Peptide backbone folding induced by the C(alpha)-tetrasubstituted cyclic alpha-amino acids 4-amino-1,2-dithiolane-4-carboxylic acid (Adt) and 1-aminocyclopentane-1-carboxylic acid (Ac5c). A joint computational and experimental study.

Article abstract of DOI:10.1039/b212247b

The conformational study of a new group of synthetic peptides containing 4-amino-1,2-dithiolane-4-carboxylic acid (Adt), a cysteine-related achiral residue, has been carried out through a joint application of computational and experimental methodologies. Molecular Dynamics simulations clearly suggest the tendency of this molecule to adopt a gamma-turn conformation in vacuum and help in analyzing the complex and crucial conformational behaviour of the dithiolane ring which appears to preferentially adopt a C(S)-like structure. Electronic structure calculations carried out in solution using the Density Functional Theory also indicate the preservation of the gamma-like folding in apolar solvents and the helix-like one in more polar solvents. A comparison with the achiral 1-aminocycloalkane-1-carboxylic acid (Ac5c) has been carried out using the same computational tools. NMR and IR data on dipeptide derivatives containing the Adt or Ac5c residue show that in chloroform solution all the models prefer a gamma-turn structure, centered at the cyclic residue, stabilized by an intramolecular H-bond, whereas in a more polar solvent, i.e. dimethyl sulfoxide, this folding is not maintained. The experimental conformational studies, extended to N-Boc protected tripeptides, clearly indicate the remarkable tendency of both the five-membered C(alpha)-tetrasubstituted cyclic amino acids Adt and Ac5c to induce the gamma-turn structure also in models able to adopt the beta-bend conformation.

Full text of DOI:10.1039/b212247b

View Article Online / Journal Homepage / Table of Contents for this issue
Peptide backbone folding induced by the C ꢀ-tetrasubstituted cyclic
ꢀ-amino acids 4-amino-1,2-dithiolane-4-carboxylic acid (Adt)
and 1-aminocyclopentane-1-carboxylic acid (Ac5c).
A joint computational and experimental study
Massimiliano Aschi,a Gino Lucente,*b,c Fernando Mazza,a Adriano Mollica,b Enrico Morera,b,c
Marianna Nallib and Mario Paglialunga Paradisib,c
a Dipartimento di Chimica, Ingegneria Chimica e Materiali - Università di L’Aquila,
I-67100 Coppito, L’Aquila, Italy
b Dipartimento di Studi Farmaceutici, Università degli Studi di Roma “La Sapienza”,
P. le A. Moro 5, I-00185 Roma, Italy
c Istituto di Chimica Biomolecolare, CNR, Sezione di Roma,
c/o Dipartimento di Studi Farmaceutici, Università di Roma “La Sapienza”, P. le A. Moro 5,
I-00185 Roma, Italy
Received 16th December 2002, Accepted 10th April 2003
First published as an Advance Article on the web 25th April 2003
The conformational study of a new group of synthetic peptides containing 4-amino-1,2-dithiolane-4-carboxylic
acid (Adt), a cysteine-related achiral residue, has been carried out through a joint application of computational
and experimental methodologies. Molecular Dynamics simulations clearly suggest the tendency of this molecule to
adopt a γ-turn conformation in vacuum and help in analyzing the complex and crucial conformational behaviour of
the dithiolane ring which appears to preferentially adopt a Cs-like structure. Electronic structure calculations carried
out in solution using the Density Functional Theory also indicate the preservation of the γ-like folding in apolar
solvents and the helix-like one in more polar solvents. A comparison with the achiral 1-aminocycloalkane-1-
carboxylic acid (Ac5c) has been carried out using the same computational tools. NMR and IR data on dipeptide
derivatives containing the Adt or Ac5c residue show that in chloroform solution all the models prefer a γ-turn
structure, centered at the cyclic residue, stabilized by an intramolecular H-bond, whereas in a more polar solvent,
i.e. dimethyl sulfoxide, this folding is not maintained. The experimental conformational studies, extended to N-Boc
protected tripeptides, clearly indicate the remarkable tendency of both the five-membered Cα-tetrasubstituted cyclic
amino acids Adt and Ac5c to induce the γ-turn structure also in models able to adopt the β-bend conformation.
concerning a γ-turn conformation induced by two six-mem-
Introduction
bered Cα-tetrasubstituted cyclic amino acids, namely: 4-amino-
tetrahydrothiopyran-4-carboxylic acid (Thp), containing a
sulfur atom in the ring, and the related cycloaliphatic model
1-aminocyclohexane-1-carboxylic acid (Ac6c). Computational
observations carried out on the model Ac-Thp-NHMe showed
that a γ-turn represents the lowest minimum; at the same time
1H-NMR and IR studies, performed on the dipeptide deriv-
atives HCO-Thp-Leu-OMe and HCO-Ac6c-Leu-OMe, showed
that such a conformation is retained in CDCl3 whereas it is lost
in the more polar (CD3)2SO solvent.
The introduction of Cα-tetrasubstituted α-amino acid residues
into peptide backbones in order to generate models character-
ized by reduced conformational flexibility, high tendency to
adopt predetermined secondary structures, and enhanced
metabolic stability, represents a versatile and well established
approach.1–8 When the biochemical relevance of peptides is
considered together with their potential role as therapeutic
agents, the importance of an efficient methodology, able to
influence the peptide conformational preferences, can be prop-
erly evaluated. Based on this approach, potent and selective
receptor ligands have been designed and valuable information
on the conformation–bioactivity relationships continues to be
obtained.9–14
The γ-turn conformation involves three consecutive
amino acid residues and can be stabilized by an intramolecular
H-bond between the CO of the first residue (i) and the NH of
the third residue (iϩ2) forming a seven-membered ring. As
shown in Fig. 1, γ-turn and inverse γ-turn are characterized by
different ꢀiϩ1 and ψiϩ1 torsion angles. Although less frequent
than β-turn, the γ-turn secondary structure is thought to play
important roles in molecular recognition,23–27 and the presence
of γ-turn in solution of several bioactive peptides has been
proposed.28
As a part of our continuing interest in the field of hetero-
cyclic α-amino acids containing a quaternary α-carbon atom
we started recently19,20 a study on a new group of synthetic
peptides containing 4-amino-1,2-dithiolane-4-carboxylic acid
(Adt), a cysteine-related achiral residue, characterized by the
presence of a five-membered ring in which a disulfide function
is contained. Although 1,2-dithiolanes can be easily obtained
by mild oxidation from the corresponding 1,3-dithiols, their
chemical reactivity is clearly distinct from that of linear and
Both theoretical and experimental studies on this subject
have been published and, in addition to amino acid residues
containing two linear side chains bound at the Cα carbon atom,
the group of Cα-tetrasubstituted residues in which the quater-
nary α-carbon forms part of a ring has been the object of exten-
sive investigation,15–18 with particular reference to the achiral
1-aminocycloalkane-1-carboxylic acids (Acnc, where n indicates
the number of carbon atoms in the cycloalkane ring).3 Interest
in the cyclic congeners in which one or more heteroatoms form
part of the ring is relatively more recent and derives from the
potentiality of these amino acid residues to combine, when
inserted into the peptide backbone, the tendency to stabilize
specifically folded conformers with the establishment of
intra- or inter-molecular interactions through the side chain
functionality.19–21 In this field we reported recently22 a study
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
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
1980

View Article Online
Two trajectories, propagated at constant energy and angular
momentum, were initiated with a Maxwellian distribution of
the velocities corresponding to a starting temperature of 298 K.
The first trajectory was started assuming for Ac-Adt-NHMe
a conformation close to a helical structure with ꢀ = 61Њ and
ψ = 16Њ; an almost opposite conformation with ꢀ = Ϫ60Њ and
ψ = Ϫ20Њ was the starting point of the second trajectory. Both
the starting structures correspond to high-energy conform-
ations having the CSSC torsion angle as large as 24Њ. This
choice ensures, during the energy-constant simulation, a suffi-
cient amount of internal energy allowing one to span the
ꢀ/ψ and the CSSC conformational space as much as possible.
The obtained trajectories were initially analyzed in terms of
their root mean square deviation (rmsd) with respect to the
starting structures. Such analysis provides the primary tool for
assessing when and how the system under investigation reaches
the actual equilibrium. Both runs revealed equilibration not
before 9.0 ps. After such a time the rmsd reaches a constant
value. Therefore the portion of the trajectory before such a
point is disregarded in the following analysis. Fig. 2 shows the
distribution of the CSSC torsion angle sampled in the last 9 ps
of each trajectory. It is evident that at the dynamic equilibrium,
the Adt ring adopts a CSSC torsion angle oscillating around 0Њ
and that the ring conformations displacing from this value play
no significant role.
Fig.
1
Schematic representation of γ-turn and inverse γ-turn
conformations.
larger ring 1,2-disulfides and represents the key structural
feature on which the activity of relevant biomolecules, such as
lipoic acid, is based;29 thus, the study of the properties of syn-
thetic peptides containing this residue appears of general inter-
est. Here we report a joint computational and experimental
study on the property as γ-turn inducer of the cyclic Cα-tetra-
substituted α-amino acid Adt and of the structurally related
Ac5c which contains, as compared with the Adt residue, a
–CH2–CH2– unit in place of the disulfide function.
In the first part of the work an exhaustive computational
analysis of the energetic and structural features of Adt and
Ac5c is addressed both in vacuum and in solution; a combined
application of classical Molecular Dynamics (MD) simulations
and quantum-chemical calculations is used for this purpose. In
the second part of the work a solution conformational analysis
of Adt- and Ac5c-containing peptide models, aimed at evalu-
ating the theoretical predictions on the experimental data,
is reported.
Fig. 2 Distribution of the CSSC torsion angle along the last 9 ps of
each trajectory.
Results and discussion
Both the trajectories were then projected on the ꢀ–ψ map,
generating a ‘dynamic’ Ramachandran map shown in Fig. 3.
The Ac-Adt-NHMe preferentially adopts the γ-turn conform-
ations. Moreover, it should be remarked that the two γ-turns
never interconvert, i.e. the two trajectories never appear to
1. Dynamic Reaction Path calculations
A preliminary analysis of the conformational features of
the Cα-tetrasubstituted achiral residue Adt in vacuum was
performed through MD simulations. In the model examined
(Ac-Adt-NHMe) the N and C ends are protected by the
CH3CO and NHMe groups, respectively.
The MD simulation represents a valuable tool for exploring
the properties of the potential energy hypersurface of a system
showing a relatively high number of internal degrees of free-
dom such as the Ac-Adt-NHMe molecule here under study. In
particular the aim is to evaluate the preferred conformation of
the dithiolane ring, together with the ꢀ and ψ values, adopted
by the Ac-Adt-NHMe molecule at the equilibrium. To this end
we adopted the DRC30 approach which produces a classical
trajectory by calculating, step by step, the potential energy and
its analytical gradient through quantum chemical procedures.
In the present case, the semiempirical AM1 Hamiltonian31 was
adopted for energy and gradient calculations. It must be
remarked that, unlike the classical MD simulations with pre-
defined force fields, the DRC simulations which build the forces
acting on each atom using quantum mechanics, are compu-
tationally rather demanding and hence can hardly be extended
for times longer than 20–30 ps.
Fig. 3 Dynamic Ramachandran plot reported as count of the ꢀ and ψ
values sampled along the last 9 ps of the trajectories.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1981

View Article Online
Table 1 Potential energy differences (E) including Zero Point Energy and selected geometrical parameters for the most relevant minima in vacuum
of the Ac-Adt-NHMe and Ac-Ac5c-NHMe species calculated at the B3LYP/6-31G(d)//B3LYP/4-31G level of theory
Ac-Adt-NHMe
Species
Type
E/kJ molϪ1
ꢀ/Њ
ψ/Њ
55.3
Ϫ55.7
Ϫ162.5
Ϫ20.5
20.5
NH–O/Å
N–H–O/Њ
CSSC/Њ
τ/Њ
I
γ
iγ
f.e.
αR
αL
0
0
4.6
6.7
6.7
Ϫ72.8
72.3
179.3
Ϫ70.0
70.1
1.800
1.802
1.960
—
152.3
152.3
114.9
—
5.3
Ϫ6.3
2.6
4.4
Ϫ4.4
113.5
113.5
103.5
113.1
113.1
II
III
IV
V
—
—
Ac-Ac5c-NHMe
Species
Type
E/kJ molϪ1
ꢀ/Њ
ψ/Њ
NH–O/Å
N–H–O/Њ
CCCC/Њ
τ/Њ
I
γ
iγ
f.e.
αR
αL
0
74.2
Ϫ72.8
178.9
Ϫ73.9
73.8
Ϫ60.9
55.4
177.6
Ϫ14.2
14.2
1.870
1.849
1.989
—
149.4
151.8
113.8
—
Ϫ9.2
Ϫ12.4
39.6
Ϫ1.9
1.9
111.3
112.3
104.0
112.3
113.4
II
III
IV
V
0.4
3.3
6.3
6.3
—
—
mutually superimpose. This could be ascribed to the presence
of a relatively large barrier created by the overlapping involving
the carbonyl oxygen of the acetyl group and the amide hydro-
gen atom of the terminal NHMe.
In conclusion, this preliminary analysis reveals for Ac-Adt-
NHMe in vacuum two deep not mutually interconverting min-
ima, corresponding to the γ-turn and inverse γ-turn conform-
ations both adopting an almost cis-planar CSSC torsion angle.
energetic features are not reported. Moreover, the f.e. structure
appears as the third minimum despite a valence angle as large as
103.5Њ, significantly distorted from the tetrahedral value.
The results referring to Ac-Ac5c-NHMe are rather similar to
those obtained for Ac-Adt-NHMe. At variance with the Adt
model, the f.e. conformer presents a more puckered penta-
atomic ring with the torsion angle, opposite to the Cα atom, of
ϩ39.6Њ. The conformer population is finally evaluated for both
Ac-Ac5c-NHMe and Ac-Adt-NHMe in the ideal gas approx-
imation using the free energy differences reported in Table 2.
The results are reported in Fig. 4 and show that, at the DFT
level, at 298 K, the γ-turn conformers are the most populated.
This should be entirely ascribed to the enthalpic gain due to the
intramolecular H-bond, since the entropy contribution remains
basically the same for all the conformers.
2. DFT calculations in vacuum
The information on the Born–Oppenheimer surface previously
obtained by the MD run was then used as a starting point for
addressing the energetic and structural features of Adt at a
more sophisticated level of theory. Thus, the Density Func-
tional Theory (DFT), which has been shown to be able to
describe the subtle effects governing the conformational equi-
libria in peptides,32 was applied. In particular, on the basis of
the MD run, we re-sampled in vacuum the preferred regions of
the potential energy surface of Adt by using the B3LYP func-
tional in conjunction with split valence basis sets (see compu-
tational details). The related results are reported in Tables 1 and
2. In the same Tables we also report analogous calculations
carried out on the Ac-Ac5c-NHMe model which refer to the
Ac5c ring. The analysis of the two sets of calculations should
reveal possible differences between the heterocyclic and the
cycloaliphatic ring systems, i.e. possible differential effects of
the nature of the ring on the conformer population.
Five deep potential energy minima, energetically close to
each other, have been found at the DFT level of theory for both
species. In the case of Ac-Adt-NHMe, in accordance with the
previous DRC results, the two lowest minima correspond to the
γ-turn and inverse γ-turn (iγ) structures, respectively. The third
one is associated with the fully extended (f.e.) conformation,
while helices (αR and αL) appear as the last two minima lying
6.7 kJ molϪ1 above the γ-turns. Such a result is paralleled by the
intramolecular H-bond formation. As can be seen in Table 1, an
H-bond is present in both the γ-turns (H–O = 1.80 Å, NH–O
angle = 152Њ) and it is also present in the f.e. conformation,
although with a longer contact (H–O = 1.96 Å) and a smaller
linear character (NH–O angle = 115Њ); no H-bonds are possible
in the two helical conformations. It is also interesting to observe
that all the above minima are associated with an almost
cis-planar CSSC torsion angle, ranging between 5.3Њ and Ϫ6.3Њ.
It must also be remarked that, as already suggested by MD
calculations, widening of the CSSC value causes an increase of
the potential energy leading to high-energy structures and to
first and higher order saddle points whose geometrical and
Fig. 4 Room temperature molar ratios in vacuum (V), chloroform (C)
and water (W) for the thermodynamically more accessible species of
Adt and Ac5c (see text) calculated using Density Functional Theory.
The roman numbers are those of Table 1.
3. DFT calculations in solution
The effect of the solvent was also addressed by carrying out the
previous DFT calculations in solution by using the polarizable
continuum model33,34 in two completely different solvents, in
order to evaluate the influence of the polarity of the medium on
the conformer relative stabilities.
The results are also reported in Tables 1 and 2. From a
structural point of view, none of the geometrical features
undergo large modifications with respect to the vacuum. For
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1982

View Article Online
Table 2 298 K Gibbs free energy differences in vacuum (∆Gv). 298 K
solvation Gibbs free energies in chloroform (∆Gc) and in water (∆Gw)
calculated at the B3LYP/6-31G(d)//B3LYP/4-31G level of theory. These
values are reported in bold type. The electrostatic, dispersion–repulsion
and cavitation contributions to the solvation free energies (see text) are
reported in italics
ascribed to an increase in the stability of γ-turn and helical
structures.
In a more polar solvent, such as water, we essentially do not
observe large differences between the two species. The helical
structures, owing to their enhanced ability to expose the polar
groups to the solvent,32 become the most stable conformations,
whereas both the γ-turns and the f.e. conformers practically
disappear at the same temperature. The presence of explicit
water molecules would probably enhance such an effect32
which, on the basis of the presently employed model, has to be
entirely ascribed to averaged bulk effects.
In conclusion, the dithiolane ring does favour the adoption
of a γ-turn-like conformation in apolar solvents where the
intramolecular hydrogen bond is the driving force for the adop-
tion of the conformations. When, on the other hand, the polar-
ity of the medium is increased, the helical structure, which
allows the polar groups to be better solvent-exposed, becomes
preferred. Interestingly, the present computational results sug-
gest that the disulfur containing ring, when compared to Ac5c,
slightly favours the adoption of a γ-turn conformation with
respect to the helical one, both in vacuum and in apolar
solvents.
Ac-Adt-NHMe
Species
I
∆Gv/kJ molϪ1
∆Gc/kJ molϪ1
∆Gw/kJ molϪ1
0
؊8.8
Ϫ14.6
Ϫ73.2
ϩ79.0
؊8.2
Ϫ14.2
Ϫ73.7
ϩ79.3
؊6.9
Ϫ14.6
Ϫ73.2
ϩ80.9
؊11.3
Ϫ19.2
Ϫ73.2
ϩ81.1
؊12.3
Ϫ20.1
Ϫ73.2
ϩ81.0
؊39.3
Ϫ53.6
Ϫ96.7
ϩ111.0
؊36.8
Ϫ51.5
Ϫ96.7
ϩ111.4
؊27.6
Ϫ41.7
Ϫ97.5
ϩ111.6
؊44.8
Ϫ61.5
Ϫ96.3
ϩ113.0
؊51.5
Ϫ67.8
Ϫ96.3
ϩ112.6
II
0
III
IV
V
؉3.1
؉6.3
؉6.3
4. Chemistry
The synthesis of dipeptides Boc-Adt-Leu-OMe (1a), Boc-Ac5c-
Leu-OMe (1b), HCO-Adt-Leu-OMe (2a), and HCO-Ac5c-Leu-
OMe (2b), and of tripeptides Boc-Adt-Leu-Ala-OMe (3a) and
Boc-Ac5c-Leu-Ala-OMe (3b), was performed according to
Scheme 1. 1a and 1b were prepared by coupling Boc-Adt-OH or
Boc-Ac5c-OH with H-Leu-OMe by the HOBt–EDC method.
The above carbodiimide procedure was also employed to obtain
the tripeptides 3a and 3b, although different strategies were
adopted: in the case of 3a the dithiolanic derivative Boc-Adt-
OH was coupled with the dipeptide H-Leu-Ala-OMe, while the
Ac5c-containing tripeptide 3b was prepared from Boc-Ac5c-
Leu-OH and H-Ala-OMe. A direct transformation of the
N-Boc derivatives 1a,b into the corresponding N-formyl
analogs 2a,b was performed by following the procedure of
Lajoie and Kraus35 by treatment with formic acid, followed by
EEDQ.
Ac-Ac5c-NHMe
Species
I
∆Gv/kJ molϪ1
∆Gc/kJ molϪ1
∆Gw/kJ molϪ1
0.4
؊8.4
Ϫ14.2
Ϫ69.1
ϩ74.9
؊9.6
Ϫ16.7
Ϫ69.3
ϩ76.4
؊13.4
Ϫ22.2
Ϫ69.1
ϩ77.9
؊14.6
Ϫ23.4
Ϫ68.6
ϩ77.4
؊7.9
؊31.0
Ϫ46.9
Ϫ90.8
ϩ106.7
؊37.2
Ϫ51.9
Ϫ90.8
ϩ105.5
؊54.4
Ϫ72.0
Ϫ90.0
ϩ107.6
؊51.1
Ϫ68.2
Ϫ90.0
ϩ107.1
؊35.6
Ϫ49.8
Ϫ92.1
ϩ106.3
II
0
III
IV
V
؉4.6
؉4.6
؉4.2
5. Solution conformation
In order to probe whether the γ-turn preference, previously
inferred through computational methods for Ac-Adt-NHMe
and Ac-Ac5c-NHMe models, is maintained in CDCl3 solution,
a 1H NMR and IR conformational analysis of compounds 1–3
was undertaken. The assignment of the NH resonances in Boc-
protected dipeptides 1a and 1b is straightforward: the Adt and
Ac5c NH groups resonate as singlets at higher field than the Leu
NH which appears in 1a as a doublet (J = 8 Hz) and in 1b as a
broad signal. The involvement of the NH groups in intra-
molecular hydrogen bonds was evaluated on the basis of the
chemical shift solvent dependence in CDCl3–(CD3)2SO mix-
tures. As shown in Figs. 5A and 5C, the Adt and the Ac5c NH
groups are significantly more sensitive to changes in solvent
composition as compared with the Leu NH group which
appears strongly solvent shielded. This behaviour clearly indi-
cates that the Leu NH is not accessible to the solvent in both
Boc dipeptides 1a and 1b and is presumably involved in an
intramolecular hydrogen bond, while the Adt and Ac5c NH
groups are solvent exposed.
Ϫ13.8
Ϫ69.9
ϩ75.8
this reason they have not been explicitely reported. On the
other hand, from the energetic point of view, the inclusion of
the bulky-solvent effect appears to sharply affect the relative
population of the conformers. Fig. 4 shows that the 298 K
conformers, poorly populated in vacuum, become thermo-
dynamically more accessible in solution. For example, in
chloroform the γ-turn remains thermodynamically populated
but, on the other hand, the helical structures increase their
stability.
In this context it can be noted that Ac-Adt-NHMe shows
some differences with respect to Ac-Ac5c-NHMe. In fact, while
the γ-turns appear as the most populated structures in chloro-
form in the case of Ac-Adt-NHMe, in the same solvent
Ac-Ac5c-NHMe preferentially adopts a helical structure (see
Fig. 4). From the collection of the free energy contributions
reported in Table 2, the systematically larger dispersion–
repulsion term does probably play a key role in such a difference
in chloroform. Interestingly, what must also be remarked is the
disappearance of the f.e. structures which can probably be
In order to confirm the above reported results, the IR spectra
in CHCl3 in the NH stretching region of 1a and 1b were exam-
ined. As is well known, a free amide NH gives rise to a sharp
absorption in the 3400–3470 cmϪ1 domain and is shifted in the
3300–3800 cmϪ1 region by hydrogen bonding.36,37 In the spec-
trum of 1a two bands are observed at 3306 and 3419 cmϪ1
,
corresponding to the H-bonded and free NH groups, respect-
ively. The ratio between the intensity of the two absorptions
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1983

View Article Online
Scheme 1 Reagents and conditions: i: HOBt, EDC, H-Leu-OMeؒHCl, DMF; ii: HCOOH, EEDQ; iii: HOBt, EDC, TEA, DMF; iv: HOBt, EDC,
H-Ala-OMe, TEA, DMF.
Table 3 Observed nuclear Overhauser effects (NOEs) in the NOESY
spectra of Boc-Adt-Leu-OMe (1a) and Boc-Ac5c-Leu-OMe (1b)a
1a
1b
Leu NH ؒ ؒ ؒ Adt NH
Leu NH ؒ ؒ ؒ Leu CαH
Leu CαH ؒ ؒ ؒ Leu CβH2
Adt NH ؒ ؒ ؒ Adt CH2
Leu NH ؒ ؒ ؒ Adt CH2
Leu CαH ؒ ؒ ؒ Leu CδH3
Leu NH ؒ ؒ ؒ Leu CβH2
m
m
s
Leu NH ؒ ؒ ؒ Ac5c NH
Leu NH ؒ ؒ ؒ Leu CαH
Leu CαH ؒ ؒ ؒ Leu CβH2
Ac5c NH ؒ ؒ ؒ Ac5c CβHA
Leu CαH ؒ ؒ ؒ Leu CδH3
Leu NH ؒ ؒ ؒ Leu CβH2
m
m
s
s
w
s
b
w and s
w
w
s
a s, strong; m, medium; w, weak. b CβHA protons (see Fig. 6) resonate at
higher field than CβHB.
Although the position of the NH absorptions cannot be
unequivocally associated with specific folded structures,12 it
should be noted that the stretching bands of intramolecularly
H-bonded Leu NH of 1a and 1b occur at frequencies very close
to those observed for the same group in the related dipeptides
HCO-Ac6c-Leu-OMe and HCO-Thp-Leu-OMe,22 which adopt
the γ-turn conformation.
All these findings, which are in accordance with the above
reported computational results, indicate that a significant popu-
lation of conformers of the two dipeptides 1a,b adopts, in
CDCl3 solution, a folded γ-turn conformation characterized by
an H-bond interaction between the Leu NH and the urethanic
carbonyl group.
Further information on the conformation adopted by the
N-Boc dipeptides 1a,b can be obtained by examining the results
of nuclear Overhauser effect (NOE) experiments summarized
in Table 3. The observation of an NH ؒ ؒ ؒ NH cross-peak
corresponding to close contact of the Adt and Leu NH protons
confirms that 1a adopts a folded γ-turn structure centered at the
Adt residue. An analogous spatial connectivity is shown by the
Ac5c and Leu NH groups of related N-Boc dipeptide 1b. A
further diagnostic feature in the NOESY spectrum of 1b con-
cerns the strong NOEs between the Ac5c NH and the four
adjacent ring protons (see Table 3 and Fig. 6): a strong effect is
observed only with the pair of protons CβHA which resonate
at higher field as compared with the CβHB protons. This is
in accordance with the presence of a γ-turn conformation in
which the Ac5c NH shows a large interproton distance only
with the CβHB pair; the two CβHB protons lie in fact on the
opposite side of the ring and experience a selective deshielding
by the cisoidal Ac5c amide carbonyl group. A similar outcome
Fig. 5 Chemical shift dependence of the NH resonances as a function
of the DMSO-d6 concentration (% v/v) in CDCl3 solution. A: N-Boc-
Adt-Leu-OMe (1a); B: HCO-Adt-Leu-OMe (2a); C: N-Boc-Ac5c-
Leu-OMe (1b); D: HCO-Ac5c-Leu-OMe (2b). Peptide concentration
10 mM.
was found to be practically concentration independent over the
range 10 to 1.0 mM. Thus, peptide self-association occurs, if at
all, to a limited extent under these experimental conditions and
the band at 3306 cmϪ1 can be assigned to the intramolecularly
H-bonded Leu NH group. The IR spectrum of 1b shows a very
close outcome in the NH stretching region: two bands are
observed at 3308 and 3436 cmϪ1, and the band at lower fre-
quency can be assigned to intramolecularly bonded Leu NH.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1984

View Article Online
Table 4 Solvent accessibility of peptide NH groups: differences (∆δ, ppm) between NH chemical shift values observed in CDCl3 solution contain-
ing (CD3)2SO (10%) and in neat CDCl3; see also Figs. 5 and 7
Compound
Adt or Ac5c NH
Leu NH
Ala NH
Boc-Adt-Leu-OMe (1a)
Boc-Ac5c-Leu-OMe (1b)
Boc-Adt-Leu-Ala-OMe (3a)
Boc-Ac5c-Leu-Ala-OMe (3b)
HCO-Adt-Leu-OMe (2a)
HCO-Ac5c-Leu-OMe (2b)
0.95
0.70
1.00
0.82
1.51
1.57
0.09
Ϫ0.17
0.26
0.01
0.06
0.62
0.35
0.005
Thp or Ac6c NH
HCO-Thp-Leu-OMea (4)
2.01
1.72
0.12
0.12
HCO-Ac6c-Leu-OMea (5)
a Ref. 22.
Fig. 6 Intramolecularly H-bonded γ-turn structure of dipeptides 1a,
2a (X = S) and 1b, 2b (X = CH2).
is exhibited by the Adt-containing dipeptide 1a; however, the
partial overlapping of the Adt CβHA and CβHB signals does not
allow, in this case, an exact evaluation of the Adt NH ؒ ؒ ؒ CH2
connectivity; this finding could reflect the different conform-
ations adopted by the two five-membered rings.
The above reported results of 1H-NMR solvent shielding
experiments can be usefully compared with those previously
obtained by us when studying the two N-formyldipeptide
models HCO-Thp-Leu-OMe (4) and HCO-Ac6c-Leu-OMe (5),
both containing a cyclic Cα-tetrasubstituted achiral residue at
the N-terminal position and exhibiting a tendency to adopt a
γ-turn conformation centered at the cyclic residue.22 For this
purpose we have summarized in Table 4 the differences (∆δ,
ppm) between the NH chemical shifts observed in CDCl3 con-
taining 10% (CD3)2SO and in neat CDCl3. It can be seen that
the Leu NH shows, as expected for intramolecularly bonded
protons, low ∆δ values for both the N-Boc dipeptides (1a and
1b) here under study as well as for the two previously studied
N-formyldipeptides 4 and 5. It can be observed, however, that
in the case of the N-Boc dipeptides 1a and 1b the Adt and Ac5c
NH groups show ∆δ values significantly lower than those exhib-
ited by the corresponding Thp and Ac6c NH of the previously
studied N-formyldipeptides (i.e.: 0.95 and 0.70 ppm in the case
of urethanic Adt and Ac5c NH, respectively; 2.01 and 1.72 ppm
for the Thp and Ac6c formamido NH groups in 4 and 5, respect-
ively). This finding indicates that the bulkiness of the Boc pro-
tecting group in 1a,b and 3a,b may significantly contribute to
lowering the solvent accessibility shown by the corresponding
NH group. In order to get information on this point, the related
N-formyl derivatives 2a,b were synthesized and examined. Data
reported in Figs. 5B and 5D and Table 4 show that the Adt and
the Ac5c formamido NH groups of 2a,b exhibit higher ∆δ
values (1.51 and 1.57 ppm, respectively) than those observed for
the corresponding Boc-protected NH groups of 1a and 1b (0.95
and 0.70 ppm, respectively). This confirms that the lower sensi-
tivity of chemical shift to the solvent composition exhibited by
urethanic NH protons of 1a and 1b is connected with the differ-
ent steric dimensions of the substituents at the NH groups.
The above described titration experiments were then
extended to tripeptides Boc-Adt-Leu-Ala-OMe (3a) and Boc-
Ac5c-Leu-Ala-OMe (3b). As shown in Fig. 7 and Table 4, in
Fig. 7 Chemical shift dependence of the NH resonances as a function
of the DMSO-d6 concentration (% v/v) in CDCl3 solution. A: N-Boc-
Adt-Leu-Ala-OMe (3a); B: N-Boc-Ac5c-Leu-Ala-OMe (3b). Peptide
concentration 10 mM.
both these derivatives the urethane NH resonance moves
downfield with increasing concentration of (CD3)2SO, while the
Leu NH group is practically unaffected by the change of the
solvent composition. An intermediate behaviour is observed in
the case of the Ala NH. These results confirm the tendency of
Adt and Ac5c residues to induce, in CDCl3 solution, γ-turn
conformations centered at the cyclic residue also in peptides
able to adopt β-bend structures. The intermediate solvent
accessibility shown by Ala NH (see Table 3) can be explained by
admitting the presence of a certain population of a β-turn or a
double γ-turn centered at the first two residues. Unfortunately,
in the NOESY spectra of both the tripeptides 3a,b the over-
lapping of Leu and Ala CαH, together with the close resonances
of the corresponding NH groups, does not allow the observ-
ation of the related diagnostic spatial connectivities. Further
information which can be obtained by examining the ∆δ values
reported in Table 4 concerns an evaluation of the relative
tendency to stabilize a γ-turn folding exhibited by the Adt and
Ac5c residues. It can be seen that the ∆δ values of the γ-turn
involved Leu NH protons show, in all three Ac5c-containing
peptides 1b–3b, a value lower than that exhibited by the corre-
sponding Adt-containing models 1a–3a. This strongly suggests
that, although both the cyclic residues induce the γ-turn fold-
ing, the tendency associated with the Ac5c presence is more
pronounced. This behaviour is not found in the case of the
cyclic residues of Thp and Ac6c in which (see Table 4) the Leu
NH presents the same ∆δ value.
Finally, in order to check whether a dipolar solvent
could disfavour the γ-turn conformations observed in CDCl3
solution, NOESY experiments have been carried out on
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1985

View Article Online
formyldipeptides 2a and 2b in (CD3)2SO solution. Only a weak
Leu NH ؒ ؒ ؒ Adt NH cross-peak was observed in the spectrum
of 2a, and the corresponding spatial connectivity between Leu
and Ac5c NH groups of 2b could not be detected. Furthermore,
the Leu NH group of both the formyldipeptides 2a and 2b
exhibit, in (CD3)2SO, relatively high temperature coefficients
(0.0051 ppm KϪ1 and 0.0043 ppm KϪ1, respectively). These data
strongly suggest that the folded γ-turn structure adopted by 2a
and 2b in CDCl3 solution is not retained in the polar solvent.
gradient based method as implemented in Gaussian 98.42
Hessian matrices were then diagonalized in order to evaluate
the actual nature of the potential energy saddle points, i.e. true
minima, first or higher order stationary points. In corre-
spondence with the deepest minima, with the use of their
harmonic frequencies and moments of inertia, the thermal
corrections to the energies as well as the entropic contributions
were then calculated at 298 K in the ideal gas approximation.
Finally the B3LYP/4-31G total energies were improved with
single point calculations carried out using the larger 6-31G(d)
basis set.43 The above calculations have been termed B3LYP/
6-31G(d)//B3LYP/4-31G.
Conclusions
NMR and IR data on the conformations adopted by dipeptides
containing the Adt (1a and 2a) or the Ac5c (1b and 2b) residue
at the N-terminal position, reported in the present paper, show
that in chloroform solution all the models prefer a γ-turn struc-
ture stabilized by an intramolecular H-bond (see Fig. 6); this
folding is not maintained in a polar solvent such as dimethyl
sulfoxide, where linear intermolecular H-bonding interactions
with the medium prevent the formation of the γ-turn structure.
The experimental conformational studies, performed in CDCl3
solution, were extended to the N-Boc-protected tripeptides
3a and 3b and the results obtained indicate the remarkable
tendency of both the five-membered Cα-tetrasubstituted cyclic
amino acids Adt and Ac5c to induce the γ-turn structure also in
models able to adopt the β-bend conformation. An analogous
behaviour was previously observed in the case of models con-
taining six-membered Cα-tetrasubstituted cyclic amino acids
Thp and Ac6c.22 The present results are in good agreement with
the conformational energy and free-energy calculations per-
formed on the Ac-Adt-NHMe and Ac-Ac5c-NHMe models.
However, while the computational data suggest that the Adt
residue, when compared to the Ac5c, is a more efficient γ-turn
inducer, both in vacuum and in apolar solvents, the spectral
data indicate that the tendency to induce this folding is more
pronounced in the case of the Ac5c residue. Such a discrepancy
is probably due to the well known limitations in the theoretical
description of the solutions ranging from the absence of
explicit solvent molecules to the impossibility of calculating an
exact molecular partition function, which obviously lower the
performances and the accuracy normally reached in the gas
phase by standard theoretical models.
The minima previously located in the vacuum calculations
were then re-optimized using the polarizable continuum model.
The solvents in such a model are represented by an infinite
dielectric medium at the temperature of interest, i.e. 298 K. The
cavity including the solute, defined in terms of interlocking
spheres, is built with the GePol procedure44 using the UAHF
atomic radii.45 The free energy of solvation is therefore evalu-
ated in correspondence with the located minima by consider-
ing the electrostatic contribution (∆Gel), i.e. the solvation
free energy arising from the electrostatic solute–solvent inter-
action, the cavitation contribution (∆Gcav) calculated through
Pierotti’s scaled particle theory46 and the dispersion–repulsion
contribution (∆Gdr).47 All these calculations were carried out at
a final B3LYP/6-31G(d)//B3LYP/4-31G level of theory.
The 298 K free energy differences between all the conformers
were finally calculated through standard thermodynamic cycles
involving the gaseous and solvated species using the ∆G298 in
vacuum and in solution. It should be remarked that the present
free energy calculations suffer from essentially two important
limitations: the solvation model was not improved by adding a
number of explicit solvent molecules and, most importantly, a
meaningful evaluation of the molecular partition function
cannot be derived with the presently available implementations
of the PCM. Nevertheless, the error induced by this latter
limitation is always assumed to be not dramatic.
Experimental measurements
Melting points were determined with a Büchi oil bath appar-
atus and are uncorrected. Optical rotations were taken at 20 ЊC
with a Schmidt-Haensch Polartronic D polarimeter in a 1 dm
cell (c 1.0 in CHCl3). Infrared spectra were recorded with a
Perkin-Elmer Spectrum 1000 FT-IR spectrometer (0.5, 2, and
Experimental
1
3 mm NaCl cells). H NMR spectra were measured with a
Dynamic Reaction Path calculations
Bruker AM 200 spectrometer in CDCl3 solution, unless other-
wise specified (tetramethylsilane as internal standard). J values
are given in Hz. NOESY spectra were recorded in CDCl3 and,
in the case of compounds 2a,b, also in DMSO-d6 on an Avance
2000 Bruker console with a 1200-ms mixing time and displayed
in the phase-sensitive mode. Usually 256 × 1024 data points
were collected and for each block 64 transients were collected
for two-dimensional experiments. The data sets were linearly
predicted to 512 × 1024 data points. A Gaussian window was
applied in both dimensions. Zero filling was used for a final
spectrum size of 1024 × 1024 data points. Elemental analyses
were performed at the Microanalysis Centre (Area della Ricerca
di Roma, CNR, Montelibretti, Italy). Column chromatography
was carried out using Merck silica gel 60 (230–400 mesh). Thin
layer chromatography was performed on silica gel Merck 60
F254 plates. The drying agent was sodium sulfate. All reagents
and solvents were purchased at highest commercial quality and
used without further purification. Light petroleum refers to the
40–60 ЊC bp fraction.
Dynamic Reaction Path calculations were carried out using the
implementation of Gordon and Taketsugu present in the pack-
age Gamess US.38 The trajectories were subsequently analyzed
with our own routines interfaced with the above package. The
initial system is ‘heated’ up to a Maxwellian distribution of
298 K and then allowed to propagate and equilibrate using a
velocity-Verlet39 algorithm with a time step of 0.1 fs by keep-
ing constant the energy and the angular momentum. Both the
trajectories were stopped after about 150 000 steps. The AM1
level of theory31 was adopted for the energy and gradient
calculations.
Density Functional Theory calculations
The Born–Oppenheimer potential energy hypersurfaces of
Ac-Adt-NHMe and Ac-Ac5c-NHMe were fully investigated
without any geometrical constraint in the framework of the
Density Functional Theory (DFT) with the use of the B3LYP
functional.40 For this purpose a split-valence 4-31G basis41 was
selected. The use of such a relatively small atomic basis set was
mainly imposed by the huge number of energy minimisations
required for an exhaustive sampling of the hypersurfaces. The
above calculations have been termed B3LYP/4-31G. Several
critical points were located by using a standard analytical-
N-Boc-Adt-Leu-OMe (1a)
To a solution of N-Boc-Adt-OMe19 (0.100 g, 0.358 mmol)
in MeOH (1.5 cm3) 2 M NaOH (0.36 cm3) was added under
stirring. After 6 h at room temperature, the solution was
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1986

View Article Online
concentrated under vacuum, acidified with 0.5 M HCl, and
extracted with EtOAc; the organic phase was washed with
water, dried, and evaporated. To the residue, dissolved in DMF
(3.0 cm3) and cooled to 0 ЊC, HOBt (0.051 g, 0.376 mmol)
and EDC (0.072 g, 0.376 mmol) were added. After stirring for
15 min at 0 ЊC and for 20 min at room temperature, H-Leu-
OMeؒHCl (0.072 g, 0.394 mmol) and TEA (0.055 cm3, 0.394
mmol) were added. The reaction mixture was stirred at room
temperature for 18 h and then diluted with water and extracted
with EtOAc. The organic phase was washed with water, 10%
citric acid solution, saturated NaHCO3, and water, dried, and
evaporated under reduced pressure. The residue (0.149 g) was
chromatographed on silica gel (4.5 g) using CHCl3 as eluent to
give title dipeptide 1a (0.125 g, 89%), mp 213–214 ЊC (from
acetone). (Found: C, 48.9; H, 7.1. C16H28N2O5S2 requires C,
49.0; H, 7.2%); [α]D = Ϫ44Њ; νmax(CHCl3)/cmϪ1 3419, 1722, 1684,
1471, and 1157; 1H NMR: δ 0.94 [6H, m, CH(CH3)2], 1.45 [9H,
s, C(CH3)3],1.44–1.75 [3H, m, CH2-CH(CH3)2], 3.34–3.80 (4H,
m, two Adt β-CH2), 3.73 (3H, s, COOCH3), 4.63 (1H, m, Leu
α-CH), 5.25 (1H, s, NH-COO), 7.03 (1H, d, J 8, Leu NH).
CβHA and two γ-CH2], 2.26–2.46 (2H, m, Ac5c CβHB), 3.73 (3H,
s, COOCH3), 4.54 (1H, m, Leu α-CH), 5.84 (1H, s, Ac5c NH),
7.25 (1H, d, J 7.3, Leu NH), 8.17 (1H, d, J 1.6, HCO).
N-Boc-Adt-Leu-Ala-OMe (3a)
To a solution of N-Boc-Leu-Ala-OMe48 (0.113 g, 0.358 mmol)
in dry MeOH (2.5 cm3) cooled at 0 ЊC thionyl chloride
(0.026 cm3, 0.358 mmol) was added and the mixture was stirred
at 50 ЊC for 3 h. Evaporation under reduced pressure of the
solution afforded H-Leu-Ala-OMeؒHCl which was used in the
next step without further purification. To the stirred suspension
of the hydrochloride in dry DMF (0.5 cm3) cooled at 0 ЊC
N-Boc-Adt-OH (0.095 g, 0.358 mmol), HOBt (0.048 g, 0.358
mmol), EDC (0.075 g, 0.358 mmol), and, after 15 min at 0 ЊC
and 20 min at room temperature, TEA (0.050 cm3, 0.358 mmol)
were added. The mixture was stirred at room temperature for
16 h and then worked up as described for 1a to give a residue
(0.156 g) which, chromatographed on silica gel (5 g) using
CHCl3–EtOAc = 8 : 2 as eluent, afforded 0.131 g (79%) of
3a, mp 154–155 ЊC (from acetone). (Found: C, 49.3; H,
7.1. C19H33N3O6S2 requires C, 49.2; H, 7.2%); [α]D = Ϫ27Њ;
νmax(CHCl3)/cmϪ1 3415, 1716, 1627, 1504, and 1157; 1H NMR:
δ 0.92 [6H, m, CH(CH3)2], 1.40 (3H, d, J 7.2 Hz, Ala β-CH3),
1.44 [9H, s, C(CH3)3], 1.55–1.76 [3H, m, CH2-CH(CH3)2], 3.30–
3.70 (4H, m, two Adt β-CH2), 3.72 (3H, s, COOCH3), 4.35–4.58
(2H, m, Ala and Leu α-CH), 5.26 (1H, s, NH-COO), 6.75 (1H,
br, Ala NH), 6.85 (1H, d, J 7.7, Leu NH).
N-Boc-Ac5c-Leu-OMe (1b)
To a solution of N-Boc-Ac5c-OH (0.200 g, 0.873 mmol) in dry
DMF (7.3 cm3) cooled at 0 ЊC HOBt (0.124 g, 0.916 mmol) and
EDC (0.176 g, 0.916 mmol) were added. After stirring 15 min
at 0 ЊC H-Leu-OMeؒHCl (0.175 g, 0.960 mmol) and TEA
(0.134 cm3, 0.960 mmol) were added. The reaction mixture was
stirred at room temperature for 19 h and then worked up as
described for 1a to give crude 1b which was further purified by
crystallization from diethyl ether (0.260 g, 84%), mp 160–161
ЊC. (Found: C, 60.7; H, 9.1. C18H32N2O5 requires C, 60.6; H,
9.0%); [α]D = Ϫ18Њ; νmax(CHCl3)/cmϪ1 3436, 3308, 1736, 1671,
1504, and 1161; 1H NMR: δ 0.91 [6H, m, CH(CH3)2], 1.45 [9H,
s, C(CH3)3],1.50–1.95 [9H, m, CH2-CH(CH3)2, Ac5c CβHA
and two γ-CH2], 2.16–2.48 (2H, m, Ac5c CβHB), 3.72 (3H, s,
COOCH3), 4.61 (1H, m, Leu α-CH), 4.80 (1H, s, NH-COO),
7.07 (1H, br, Leu NH).
N-Boc-Ac5c-Leu-Ala-OMe (3b)
To a solution of 1b (0.346 g, 1 mmol) in MeOH (5.5 cm3), 2 M
NaOH (1 ml) was added under stirring. After 6 h at room
temperature, the solution was concentrated under vacuum
below 30 ЊC, acidified with 0.5 M HCl, and extracted with
EtOAc; the organic phase was washed with water, dried, and
evaporated. The resulting residue was used in the next step
without further purification.
To the previous residue dissolved in DMF (5.5 cm3) and
cooled to 0 ЊC, HOBt (0.170 g, 1.1 mmol) and EDC (0.211 g,
1.1 mmol) were added. After stirring 15 min at 0 ЊC and 20 min
at room temperature, H-Ala-OMeؒHCl (0.154 g, 1.1 mmol) and
TEA (0.155 cm3, 1.1 mmol) were added. The reaction mixture
was stirred at room temperature for 18 h and then worked up as
described for 1a to give a residue (0.42 g) which, chromato-
graphed on silica gel (16 g) using CH2Cl2–EtOAc = 8 : 2 and
6 : 4 as eluent, afforded 0.341 g of the title product 3b (80%),
mp 140–141 ЊC (from EtOAc–light petroleum). (Found: C, 59.0;
H, 8.8. C21H37N3O6 requires C, 59.0; H, 8.7%); [α]D = Ϫ35Њ;
HCO-Adt-Leu-OMe (2a)
A solution of 1a (0.102 g, 0.26 mmol) in 1 cm3 of formic acid
was stirred at room temperature for 18 h. After removal of the
excess of formic acid under vacuum, the residue was dissolved
in dry DMF (0.7 cm3) and EEDQ (0.078 g, 0.312 mmol) was
added. The solution was stirred at room temperature for
24 h. Evaporation under reduced pressure afforded a residue
(0.080 g) which was chromatographed on silica gel (5 g), eluting
with CHCl3–MeOH (99 : 1) to give pure 2a (0.068 g, 82%),
mp 72–73 ЊC (from CH2Cl2–n-hexane). (Found: C, 44.8; H, 6.2.
C12H20N2O4S2 requires C, 45.0; H, 6.3%); [α]D = Ϫ29Њ;
1
νmax(CHCl3)/cmϪ1 3431, 1739, 1714, 1670, 1501, and 1161; H
NMR: δ 0.93 [6H, m, CH(CH3)2], 1.41 (3H, d, J 7.2 Hz, Ala
β-CH3), 1.44 [9H, s, C(CH3)3], 1.50–1.98 [9H, m, CH2-CH-
(CH3)2, Ac5c CβHA and two γ-CH2], 2.16–2.43 (2H, m, Ac5c
CβHB), 3.73 (3H, s, COOCH3), 4.40–4.57 (2H, m, Ala and Leu
α-CH), 4.85 (1H, s, NH-COO), 6.78 (1H, d, J 7.7, Leu NH),
6.98 (1H, br, Ala NH).
1
νmax(KBr)/cmϪ1 3314, 1743, 1676, 1547, and 1212; H NMR:
δ 0.93 [6H, d, J = 5.9, CH(CH3)2], 1.57–1.70 [3H, m, CH2-
CH(CH3)2], 3.50–3.67 (4H, m, two Adt β-CH2), 3.73 (3H, s,
COOCH3), 4.56 (1H, m, Leu α-CH), 6.49 (1H, s, Adt NH), 7.23
(1H, d, J 7.3, Leu NH), 8.25 (1H, d, J 1.3, HCO).
Abbreviations
HCO-Ac5c-Leu-OMe (2b)
A solution of 1b (0.100 g, 0.28 mmol) in 0.7 cm3 of formic acid
was stirred at room temperature for 20 h. After removal of the
excess of formic acid under vacuum, the residue was dissolved
in dry DMF (1.2 cm3) and EEDQ (0.083 g, 0.336 mmol) was
added. The solution was stirred at room temperature for 18 h.
Evaporation under reduced pressure afforded a residue (0.080
g) which was chromatographed on silica gel (5 g), eluting with
CHCl3–EtOAc (1 : 1) to give pure 2b (0.070 g, 88%), mp 127–
127.5 ЊC (from CH2Cl2–diethyl ether). (Found: C, 59.2; H, 8.4.
C14H24N2O4 requires C, 59.1; H, 8.5%); [α]D = Ϫ5Њ; νmax(KBr))/
cmϪ1: 3292, 1751, 1736, 1665, 1543, and 1157; 1H NMR: δ 0.95
[6H, m, CH(CH3)2], 1.54–2.15 [9H, m, CH2-CH(CH3)2, Ac5c
B3LYP
hybrid functional for Density Functional
Theory calculations
tert-butyloxycarbonyl
dicyclohexylcarbodiimide
N,N-dimethylformamide
dimethyl sulfoxide
Dynamic Reaction Coordinate
N-(3-dimethylaminopropyl)-NЈ-ethylcarbo-
diimide hydrochloride
Boc
DCC
DMF
DMSO
DRC
EDC
EEDQ
EtOAc
ethyl 2-ethoxy-1,2-dihydro-1-quinoline-
carboxylate
ethyl acetate
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1987

View Article Online
22 M. Paglialunga Paradisi, I. Torrini, G. Pagani Zecchini, G. Lucente,
E. Gavuzzo, F. Mazza and G. Pochetti, Tetrahedron, 1995, 51, 2379–
2386.
23 T. P. Curran, N. M. Chandler, R. J. Kennedy and M. T. Keaney,
Tetrahedron Lett., 1996, 37, 1933–1936.
24 A. F. Spatola, M. K. Anwer, A. L. Rockwell and L. M. Gierasch,
J. Am. Chem. Soc., 1986, 108, 825–831.
f.e.
HOBt
MD
NMM
TEA
THF
fully extended conformation
1-hydroxybenzotriazole
Molecular Dynamics
N-methylmorpholine
triethylamine
tetrahydrofuran
25 A. M. Sapse, L. Mallah-Levy, S. B. Daniels and B. W. Erickson,
J. Am. Chem. Soc., 1986, 109, 3256–3262.
Acknowledgements
26 K. Burgess and K. K. Ho, J. Am. Chem. Soc., 1994, 116, 799–805.
27 R. Haubner, R. Gracias, B. Diefenbach, S. L. Goodman, A. Jonczyk
and H. Kessler, J. Am. Chem. Soc., 1996, 118, 7461–7466.
28 S. Herrero, M. T. Garcìa-Lòpez, M. Latorre, E. Cenarruzabeitia,
J. Del Rio and R. Herranz, J. Org. Chem., 2002, 67, 3866–3873 and
references therein.
We would like to thank Professor Alfredo Di Nola (University
of Rome - La Sapienza) and CASPUR (Supercomputing
Centre of Rome) for providing all the computational facilities.
29 L. Teuber, Sulfur Rep., 1990, 9, 257–349 and references therein.
30 J. J. P. Stewart, L. P. Davis and L. W. Burggraf, J. Comput. Chem.,
1987, 8, 1117–1123.
31 M. J. S. Dewar and W. Thiel, J. Am. Chem. Soc., 1977, 99, 4999–
5005.
32 See for example: R. Improta, N. Rega, C. Aleman and V. Barone,
Macromolecules, 2001, 34, 7550–7557.
33 S. Miertus, E. Scrocco and J. Tomasi, J. Chem. Phys., 1981, 55,
117–122.
References
1 C. Toniolo, M. Crisma, F. Formaggio and C. Peggion, Biopolymers
(Peptide Sciences), 2001, 60, 396–419 and references therein.
2 P. Balaram, Curr. Opin. Struct. Biol., 1992, 2, 845–851.
3 C. Toniolo, Janssen Chim. Acta, 1993, 11, 10–16.
4 P. Sudhanand, R. Balaji and P. Balaram, Biopolymers, 1995, 35,
11–20.
5 I. Torrini, M. Paglialunga Paradisi, G. Pagani Zecchini and
G. Lucente, Synth. Commun., 1994, 24, 153–158.
6 M. Cirilli, V. M. Coiro, A. Di Nola and F. Mazza, Biopolymers,
1998, 46, 239–244.
7 C. Cativiela and M. D. Diaz-de-Villegas, Tetrahedron: Asymmetry,
1998, 9, 3517–3599.
8 R. Kaul, S. Banumathi, D. Velmurugan, R. Balaji Rao and
P. Balaram, Biopolymers, 2000, 54, 159–167.
9 D. C. Horwell, Bioorg. Med. Chem., 1996, 4, 1573–1576.
10 I. Torrini, G. Pagani Zecchini, M. Paglialunga Paradisi, G. Lucente,
G. Mastropietro, E. Gavuzzo, F. Mazza, G. Pochetti, S. Traniello
and S. Spisani, Biopolymers, 1996, 39, 327–337.
11 E. Mossel, F. Formaggio, M. Crisma, C. Toniolo, Q. B. Broxterman,
W. H. J. Boesten, J. Kamphuis, P. J. L. M. Quaedflieg and P. Temussi,
Tetrahedron: Asymmetry, 1997, 8, 1305–1314.
12 I. Torrini, M. Paglialunga Paradisi, G. Pagani Zecchini, G. Lucente,
E. Gavuzzo, F. Mazza, G. Pochetti, S. Traniello and S. Spisani,
Biopolymers, 1997, 42, 415–426.
13 A. G. S. Blommaert, H. Dhotel, B. Ducos, C. Durieux,
N. Goudreau, A. Bado, C. Garbay and B. P. Roques, J. Med. Chem.,
1997, 40, 647–658.
34 J. Tomasi and M. Persico, Chem. Rev., 1994, 94, 2027–2035.
35 G. Lajoie and J. L. Kraus, Peptides, 1984, 5, 563.
36 P. A. Ray and P. Balaram, Biopolymers, 1985, 24, 1131–1146.
37 S. Zerkout, V. Dupont, A. Aubry, J. Vidal, A. Collet, A. Vicherat and
A. Marraud, Int. J. Pept. Protein Res., 1994, 44, 378–387.
38 Gamess (version R1). M. W. Schmidt, K. K. Baldridge, J. A. Boatz,
S. T. Elbert, M. S. Gordon, J. H. Jensen, S. Koseki, N. Matsunaga,
K. A. Nguyen, S. J. Su and T. L. Windus, J. Comput. Chem., 1993,
14, 1347–1363.
39 M. P. Allen and D. J. Tildesley, Computer Simulations of Liquids,
Oxford University Press, Oxford, 1987.
40 A. D. Becke, J. Chem. Phys., 1992, 98, 5648–5652.
41 R. Ditchfield, W. J. Hehre and J. A. Pople, J. Chem. Phys., 1971, 54,
724–728.
42 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery,
Jr., R. E. Stratmann, J. C. Burant, S. Dapprich, J. M. Millam,
A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi,
V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala,
Q. Cui, K. Morokuma, P. Salvador, J. J. Dannenberg, D. K. Malick,
A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski,
J. V. Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko,
P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox,
T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara,
M. Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen,
M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-Gordon,
E. S. Replogle and J. A. Pople, GAUSSIAN 98 (Revision A.11),
Gaussian, Inc., Pittsburgh, PA, 2001.
14 S. D. Bryant, R. Guerrini, S. Salvadori, C. Bianchi, R. Tomatis,
M. Attila and L. H. Lazarus, J. Med. Chem., 1997, 40, 2579–2587.
15 C. Toniolo, Int. J. Pept. Protein Res., 1990, 35, 287–300.
16 M. Saviano, R. Iacovino, V. Menchise, E. Benedetti, G. M. Bonora,
M. Gatos, L. Graci, F. Formaggio, M. Crisma and C. Toniolo,
Biopolymers, 2000, 53, 200–212.
17 C. Cativiela and M. D. Diaz-de-Villegas, Tetrahedron: Asymmetry,
2000, 11, 645–732.
18 A. Moretto, F. Formaggio, M. Crisma, C. Toniolo, M. Saviano,
R. Iacovino, R. M. Vitale and E. Benedetti, J. Pept. Res., 2001, 57,
307–315.
19 E. Morera, M. Nalli, F. Pinnen, D. Rossi and G. Lucente,
Bioorg. Med. Chem. Lett., 2000, 10, 1585–1588 and references
therein.
20 E. Morera, G. Lucente, G. Ortar, M. Nalli, F. Mazza, E. Gavuzzo
and S. Spisani, Bioorg. Med. Chem., 2002, 10, 147–157.
21 E. Morera, F. Pinnen and G. Lucente, Org. Lett., 2002, 4,
1139–1142.
43 M. M. Francl, W. J. Petro, W. J. Hehre, J. S. Binkley, M. S. Gordon,
D. J. DeFree and J. A. Pople, J. Chem. Phys., 1982, 77, 3654–
3659.
44 J. Pasqual-Ahuir, E. Silla, J. Tomasi and R. Bonaccorsi, J. Comput.
Chem., 1996, 17, 56–71.
45 V. Barone, M. Cossi and J. Tomasi, J. Chem. Phys., 1997, 107,
3210–3215.
46 R. A. Pierotti, Chem. Rev., 1976, 76, 717–730.
47 F. M. Floris and J. Tomasi, J. Comput. Chem., 1986, 10, 616–630.
48 K. Pudhom and T. Vilaivan, Synth. Commun., 2001, 31, 61–70.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 1 9 8 0 – 1 9 8 8
1988