkyl group at the nitrogen atom, whatever its role might be
in inducing the conformation required for internal nucleophilic
attack (by the N-terminal carbonyl oxygen atom to the carb-
onyl group at the C-terminus). However, conversion of amide
7h into the oxazolone 5f in high yield without any visible cleav-
age of the N-acyl group shows that with α,α-dibenzyl glycine
derivatives (i) not only does such a requirement not apply any
more but (ii) also the C-terminal amide bond is more labile than
Cleavage of 1 with neat TFA: synthesis of N-acyl-ꢀ,ꢀ-dialkyl
glycines 6
Having been able to convert six of the above eight adducts 1
into the corresponding N-acyl-N,α,α-trialkyl glycines in good
yields, we set out to obtain all corresponding N-acyl-α,α-dialkyl
derivatives 6 by refluxing the above adducts with neat TFA
(Scheme 3). By working up the reaction mixture with aqueous
base the N-acylamino acids 6a,6b and 6e were obtained (from
1a and 1c,1b and 1d, and 1e and 1g, respectively) in yields fall-
ing within the range 59–95% (Table 3). Compound 1f gave the
expected product 6f in only very low yield (9%), the major
product being oxazolone 5f (79%); paths B2 and B1 of Scheme
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 0 4 – 3 8 1 0
3806

View Article Online
that at the N-terminus, possibly because the bulky benzyl
groups at the α-carbon atom induce the conformation that pro-
vides the proximity required for internal nucleophilic attack.
This finding also corroborates the above mentioned competi-
tion, as the oxazolone obtained in 5 minute reactions with a
yield of about 42% (Table 3, footnote to entry 1h) could not
result solely from cleavage of amide 7h (Path A, Scheme 3),
which requires a much longer reaction time, but would be the
product of cleavage of the N-alkyl group of oxazolonium 2
(Path B). Comparing with what was observed in the case
of reactions in diluted TFA, in neat TFA the N-alkyl group
of the bulkier N-benzoyl derivatives cleaves faster than their
amide bond, which seems to indicate that the composition
of the products of these reactions is kinetically controlled by
the bulkiness of the substituents at R1 and R4, and by the
concentration of acid.
typical stretching peaks for C᎐O and C᎐N bonds at 1815 and
᎐ ᎐
1655 cmϪ1, respectively; in addition, a high resolution electron
impact mass spectrum showed a good agreement with the
molecular weight expected for this compound. In the 1H NMR
spectra of some of the above α,α-dibenzylglycine derivatives
recorded at 25 ЊC, the AB-type pattern for the methylene groups
bonded to the α-carbon atom was not symmetrical as one
would expect, but distorted with the doublet at higher field well
resolved and the other very broad; nevertheless, when these
spectra were inspected at 50 ЊC the broad doublet sharpened
and the AB-type pattern became almost symmetrical (Fig. 2).
This was observed with all N-acyl-N-(4-methoxybenzyl)-α,α-
dibenzyl glycine amides 1e–1h both in CDCl3 and DMSO-d6
but not with the oxazolone 5f or the amide 7h (Fig. 1), which
showed always a symmetric pattern; the fully cleaved N-acyl-
amino acids 6e and 6f also showed well resolved symmetric
patterns in both solvents. However, for 3e and 3f, which are the
partially cleaved analogues of compounds 6 above having a
4-methybenzyl group bonded to the nitrogen atom, the AB-type
pattern was asymmetric in CDCl3 and symmetric in the more
polar DMSO-d6 (Fig. 2), the separation of the two doublets
falling within the range indicated above for compounds 1e–1h.
These effects suggest slow rotamer interconversion and the
compounds where they were observed were those and only
those having a substituent at the nitrogen atom of the
α,α-dibenzyl glycine residue; somehow this correlates with the
deviating behaviour towards acidic cleavage found with some of
these compounds.
Spectroscopic characterisation of oxazolone 5f and other
ꢀ,ꢀ-dibenzyl glycine derivatives.
The α-carbon atom of symmetrical α,α-dialkyl glycines and
their derivatives is always prochiral10 and, thus, the two hydro-
gen atoms of each of the benzyl methylene groups are non-
equivalent; as a result, separate absorption peaks for each of
the two protons of either methylene group should be expected
to be observed in the 1H NMR spectra of these compounds. In
practice, this results in two distorted doublets, one for each
proton, with a pattern resembling an AB quartet. In DMSO-d6
the two doublets are separated by 0.31 to 0.40 ppm in all
N-(4-methoxybenzyl)-α,α-dibenzyl glycine derivatives 1e–1h,
close to 0.5 ppm in the N-acyl-α,α-dibenzyl glycines 6e and 6f
and as much as 0.80 ppm in the N-benzoyl-α,α-dibenzylglycine
1
cyclohexylamide 7h (Fig. 1). However, the H NMR spectrum
of oxazolone 5f shows the methylene proton absorptions as 4
peaks shaped as a very narrow AB-type pattern in which the
absorption lines are separated by only as little as 0.02 ppm
(Fig. 1). This almost isochronous behaviour seems to be typical
of α,α-dibenzylglycine oxazolones, was reported for the first
time thirty years ago4 and is a reliable means of identifying
these cyclic species. In all compounds, including 5f, the coup-
ling constant lay within Ϫ11.0 and Ϫ13.5 Hz, which is typical
of a vicinal coupling. The IR spectrum of oxazolone 5f showed
Fig. 2 Partial 1H NMR spectrum of 1f run in DMSO-d6 at 25 ЊC
showing an asymmetric AB-type pattern for the benzyl methylene
protons, which sharpens and becomes almost symmetric at 50 ЊC,
together with that of 3f in CDCl3 at 25 ЊC showing an also asymmetric
AB-type pattern, which is almost symmetric in DMSO-d6.
Conclusions
All α,α-dimethyl glycine Ugi–Passerini adducts exhibited a con-
sistent pattern of behaviour according to which treatment with
diluted TFA for several hours only cleaves their C-terminal
amide bond, while boiling with neat TFA for 5 to 10 minutes
cleaves the amide bond and also the N-alkyl group. A similar
behaviour was observed with the N-phenylacetyl derivatives of
α,α-dibenzyl glycine, but the N-benzoyl analogues behaved dif-
ferently; with diluted TFA selective amide cleavage with regard
to the N-alkyl group could not be fully accomplished, as the
major product was a very stable oxazolone. To our knowledge
such cyclic compound had never been obtained under these
Fig. 1 Partial 1H NMR spectra of 1h,5f,6f and 7h showing typical
AB-type patterns for the benzyl methylene protons. The spectrum of 1h
was recorded at 50 ЊC, while the others were run at 25 ЊC.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 0 4 – 3 8 1 0
3807

View Article Online
circumstances and conditions; its formation supports Good-
man’s proposal6 that amide cleavage is preceded by a cyclic
intermediate. With neat TFA the same compounds behaved
again differently, the bulkier yielding two major products, viz,
oxazolone and N-acylamino acid amide, both of them exhibit-
ing no 4-methoxybenzyl group at the nitrogen atom of the of
α,α-dibenzyl glycine residue. The fact that the latter compound
could also undergo amide cleavage on boiling with neat TFA
under forcing conditions to give the oxazolone as the only
product leads us to conclude also that (i) no N-alkyl group is
necessary to assist selective amide cleavage within this com-
pound, although for such purpose long reaction times are
required, (ii) the oxazolone formed in the 5 minute reactions did
not arise from the above amide but must have resulted directly
from the Ugi–Passerini adduct, which indicates the occurrence
of two competitive reactions, (iii) the apparently facilitated loss
of the N-alkyl group from both above products suggests assist-
ance by the two bulky benzyl groups at the α-carbon atom and
(iv) formation of oxazolone under these circumstances also
suggests assistance by the same groups in generating the con-
formation required for internal nucleophilic attack, which in
the α,α-dimethyl glycine series seemed to require assistance also
by the N-alkyl group. This was corroborated by the evidence of
slow conformational interconversion observed in the NMR
spectra of the N,α,α-tribenzyl glycine derivatives, which seems
to be related to the presence of the methoxybenzyl group at the
amino acid nitrogen atom.
trated to a small volume (50 ml) and set aside overnight at
room temperature. The precipitate thus formed was collected by
filtration and recrystallised from methanol-diethyl ether.
N-Phenylacetyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dimethylglycine 4-
methoxybenzylamide 1a. The reaction was carried out on a
0.084 molar scale and 1a (18.5 g, 48%) was isolated as a white
solid, mp 119.4–119.6 ЊC. Anal. Found: C, 72.91; H, 7.07; N,
6.06. Calc. for C28H32N2O4: C, 73.02; H, 7.00; N, 6.08%.
N-Benzoyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dimethylglycine 4-meth-
oxybenzylamide 1b. The reaction was carried out on a 0.02
molar scale and 1b (6.90 g, 77%) was isolated as a white solid,
mp 64.4–65 ЊC. Anal. Found: C, 72.34; H, 6.82; N, 6.23. Calc.
for C27H30N2O4: C, 72.62; H, 6.77; N, 6.27%.
N-Phenylacetyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dimethylglycine
cyclohexylamide 1c. The reaction was carried out on a 0.06
molar scale and 1c (19.28 g, 76%) was isolated as a white solid,
mp 168.4–169.8 ЊC. Anal. Found: C, 73.92; H, 7.91; N, 6.70.
Calc. for C26H34N2O3: C, 73.90; H, 8.11; N, 6.63%.
N-Benzoyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dimethylglycine cyclo-
hexylamide 1d. The reaction was carried out on a 0.03 molar
scale and 1d (10.19 g, 83%) was isolated as a white solid, mp
140.3–140.8 ЊC. Anal. Found: C, 73.09; H, 7.68; N, 6.81. Calc.
for C25H32N2O3: C, 73.50; H, 7.89; N, 6.86%.
General method 2: synthesis of Ugi–Passerini adducts
(for compounds 1e–h)
Experimental
Freshly distilled N-(4-methoxybenzyl)-1,3-diphenyl-2-propan-
imine and the acid (0.055 mol) were added to a flask contain-
ing dry methanol (20 ml). After stirring at room temperature
for 10 min to dissolve the acid, the isonitrile (0.055 mol) was
added. The reaction mixture was stirred in the dark at room
temperature under nitrogen for two to three weeks, after which
the solvent was evaporated and the product chromatographed
with silica, using the following eluents: DCM–hexane (1 : 2),
DCM–hexane (1 : 1), neat DCM or DCM–methanol (50 : 1).
The pure compounds were obtained by evaporation of the
combined fractions.
N-(4-Methoxybenzyl)-1,3-diphenyl-2-propanimine and 4-meth-
oxybenzyl isonitrile were prepared as described elsewhere.9
Acetone was freshly distilled after drying over CaCl2. Methanol
and toluene were dried by standard procedures. All other sol-
vents and reagents, including cyclohexyl isonitrile, were used as
obtained from commercial sources. TLC analyses were carried
out on 0.25 mm thick pre-coated silica plates (Merck Fertig-
platten Kieselgel 60F254) and spots were visualised under UV
light or by exposure to vaporised iodine. Preparative chromato-
graphy was carried out on Merck Kieselgel 60 (230–400 mesh).
All melting points were measured on a Gallenkamp melting
1
point apparatus and are uncorrected. H NMR Spectra were
N-Phenylacetyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dibenzylglycine 4-
methoxybenzyl amide 1e. The reaction was carried out on a
0.0455 molar scale and 1e (12.56 g, 45%) was isolated as a white
solid, mp 90.4–91.0 ЊC. Anal. Found: C, 78.18; H, 6.51; N, 4.66.
Calc. for C40H40N2O4: C, 78.41; H, 6.58; N, 4.57%.
recorded at 25 ЊC in ∼5% CDCl3 or DMSO-d6 solution on
a Varian 300 Unity Plus spectrometer; all shifts are given in
δ ppm using δH Me4Si = 0 and J-values are given in Hz, and
assignments were made by comparison of chemical shifts, peak
multiplicity and J-values. 13C NMR Spectra were recorded with
the same instrument at 75.4 MHz and using the solvent peak as
internal reference; assignments were carried out by the DEPT
135, HMBC and/or HMQC techniques. IR Spectra were run on
a FTIR Perkin-Elmer 1600 spectrophotometer. Elemental
analyses were carried out on a Leco CHNS 932 instrument.
Comparison of compounds with genuine samples was carried
out by mixed melting points.
N-Benzoyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dibenzylglycine 4-meth-
oxybenzyl amide 1f. The reaction was carried out on a 0.0455
molar scale and 1f (16.0 g, 59%) was isolated as a white solid,
mp 100.8–101.1 ЊC. Anal. Found: C, 78.20; H, 6.44; N, 4.77.
Calc. for C39H38N2O4: C, 78.24; H, 6.40; N, 4.68%.
N-Phenylacetyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dibenzylglycine
cyclohexyl amide 1g. The reaction was carried out on a 0.039
molar scale and 1g (17.95 g, 80%) was isolated as a white solid,
mp 87.3–87.9 ЊC. Anal. Found: C, 79.37; H, 7.63; N, 5.04. Calc.
for C38H42N2O3: C, 79.41; H, 7.37; N, 4.87%.
General method 1: synthesis of Ugi–Passerini adducts
(for compounds 1a–d)
Freshly distilled acetone (20 ml) was added to a flask, followed
by anhydrous sodium sulfate (1 g) and 4-methoxybenzylamine
(4.12 g, 0.03 mol). After stirring at room temperature for
15 min, the acid (0.03 mol) was added, which caused a large
amount of a white solid to separate. After stirring for 10 min,
the isonitrile (0.03 mol) was added. The reaction mixture was
stirred in the dark at room temperature under nitrogen and
monitored every day by TLC. When no more starting material
could be observed, which required two to three weeks depend-
ing on the nature of the reagents, the solvent was evaporated
and the residue taken up in methanol (200 ml). The mixture was
filtered to separate the sodium sulfate and the filtrate concen-
N-Benzoyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dibenzylglycine cyclo-
hexyl amide 1h. The reaction was carried out on a 0.045 molar
scale and 1h (12.0 g, 48%) was isolated as a white solid, mp
100.9–101.5 ЊC. Anal. Found: C, 78.85; H, 7.07; N, 4.99. Calc.
for C37H40N2O3: C, 79.25; H, 7.19; N, 5.00%.
General method 3: cleavage of Ugi–Passerini adducts with
diluted TFA
The Ugi–Passerini adduct 1 was dissolved in dry aceto-
nitrile, which was followed by addition of the amount of TFA
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 0 4 – 3 8 1 0
3808

View Article Online
necessary to give a 0.02 M solution of substrate containing the
acid at a concentration of 2%. This was kept at room temper-
ature until TLC (eluent: dichloromethane–MeOH, 9 : 1)
showed no more starting material (1.5–36 hours, depending on
the substrate). The solvent was then evaporated at 30 ЊC and the
residue taken up in 6 M NaOH (10 ml). In the case of com-
pounds 1a–1g about 30 min later the pH was adjusted to 1.
Then, dichloromethane (DCM) (20 ml) was added and the
aqueous layer extracted with DCM (2 × 15 ml). The combined
organic layers were washed with brine (3 × 10 ml), and dried
over anhydrous MgSO4. This was removed by filtration and the
filtrate concentrated under vacuum. The residue thus obtained
was purified by column chromatography (DCM–MeOH, 50 : 1)
to yield N-acylamino acids 3 by evaporation of the corre-
sponding fractions. For compounds 1f and 1h the pH was first
adjusted to 7 and the aqueous phase extracted with DCM; the
combined extracts were dried over magnesium sulfate and con-
centrated to yield oxazolone 5f. The pH of the aqueous layer
was then adjusted to 1 and the work-up continued as described
above for the other compounds.
N-Phenylacetyl-ꢀ,ꢀ-dimethylglycine 6a (from 1a and 1c).
Adduct 1a (1.0 g, 2.2 mmol) gave the N-protected amino acid
6a (0.32 g, 66%) as a white solid, mp 190.3–191.0 ЊC. Anal.
Found: C, 65.17, H, 6.76, N, 6.45. Calc. for C12H15NO3: C,
65.14, H, 6.83, N, 6.33%. Compound 6a (0.38 g, 72%) was also
obtained from 1c (1.1 g, 2.4 mmol).
N-Benzoyl-ꢀ,ꢀ-dimethylglycine 6b (from 1b and 1d). Adduct
1b (1.0 g, 2.2 mmol) gave the N-protected amino acid 6b (0.27 g,
59%) as a white solid, mp 197.3–198.0 ЊC. Anal. Found: C,
63.58; H, 6.29; N, 6.79. Calc. for C11H13NO3: C, 63.76; H, 6.32;
N, 6.76%. Compound 6b (0.35 g, 70%) was also obtained from
1d (1.0 g, 2.45 mmol).
N-Phenylacetyl-ꢀ,ꢀ-dibenzylglycine 6e (from 1e and 1g).
Adduct 1e (1.0 g, 1.6 mmol) gave the N-protected amino acid 6e
(0.58 g, 95%) as a white solid, mp 210.6–211.6 ЊC. HRMS (EI)
calcd. for C24H23NO3: 373.1678; found: 373.1676. Compound
6e (0.517 g, 80%) was also obtained from 1g (1.0 g, 1.7 mmol).
N-Benzoyl-ꢀ,ꢀ-dibenzylglycine 6f (from 1f). Adduct 1f (1.0 g,
1.7 mmol) gave the N-protected amino acid 6f (0.054 g, 9%)
as a white solid, mp 208.6–209.2 ЊC. HRMS (EI) calcd. for
C23H21NO3: 359.1521; found: 359.1506. However, oxazolone 5f
(0.45 g, 79%) was the major product, comparing well with a
genuine sample as described below.
N-Phenylacetyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dimethylglycine 3a
(from 1a and 1c). Adduct 1a (0.92 g, 2.0 mmol) gave compound
3a (0.60 g, 88%) as a white solid, mp 168.6–169.2 ЊC. Anal.
Found: C, 69.75; H, 6.82; N, 4.06. Calc. for C20H23NO4ؒ¼H2O:
C, 69.45; H, 6.85; N, 4.05%. Compound 3a (0.51 g, 75%) was
also obtained from 1c (0.85 g, 2.0 mmol).
N-Benzoyl-ꢀ,ꢀ-dibenzylglycine cyclohexylamide 7h (from 1h).
In a 10 minute reaction adduct 1h (1.0 g, 1.8 mmol) gave com-
pound 7h (0.35 g, 44%) as a white solid, mp 205.9–206.8ЊC.
Anal. Found: C, 77.74, H, 7.32, N, 6.33. Calc. for C29H32N2O2ؒ
½ H2O: C, 77.47, H, 7.40, N, 6.23%. Oxazolone 5f (0.258 g,
40%) was also obtained together with a small amount of com-
pound 6f (0.077 g, 12%), both comparing well with genuine
samples as they are described below. In a 5 minute repeat of this
reaction no free acid 6f was isolated, while compounds 5f and
7h were obtained in slightly better yields than above (42 and
52%, respectively).
N-Benzoyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dimethylglycine
3b
(from 1b and 1d). Adduct 1b (0.89 g, 2.0 mmol) gave compound
3b (0.55 g, 84%) as a white solid, mp 154.9–155.8 ЊC. HRMS
(EI) calcd. for C19H21NO4: 327.1471; found: 327.1460. Com-
pound 3b (0.53 g, 81%) was also obtained from 1d (0.92 g, 2.0
mmol).
N-Phenylacetyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dibenzylglycine 3e
(from 1e and 1g). Adduct 1e (1.23 g, 2.0 mmol) gave compound
3e (0.59 g, 60%) as a white solid, mp 158.2–159.2 ЊC. Anal.
Found: C, 77.43; H, 6.47; N, 2.95. Calc. for C32H31NO4: C,
77.87; H, 6.33; N, 2.84%. Compound 3e (0.69 g, 70%) was also
obtained from 1g (1.15 g, 2.0 mmol).
Other preparations
2-Phenyl-4,4-dibenzyl-1,3(4H)-oxazol-5-one 5f
N-Benzoyl-N-(4-methoxybenzyl)-ꢀ,ꢀ-dibenzylglycine 3f (from
1f and 1h). Adduct 1f (120 mg, 0.2 mmol) gave compound 3f (45
mg, 47%) as a white solid, mp 154.6–155.4 ЊC. Anal. Found: C,
74.67; H, 6.47; N, 2.75. Calc. for C31H29NO4: C, 74.83; H, 6.28;
N, 2.82%. A large amount of oxazolone 5f (30 mg, 44%)
was also obtained, comparing well with a genuine sample as
described below. Compound 1h (112 mg, 0.2 mmol) gave only a
small amount of 3f (14 mg, 15%), the major product being
oxazolone 5f (52 mg, 76%).
From 7h with neat TFA. Amide 7h (0.11 g, 0.25 mmol) was
refluxed with neat TFA for 2 hours after which the residue from
evaporation of the solvent was taken up in 6 M NaOH (10 ml).
Then, oxazolone 5f (0.061 g, 72%) was isolated by the pro-
cedure described in general method 4 above as a white solid, mp
118.1–119.1 ЊC. HRMS (EI) calcd. for C23H19NO2: 341.1416;
found: 341.1413.
From 1f and 1h with 2% TFA. Cf. preparation of 3f above by
general method 3.
General method 4: cleavage of Ugi–Passerini adducts with neat
TFA
From 1f and 1h with neat TFA. Cf. preparation of 6f above by
The Ugi–Passerini products 1 (1 g) were dissolved in neat TFA
(5 ml) and refluxed for 10 min, which made the colour of the
solution turn red. The acid in excess was evaporated and the
residue taken up in 6 M NaOH. After stirring for 30 min, the
white solid formed was filtered off and the filtrate extracted
twice with Et2O. In the case of compounds 1a–1g the pH of the
aqueous layer was adjusted to 1 to precipitate the corre-
sponding product 6, which was filtered off, washed with water
and dried. For compounds 1f and 1h DCM was used instead of
Et2O and the pH first adjusted to 7. The aqueous phase was
extracted with DCM, the combined extracts being dried over
magnesium sulfate and concentrated. The residue was purified
by column chromatography to yield compounds 5 and/or 7.
Then, the pH of the aqueous layer was adjusted to 1 and the
work-up continued as described above for the other
compounds.
general method 4.
N-Benzoyl-N-ꢀ,ꢀ-dibenzylglycine 6f (from 5f)
Oxazolone 5f (100 mg, 0.29 mmol) was suspended in 10 ml of 6
M HCl and stirred at 100 ЊC for 4 h; the precipitate thus formed
was separated by filtration, washed well with water and dried to
give the open-chain product 6f (98 mg, 94%), which compared
well with a genuine sample as described above.
Acknowledgements
We thank the Foundation for Science and Technology (Portu-
gal) for financial support to the Institute of Biotechnology and
Fine Chemistry (University of Minho) and for scholarship
no. SFRH/BPD/1544/2000 to W.-Q J.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 0 4 – 3 8 1 0
3809

View Article Online
4 H. L. Maia, B. Ridge and H. N. Rydon, J. Chem. Soc., 1973, 98.
5 C. Toniolo, Janssen Chem. Acta, 1993, 11, 28.
References
1 C. Toniolo, Janssen Chim. Acta, 1993, 11, 10 (review article); J.-P.
Mazeleyrat, K. Wright, M. Wakselman, F. Fromaggio, M. Crisma
and C. Toniolo, Eur. J. Org. Chem., 2001, 1821.
2 V. De Filippis, F. De Antoni, M. Frigo, P. P. Laureto and
A. Fontana, Biochemistry, 1998, 37, 1686; H. Medzihradszky-
Schweiger, K. Medzihradszky, H. Nádasi and H. Suli-Vargha, in
Peptides 1998, Proceedings of the 25th European Peptide Symposium,
eds. S. Bajusz and F. Hudecz, Akadémiai Kiadó, Budapest, 1999,
p. 6063.
6 C. J. Creighton, T. T. Romoff, J. H. Bu and M. Goodman, J. Am.
Chem. Soc., 1999, 121, 6786.
7 T. Johnson, M. Quibell, D. Owen and R. C. Sheppard, J. Chem. Soc.,
Chem. Commun., 1992, 1573.
8 G. Gokel, P. Hoffmann, H. Kleimann, H. Klusacek, G. Ludke,
D. Marquaerding and I. Ugi, in Isonitrile Chemistry, ed. I. Ugi,
Academic Press, New York and London, 1971, p. 201.
9 S. P. G. Costa, H. L. S. Maia and S. M. M. A. Pereira-Lima,
Org. Biomol. Chem., 2003, 1, 1475.
3 D. S. Jones, G. W. Kenner, J. Preston and R. C. Sheppard, J. Chem.
Soc., 1965, 6227.
10 Hors Friebolin, Basic One- and Two-Dimensional NMR
Spectroscopy, 3rd edn., Wiley-VCH, Weinheim, 1998, p. 77.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 3 8 0 4 – 3 8 1 0
3810