Synthesis of conformationally constrained peptidomimetics using multicomponent reactions

Article abstract of DOI:10.1021/jo802052j

A novel modular synthetic approach toward constrained peptidomimetics is reported. The approach involves a highly efficient three-step sequence including two multicomponent reactions, thus allowing unprecedented diversification of both the peptide moietie

Full text of DOI:10.1021/jo802052j

Synthesis of Conformationally Constrained Peptidomimetics using
Multicomponent Reactions
Rachel Scheffelaar,† Roel A. Klein Nijenhuis,† Monica Paravidino,† Martin Lutz,‡
Anthony L. Spek,‡ Andreas W. Ehlers,† Frans J. J. de Kanter,† Marinus B. Groen,†
Romano V. A. Orru,*,† and Eelco Ruijter*,†
Department of Chemistry and Pharmaceutical Sciences, Vrije UniVersiteit Amsterdam, De Boelelaan 1083,
1081 HV Amsterdam, The Netherlands and BijVoet Center for Biomolecular Research, Crystal and
Structural Chemistry, Utrecht UniVersity, Padualaan 8, 3584 CH Utrecht, The Netherlands
orru@few.Vu.nl; e.ruijter@few.Vu.nl
ReceiVed September 17, 2008
A novel modular synthetic approach toward constrained peptidomimetics is reported. The approach involves
a highly efficient three-step sequence including two multicomponent reactions, thus allowing unprecedented
diversification of both the peptide moieties and the turn-inducing scaffold. The turn-inducing properties
of the dihydropyridone scaffold were evaluated by molecular modeling, X-ray crystallography, and NMR
studies of a resulting peptidomimetic. Although modeling studies point toward a type IV ꢀ-turn-like
structure, the X-ray crystal structure and NMR studies indicate an open turn structure.
Introduction
A major challenge in the field of peptidomimetics is the
design and synthesis of conformationally constrained analogues
that mimic secondary structural features of biologically active
peptides or proteins. Among these structural motifs, the reverse
turns commonly found in bioactive peptides are believed to be
essential for protein folding.1 Furthermore, they function as
molecular recognition elements for the interaction between
peptides and their receptors but also play an important role in
antibody recognition.2
FIGURE 1. Schematic diagram of a ꢀ-turn (A) and the Freidinger
lactams (B).
A well studied subset of the reverse turns are the ꢀ-turns
(Figure 1A), which are defined as any tetrapeptide sequence
occurring in the nonhelical region of a protein in which the
CRi-CRi+3 spatial distance is e7 Å. In most cases, ꢀ-turns are
stabilized by an intramolecular hydrogen bond between residues
(i) and (i + 3) forming a pseudo-10-membered ring.1 Moreover,
ꢀ-turns are classified according to the geometry of their peptide
backbone, determined by the torsion angles ꢁ and Ψ. Conse-
quently, a range of different turn types (I-VIII) have been
described.1,3
Topographical mimics of ꢀ-turns receive considerable atten-
tion in medicinal chemistry, and synthetic strategies for their
preparation have evolved markedly in this area.4 The incorpora-
† Vrije Universiteit Amsterdam.
‡ Utrecht University.
(1) Rose, G. D.; Gierasch, L. M.; Smith, J. A. AdV. Protein Chem. 1985, 37,
1–109.
(3) (a) Venkatachalam, C. M. Biopolymers 1968, 6, 1425–1436. (b) Chou,
P. Y.; Fasman, G. D. J. Mol. Biol. 1977, 115, 135–175. (c) Wilmot, C. M.;
Thornton, J. M. J. Mol. Biol. 1988, 203, 221–232. (d) Ball, J. B.; Hughes, R. A.;
Alewood, P. F.; Andrews, P. R. Tetrahedron 1993, 49, 3467–3478.
(2) Burgess, K. Acc. Chem. Res. 2001, 34, 826–835.
660 J. Org. Chem. 2009, 74, 660–668
10.1021/jo802052j CCC: $40.75 2009 American Chemical Society
Published on Web 12/04/2008

Synthesis of Conformationally Constrained Peptidomimetics
SCHEME 1. MCR toward 3,4-DHP-2-ones and Follow-Up Passerini Chemistry
SCHEME 2. Retrosynthetic Analysis of MCR-Derived
Constrained Peptidomimetics
tion of conformational constraints can enhance the potency,
receptor selectivity, and (peptidase) stability of the parent
peptide.4 Furthermore, it can provide useful information on the
biologically active conformation of the peptide.5 Among the
various different strategies, of particular interest is the introduc-
tion of constrained dipeptide scaffolds, i.e., dipeptide isosters,
in the parent peptide that are able to induce a folding of the
peptide chain.6 A pioneer in this field is Freidinger, who
developed dipeptide γ-, δ-, ε-lactams (Figure 1B) as constrained
scaffolds to stabilize turn conformations. He elegantly validated
the concept by preparing a potent analogue of LHRH7 (incor-
poration of a γ-lactam).5,8
In the course of a program directed toward the search for
new multicomponent methodologies to access pharmaceutically
relevant heterocyclic scaffolds,9 we recently developed a novel
four-component reaction (4CR) of phosphonate 1, nitriles 2,
aldehydes 3, and R-aryl isocyanoacetates 4 to obtain 3,4-
dihydropyridin-2-ones 5 (3,4-DHP-2-ones). This 4CR generally
proceeds in good to excellent yields and with complete
diastereoselectivity in favor of the cis-diastereomer.10 The 4CR
is highly chemoselective, leaving the rather reactive isonitrile
functionality intact. This exceptional chemoselectivity allows
isonitrile-based follow-up chemistry. This was demonstrated by
performing a Passerini 3CR to construct highly functionalized
and conformationally constrained depsipeptides 6 in a novel 6CR
(Scheme 1).11 The structural resemblance of the DHP-2-one
scaffold in 5 to the Freidinger lactams inspired us to investigate
its potential as a turn-inducing moiety. Our strategy involves a
modular MCR-alkylation-MCR sequence. After the initial
4CR to access the DHP-2-one scaffold as the Freidinger-type
lactam, construction of the dipeptide isoster (given in gray,
Scheme 2) can be realized by introduction of a peptide chain
onto the DHP-2-one core by N-alkylation. After this a second
MCR, e.g., a Passerini 3CR or Ugi 4CR (Scheme 2), is
envisioned. This introduces a second (depsi)peptide chain. In
case of an Ugi-4CR this would complete a tetrapeptide (or
larger) sequence, in which the turn conformation can in principle
be stabilized by two intramolecular hydrogen bonds. Most
synthetic methods for Freidinger-type lactams lack flexibility
and are therefore not suited to introduce diversity.12 This is,
however, important because the lactam ring is often positioned
in the pharmacophoric region of the peptide. Flexible synthetic
methodology toward Freidinger-type lactams and the corre-
sponding tetrapeptide turn-mimics are of considerable interest
to facilitate optimization studies.12 Our 4CR for the synthesis
of the DHP-2-one core is very flexible, and variation of all
components proved possible.10 In addition, the modular
MCR-alkylation-MCR strategy as discussed above further
increases the number of diversity points, making our approach
ideally suited for the generation of libraries of conformationally
constrained peptidomimetics and potential ꢀ-turn mimetics.13
In this paper we discuss the results of our synthetic studies and
the turn-inducing properties of the resulting peptidomimetics
in full detail.
(4) For reviews on peptidomimetics, see: (a) Ball, J. B.; Alewood, P. F. J.
Mol. Recognit. 1990, 3, 55–64. (b) Giannis, A.; Kolter, T. Angew. Chem., Int.
Ed. Engl. 1993, 32, 1244–1267. (c) Gante, J. Angew. Chem., Int. Ed. Engl. 1994,
33, 1699–1720. (d) Hanessian, S.; McNaughton-Smith, G.; Lombart, H.-G.;
Lubell, W. D. Tetrahedron 1997, 53, 12789–12854. (e) Giannis, A.; Ru¨bsam,
F. AdVances in Drug Research; Test, B., Ed.; Academic Press: London, 1997;
Vol. 29, pp 1-78. (f) Synthesis of Peptides and Peptidomimetics. In Houben-
Weyl, Methods of Organic Chemistry; Felix, A., Moroder, L., Toniolo, C., Eds.;
Thieme: Stuttgart, New York, 2003; Vol. E22c. (g) Kee, S.; Jois, S. D. S. Curr.
Pharm. Des. 2003, 9, 1209–1224. (h) Perdih, A.; Kikelj, D. Curr. Med. Chem.
2006, 13, 1525–1556.
(5) A good illustration of this concept is the constrained analogue of LHRH
(see ref 7) by Freidinger: Freidinger, R. M.; Veber, D. F.; Perlow, D. S.; Brooke,
J. R.; Saperstein, R. Science 1980, 210, 656–658.
(6) See also ref 4. For recent examples, see: (a) Palomo, C.; Aizpurua, J. M.;
Benito, A.; Miranda, J. I.; Fratila, R. M.; Matute, C.; Domercq, M.; Gago, F.;
Martin-Santamaria, S.; Linden, A. J. Am. Chem. Soc. 2003, 125, 16243–16260.
(b) Ecija, M.; Diez, A.; Rubiralta, M.; Casamitjana, N. J. Org. Chem. 2003, 68,
9541–9553. (c) Dietrich, S. A.; Banfi, L.; Basso, A.; Damonte, G.; Guanti, G.;
Riva, R. Org. Biomol. Chem. 2005, 3, 97–106. (d) Van Rompaey, K.; Ballet,
S.; To¨mbo¨ly, C.; De Wachter, R.; Vanommeslaeghe, K.; Biesemans, M.; Willem,
R.; Tourwe´, D. Eur. J. Org. Chem. 2006, 2899–2911. (e) Einsiedel, J.; Lanig,
H.; Waibel, R.; Gmeiner, P. J. Org. Chem. 2007, 72, 9102–9113.
(7) LHRH: luteinizing hormone-releasing hormone.
Results and Discussion
(8) (a) Freidinger, R. M.; Veber, D. F.; Hirschmann, R.; Paege, L. M. Int. J.
Pept. Protein Res. 1980, 16, 464–470. (b) Freidinger, R. M.; Perlow, D. S.;
Veber, D. F. J. Org. Chem. 1982, 47, 104–109. (c) Freidinger, R. M. J. Med.
Chem. 2003, 46, 5553–5566.
Modeling Studies. Hydrophobic and in particular aromatic
elements in receptor ligands often play an important role in
recognition and activation of receptors possessing lipophilic
domains, and that is why many ligand mimics contain these
lipophilic elements.14 Numerous effective benzo-fused Fre-
idinger lactams and benzodiazepine mimics (mostly antagonists)
(9) (a) Bon, R. S.; Van Vliet, B.; Sprenkels, N. E.; Schmitz, R. F.; de Kanter,
F. J. J.; Stevens, C. V.; Swart, M.; Bickelhaupt, F. M.; Groen, M. B.; Orru,
R. V. A. J. Org. Chem. 2005, 70, 3542–3553. (b) Vugts, D. J.; Koningstein,
M. M.; Schmitz, R. F.; de Kanter, F. J. J.; Groen, M. B.; Orru, R. V. A. Chem.
Eur. J. 2006, 12, 7178–7189. (c) Elders, N.; Schmitz, R. F.; de Kanter, F. J. J.;
Ruijter, E.; Groen, M. B.; Orru, R. V. A. J. Org. Chem. 2007, 72, 6135–6142.
(d) Groenendaal, B.; Vugts, D. J.; Schmitz, R. F.; de Kanter, F. J. J.; Ruijter, E.;
Groen, M. B.; Orru, R. V. A. J. Org. Chem. 2008, 73, 719–722. (e) Elders, N.;
de Kanter, F. J. J.; Ruijter, E.; Groen, M. B.; Orru, R. V. A. Chem. Eur. J.
2008, 14, 4961–4974. (f) Groenendaal, B.; Ruijter, E.; de Kanter, F. J. J.; Lutz,
M.; Spek, A. L.; Orru, R. V. A. Org. Biomol. Chem. 2008, 6, 3158–3165.
(10) Paravidino, M.; Bon, R. S.; Scheffelaar, R.; Vugts, D. J.; Znabet, A.;
Schmitz, R. F.; de Kanter, F. J. J.; Lutz, M.; Spek, A. L.; Groen, M. B.; Orru,
R. V. A. Org. Lett. 2006, 8, 5369–5372.
(11) Paravidino, M.; Scheffelaar, R.; Schmitz, R. F.; de Kanter, F. J. J.; Groen,
M. B.; Ruijter, E.; Orru, R. V. A. J. Org. Chem. 2007, 72, 10239–10242.
(12) Golebiowski, A.; Jozwik, J.; Klopfenstein, S. R.; Colson, A.-O.; Grieb,
A. L.; Russell, A. F.; Rastogi, V. L.; Diven, C. F.; Portlock, D. E.; Chen, J. J.
J. Comb. Chem. 2002, 4, 584–590.
(13) For a review on ꢀ-turn mimetic library synthesis, see: Souers, A. J.;
Ellman, J. A. Tetrahedron 2001, 57, 7431–7448.
J. Org. Chem. Vol. 74, No. 2, 2009 661

Scheffelaar et al.
FIGURE 2. Representation of the lowest energy conformers of model systems 9 and 10.
TABLE 1. Two Lowest Energy Conformers of 9 and 10 Adopting a ꢀ-Turn Geometry and Calculated Values of the Different Geometrical
Parameters
a
b
c
d
d(CRi-CRi+3) [Å] d(COi-NHi+3) [Å] ꢁi+1 [deg] Ψi+1 [deg] ꢁi+2 [deg] Ψi+2 [deg] Eaq (kcal/mol) ꢀ-turn type
9 (conformer 1)
9 (conformer 3)
10 (conformer 1)
10 (conformer 2)
8.87
5.36
5.06
4.87
4.75
2.36
2.15
2.14
28
45
49
34
77
73
73
71
-86
84
84
67
-58
-45
-48
-176.14
-175.23
-136.72
-135.40
IV
IV
IV
87
a ꢁi+1 ) torsion angle Ci-Ni+1-CRi+1-Ci+1
. . .
b Ψi+1 ) torsion angle Ni+1-CRi+1-Ci+1-Ni+2 c ꢁi+2 ) torsion angle Ci+1-Ni+2-CRi+2-Ci+2 d Ψi+2
) torsion angle Ni+2-CRi+2-Ci+2-Ni+3
.
have been reported.4h Therefore, we chose, the DHP-2-one
scaffold 5a (R1 ) R2 ) R3 ) Ph) as the standard heterocyclic
core and potential turn-inducing element in model systems 9
and 10. After the 4CR to access 5a, depsipeptide 9 and amide
variant 10 can, in principle, be prepared by combining an
alkylation with the Passerini 3CR and Ugi 4CR, respectively,
as outlined above. However, to determine the propensity of the
3,4-DHP-2-one scaffold to indeed induce a ꢀ-turn-like confor-
mation in 9 and 10, modeling studies using Spartan 02 were
performed first (Figure 2). The procedure consists of a MM-
SYBYL conformational search followed by an AM1 geometry
optimization and single point energy calculation using AM1
including solvent effects (SM5). The turn-inducing propensity
of the lowest energy conformers of 9 and 10 was evaluated by
analyzing different geometrical features: (1) the CRi-CRi+3
distance (d) (<7 Å); (2) the ꢁ and Ψ backbone torsion angles
in residues 2 and 3; (3) the distance between the carbonyl
oxygen of residue 1 and the amide proton of residue 4 indicative
for intramolecular hydrogen bonding (<2.5 Å).
lowest energy conformer that adopts a turn conformation
(type IV, the third conformation, Figure 2B) is 0.91 kcal/
mol above the global minimum. Considering the error
margins of the method used, this is a feasible conformation.
In contrast, the lowest energy conformer of tetrapeptide 10
shows to be a ꢀ-turn of type IV according to the backbone
torsion angle classification (Table 1 and Figure 2C). In
addition to a turn-stabilizing hydrogen bond forming a
pseudo-10-membered ring, a second stabilizing intramolecular
hydrogen bond is present in conformer 2C of tetrapeptide
10, forming a pseudo-14-membered ring. The second lowest
energy conformer (1.32 kcal/mol above the global minimum,
Figure 2D) is also a type IV turn, although it lacks the second
intramolecular hydrogen bond. Moreover, within 4.0 kcal/
mol several more turn structures were found. Thus, the
tetrapeptide model system 10 has a higher probability
(potential) to indeed adopt a turn conformation.
Synthesis. After these promising results, we decided to study
the follow-up Ugi 4CR on the 3,4-DHP-2-one scaffold. In
principle, Freidinger lactam-like structures as 10 are available
from MCR product 5a (R1 ) R2 ) R3 ) Ph) by two possible
synthetic routes, i.e., Ugi 4CR followed by N1-alkylation or
N1-alkylation followed by Ugi 4CR. Since reaction of 5a with
The lowest energy conformer of depsipeptide 9 has a linear
conformation and is therefore not a ꢀ-turn (Figure 2A). The
(14) (a) Marshall, G. R. Curr. Opin. Struct. Biol. 1992, 2, 904–919. (b)
Giannis, A.; Kolter, T. Angew. Chem., Int. Ed. Engl. 1993, 32, 1244–1267.
662 J. Org. Chem. Vol. 74, No. 2, 2009

Synthesis of Conformationally Constrained Peptidomimetics
TABLE 2. Alkylation of the DHP-2-one 5a and Ugi Follow-Up Chemistry on Alkylated DHP-2-one 11
a Unless otherwise noted (see Supporting Information), the Ugi 4CRs were carried out using 0.88 M of DHP-2-one 11 with 1.1 equiv of carboxylic
acid and 1.5 equiv of both aldehyde and amine b Isolated yields are reported. c A 1:1 mixture of diastereomers was obtained. d Along with the product
also 45% of side product 15 was isolated. e Along with the product also 23% of side product 14 was isolated. DMB ) dimethoxybenzyl.
isobutyraldehyde, benzylamine and propionic acid did not afford
the expected Ugi product, the former route was abandoned. On
the other hand, 5a could be alkylated with tert-butyl bromo-
acetate after deprotonation with NaH in excellent yield (93%)
to give 11 (Table 2). To our delight, the Ugi 4CR involving
11, isobutyraldehyde, benzylamine, and propionic acid indeed
succeeded, affording the Ugi product 12g in nearly quantitative
yield (Entry 7, Table 2). To determine the generality of this
approach, the scope of the Ugi reaction was studied in more
detail. Isocyanide 11 was reacted with several commercially
available aldehydes, amines, and acids under standard Ugi
conditions (MeOH, rt). In this study, the DHP-2-one starting
material 11 (R1 ) R2 ) R3 ) Ph) was not varied as we have
shown this extensively in our previous work.10,11
formaldehyde (or paraformaldehyde) are rare and afford the
desired product in moderate yield at best.15 The initial products
of the reaction between formaldehyde and aromatic amines are
carbinolamines,16 which will react further to the Ugi product
most likely via an imine or iminium ion intermediate.17 Imine
formation between (para)formaldehyde and aniline is expected
to be a slow process compared to the formation of other imines
due to the limited nucleophilic character of aniline and the rather
high stability of the carbinolamine.16 This also explains the long
reaction time of this particular Ugi reaction and the formation
of side products such as the iminoimidazolidine 15. The
formation of 15 is most likely caused by relatively high
concentration of free aniline during the reaction, resulting from
slow imine formation and the rather low basicity of aniline.
The expected Ugi products 12a-k were obtained in moderate
to excellent yields (32-100%) with reaction times varying from
6 h to 7 days (Table 2). When aldehydes other than paraform-
aldehyde were applied, approximately 1:1 mixtures of diaster-
eomers were obtained. The diastereomers could only be
separated for the example applying benzaldehyde (12h, entry
8, Table 2). Table 2 shows the broad substrate scope of the
reaction, and no limitations were found in any of the compo-
nents. When paraformaldehyde and aniline (entry 9, Table 2)
were used as components, a lower yield was obtained as a result
of the formation of an iminoimidazolidine side product 15 (45%,
Scheme 3). Ugi reactions involving aniline in combination with
(15) (a) Umkehrer, M.; Kalinski, C.; Kolb, J.; Burdack, C. Tetrahedron Lett.
2006, 47, 733–737. (b) Kalinski, C.; Umkehrer, M.; Ross, G.; Kolb, J.; Burdack,
C.; Hiller, W. Tetrahedron Lett. 2006, 47, 2391–2393. (c) Maison, W.;
Schlemminger, I.; Westerhoff, O.; Martens, J. Bioorg. Med. Chem. 2000, 8, 1343–
1360.
(16) (a) Abrams, W. R.; Kallen, R. G. J. Am. Chem. Soc. 1976, 98, 7777–
7789. (b) Atherton, J. H.; Brown, K. H.; Crampton, M. R. J. Chem. Soc., Perkin
Trans. 2 2000, 941–946. The rather high stability of the carbinolamine and
therefore slow formation of the imine most likely explains the long reaction
time of this particular Ugi reaction and the formation of side products such as
the iminoimidazolidine.
(17) There is evidence that pure N-methyleneaniline occurs as a cyclic trimer.
However, since nucleophilic species are present the formation of this trimer is
not likely. See: Miller, J. G.; Wagner, E. C. J. Am. Chem. Soc. 1932, 54, 3698–
3706.
J. Org. Chem. Vol. 74, No. 2, 2009 663

Scheffelaar et al.
SCHEME 3. Formation of r-Amino Amidine and Iminoimidazolidine Side Products Using Aniline in the Ugi Reaction of 11
SCHEME 4. Synthetic Route toward Model Structures 9 and 10
The higher nucleophilicity of aniline compared to the carboxy-
late anion then results in the intermediate R-amino amidine 13
rather than the Ugi product 12i.18 Since paraformaldehyde is
quite unhindered, a subsequent ring closure is possible by
reaction of intermediate 13 with a second equivalent of
(para)formaldehyde to give 15. The formation of R-amino
amidines via this Ugi-type condensation has been reported
earlier by McFarland, who needed 1 equiv of dimethylamine
hydrochloride to facilitate the reaction.19 Later, Keung et al.
optimized this procedure by replacing the stoichiometric am-
monium salt by a catalytic amount of scandium triflate.20 The
presence of acid and free amine in the reaction above (entry 9,
Table 2) can account for observation of this side reaction.
When isobutyraldehyde, aniline, and propionic acid were
employed in the Ugi reaction, a higher yield of Ugi product
12j was obtained (58%), albeit accompanied by R-amino
amidine 14 (23%). In this case, cyclization to the corresponding
iminoimidazolidine product was not observed, most likely
because isobutyraldehyde is too sterically hindered. In conclu-
sion, in the majority of combinations that were tested the Ugi
reaction can be performed successfully after alkylation of the
DHP-2-one scaffold 5a.
To investigate the actual turn-inducing capacities of the DHP-
2-one scaffold, model structures 9 and 10 were synthesized
(Scheme 4) using the MCR-alkylation-MCR sequence. Thus
after initial 4CR and alkylation, chemoselective condensation
of the isocyano function in DHP-2-one 11 using a Passerini
3CR reaction with paraformaldehyde and propionic acid gave
16 in quantitative yield. tert-Butyl deprotection and subsequent
EDC/HOBt-mediated coupling proceeded smoothly in 65%
overall yield to afford constrained depsipeptide 9. A similar
strategy employing a 4CR-alkylation-Ugi 4CR sequence gave
(18) The formation of this intermediate could be proven by stirring 11 with
paraformaldehyde (3 equiv) and aniline (3 equiv) to give 45% of 15 and 10% of
amidine 13 (reaction time 77) h). When 1.0 equiv of Et3N · HCl was added, the
amidine was not present after 77 h and the yield of 15 is 73%. During the reaction,
however, 13 was observed on TLC, supporting its involvement as an intermediate.
(19) McFarland, J. W. J. Org. Chem. 1963, 28, 2179–2181.
(20) Keung, W.; Bakir, F.; Patron, A. P.; Rogers, D.; Priest, C. D.;
Darmohusodo, V. Tetrahedron Lett. 2004, 45, 733–737.
664 J. Org. Chem. Vol. 74, No. 2, 2009

Synthesis of Conformationally Constrained Peptidomimetics
SCHEME 5. Synthesis of Pentapeptide Mimics 19a-d
the Ugi product 12d (entry 4, Table 2), which was subsequently
subjected to tert-butyl deprotection and subsequent EDC/HOBt-
mediated coupling to obtain the corresponding tetrapeptide
mimic 10. The DMB group was removed simultaneously with
the tert-butyl group under acidic conditions. A different solvent
system was required (MeCN/DMF 5:1 instead of CH2Cl2) to
dissolve the free acid for the coupling reaction to allow the
isolation of 10 in 70% yield.
Our modular MCR-alkylation-MCR strategy can also be
used to directly attach peptide moieties to both positions of
the DHP-2-one scaffold. This further improves the flexibility,
ease of handling, and general applicability of our method to
arrive at potential turn mimics based on Freidinger-type lactams.
Thus, alkylation of 5a with BrCH2C(O)-Gly-OBn (17) cleanly
gave isocyanide 18, which could be used in subsequent Ugi or
Passerini reactions to give constrained pentapeptides 19a and
19b and depsipeptide 19c (Scheme 5). The DMB group in 19b
could then be cleaved by treatment with TFA to give 19d in
70% yield.
calculations. This Eaq appeared to be -132.3 kcal/mol, which
is 4.4 kcal/mol above the modeled lowest energy conformer
(Eaq ) -136.72 kcal/mol, see also Table 1) of 10.
The two somewhat contradictory results of the modeling
studies and the X-ray crystal structure can be rationalized by
the different media in which the experiments are performed.
The modeling experiments deal with gas phase conformations
and therefore intermolecular interactions are not taken into
account. On the other hand, in the X-ray crystal structure
packing forces become important, which is of course facilitated
by intermolecular hydrogen bonds and which may lead to a
change in secondary structure. Intermolecular forces are of
primary importance in crystal formation.21 Both methods by
definition do not give the correct representation of the secondary
conformation of 10 in solution, so further solution phase studies
by NMR (low concentration) were required.
Conformational Analysis. The results from the computa-
tional modeling studies showed that the DHP-2-one core in both
the depsipeptide 9 and the tetrapeptide 10 indeed could serve
as a turn-inducing element and that for both model systems a
ꢀ-turn (type IV) is a viable conformation. With 9 and the 10 in
hand we now decided to undertake a detailed spectroscopic study
on the secondary structure of these model systems.
X-ray Crystal Structure Determination. We were able to
produce crystals of the tetrapeptide 10 that were of sufficient
quality for a detailed X-ray diffraction. Figure 3 shows the
molecular structure of 10 in the crystal, clearly adopting a turn
conformation. However, no intramolecular hydrogen bond is
present between COi and NHi+3. Instead, there is an intramo-
lecular hydrogen bond between Ni+1 and COi+1 and the other
two amide protons (NHi and NHi+3) participate in a one-
dimensional chain in the direction of the crystallographic a-axis
by intermolecular hydrogen bonds. The presence of the in-
tramolecular hydrogen bond between Ni+1 and COi+1 excludes
the possibility of an intramolecular hydrogen bond between COi
and NHi+3. The peptide chains are rotated to such an extent
that the interstrand distance becomes to long for hydrogen
bonding contacts. An important criterion for ꢀ-turns is that the
CRi-CRi+3 distance should be smaller than 7 Å. The X-ray
crystal structure shows a CRi-CRi+3 distance of 8.949(2) Å,
and although 10 clearly adopts a turn structure, it can not be
classified as a ꢀ-turn.
FIGURE 3. Displacement ellipsoid plot of 10. Drawn at 50% prob-
ability level. The compound was crystallized from a benzene/CH2Cl2
mixture (1:1) by slow evaporation as a racemate in the centrosymmetric
space group P21/c. Only N-H hydrogen atoms of the amide groups
are drawn, for clarity. d(CRi-CRi+3) ) 8.949(2) Å; torsion angles with
standard uncertainties: ꢁi+1 ) 175.75(14)°, Ψi+1 ) 167.27(12)°, ꢁi+2
) -68.44(16)°, Ψi+2 ) 159.23(12)°; Eaq) -132.3 kcal/mol.
To compare these results to the modeling results on the
conformation of 10 described above, the energy (Eaq) of the
conformation in the X-ray crystal structure was calculated using
the same geometry optimization//AM1-SM5 single point energy
FIGURE 4. Model compounds 9 and 10 used for the conformational
analysis including amide proton numbering.
J. Org. Chem. Vol. 74, No. 2, 2009 665

Scheffelaar et al.
FIGURE 5. Chemical shift values at different temperatures for the different amide protons of model compounds 9 (A) and 10 (B).
TABLE 3. Chemical Shifts in CDCl3, Solvent Effect (in Brackets), and Temperature Coefficients of Amide Protons in 9 and 10a
compd
δΗ1 (ppm)
δΗ2 (ppm)
δΗ3 (ppm)
∆δH1/∆T (ppb/K)b
∆δH2/∆T (ppb/K)b
∆δH3/∆T (ppb/K)b
9
10
7.80 (0.14)
7.39 (0.49)
5.88 (2.34)
5.92 (2.32)
-2.22 to -0.98c
-1.21 to -0.42c
5.86 (2.16)
-2.38 to -0.05d
-0.94 to -0.53e
-3.57 to -2.88e
a All NMR experiments were performed using 1.0-2.0 mM CDCl3 solutions; chemical shift values (δ) reported are the values at 300 K. Data in
brackets correspond to δ(DMSO) - δ(CDCl3) (ppm, 300 K). b Temperature coefficients were determined between all successive data points; highest and
lowest values are reported. c Temperature coefficients were determined between 233 and 330 K. d Temperature coefficient were determined between 233
and 300 K. e Temperature coefficient were determined between 294 and 330 K.
1H NMR Studies. The conformational behavior of model
structures 9 and 10 (Figure 4) in solution was also investigated
using 1-D and 2-D NMR techniques.
The temperature coefficient for H1 of both compounds is very
small (<3 ppb/K, in absolute value), which supports (together
with the high chemical shift values) a hydrogen bonded state.
For compound 10, signals for H2 and H3 overlap significantly
in the 233-283 K range. Accordingly, temperature coefficients
could be most accurately determined for the 294-330 K range,
and these values are reported in Table 3.26 The coefficient for
H2 of both compounds is small, but since the chemical shift
value is below 6.0 ppm this indicates a non-hydrogen-bonded
state. The H3 amide proton of 10 has a slightly higher
temperature coefficient (average of -3 ppb/K), which may
indicate a small amount of hydrogen bonding. However,
considering the chemical shift value of this proton, this is most
likely very small and negligible.
A third indication of intramolecular hydrogen bonding is the
effect on the chemical shift upon switching from the non-
hydrogen-bonding solvent CDCl3 to the hydrogen-bond-acceptor
solvent DMSO. Strongly hydrogen-bonded amide protons are
not accessible to the solvent and have a small chemical shift
difference (∆δ < 0.2 ppm in absolute value). Non-hydrogen-
bonded amide protons are accessible to the solvent and have a
large chemical shift difference (∆δ > 0.2 ppm in absolute value,
usually larger than 1.0 ppm).24 The solvent effect on the
chemical shifts of protons H1, H2, and H3 of compounds 9 and
10 are given in brackets (Table 3).24 The chemical shift
difference of H1 in 9 is small (∆δ ) 0.14 ppm in absolute value)
when the solvent is changed to DMSO, but a large chemical
shift difference is observed for H2 (∆δ ) 2.34 ppm) indicating
Full assignment of the different methylene and the amide NH
protons was achieved using gs-HMBC and gs-NOESY NMR
1
measurements. 2D H NOESY NMR data did not suggest the
close proximity of the two peptide chains, since no interstrand
correlations, but only intrastrand contacts, were observed.
The presence of intramolecular hydrogen bonding was
determined by evaluation of the chemical shift values of the
amide NH protons and their temperature coefficients (∆δΗ/
∆T). Hydrogen-bonded amide protons have a high chemical shift
value (δ g 7.0 ppm) and display a very small temperature
coefficient (0 to -3.0 ppb/K). Non-hydrogen-bonded protons
usually show a low chemical shift value (δ e 7.0 ppm) and
also a very small temperature coefficient (0 to -3.0 ppb/K).
Large temperature coefficients (between -4 and -8 ppb/K)
indicate equilibrium between hydrogen-bonded and non-hydrogen-
bonded states.22
Figure 5 shows the temperature dependence of the chemical
shifts of the amide protons of 9 and 10. Spectra were recorded
from 224 to 330 K with increments of approximately 10 K using
1-2 mM solutions in CDCl3 to avoid intermolecular interac-
tions.23 The chemical shift values of the H1 amide protons of
both compounds appear downfield (δ > 7.0 ppm), which is an
indication of hydrogen bonding. On the other hand, amide
protons H2 of both compounds and H3 of 10 appear upfield
(δ < 6.0 ppm), indicative for a non-hydrogen-bonded state.
Table 3 shows the chemical shifts (ppm) values at 300 K
and temperature coefficients (ppb/K) for the amide NH protons
of 9 and 10. Temperature coefficients were determined between
all successive data points (highest and lowest values are reported
in Table 3).24,25
(24) For all data, see Supporting Information.
(25) In general, the coefficients have the largest values at lower temperatures
and become more or less constant at room temperature, which can be explained
by an increased amount of intermolecular hydrogen bonding at lower temper-
atures. Specifically high coefficients are found in the 224-233 K interval,
accompanied with line broadening of the 1H NMR signals (especially for 10).
These values are therefore not reliable and were excluded from the table.
(26) Although the signals for H2 and H3 overlap in the 233-283 K range,
it is clear from Figure 5B that the temperature coefficients for both protons are
significantly larger than in the 294-330 K range. However, it must be noted
that high temperature coefficients are especially observed in the low temperature
range, where line broadening (possibly due to aggregation) is observed.
(21) See ref 6a and references therein.
(22) Belvisi, L.; Gennari, C.; Mielgo, A.; Potenza, D.; Scolastico, C. Eur. J.
Org. Chem. 1999, 389–400.
(23) Boussard, G.; Marraud, M. J. Am. Chem. Soc. 1985, 107, 1825–1828.
666 J. Org. Chem. Vol. 74, No. 2, 2009

Synthesis of Conformationally Constrained Peptidomimetics
a non-hydrogen-bonded situation. Amide proton H1 of the Ugi
derivative 10 has a slightly higher solvent coefficient (∆δ )
0.49 ppm in absolute value) indicating that this proton is
accessible to the solvent. Moreover, both amide protons H2 and
H3 have very large solvent coefficients and are therefore clearly
non-hydrogen-bonded. According to the chemical shift value
and temperature coefficient experiments H1 was locked in a
hydrogen bonded conformation (in CDCl3). However, apparently
pure DMSO is a too strong competitive hydrogen bond acceptor.
DMSO titration studies in CDCl3 indicated that the chemical
shift of H1 in 10 is constant up to 20% of DMSO and can
amine (0.33 mmol), and a carboxylic acid (0.24 and 0.22 mmol
when 2,4-dimethoxybenzylamine is used28). The mixture was stirred
at rt until the reaction was completed (determined by TLC). The
resulting mixture was concentrated in vacuo and purified by column
chromatography.
General Procedure II for the Passerini Reaction on
Alkylated DHP-2-ones. To a solution of alkylated DHP-2-one 11
(0.22 mmol) in CH2Cl2 (367 µL) were added an aldehyde (0.33
mmol) and a carboxylic acid (0.24 mmol). The mixture was stirred
at room temperature until the reaction was completed (determined
by TLC). The resulting mixture was concentrated in vacuo and
purified by column chromatography.
therefore be considered as hydrogen-bonded.24 All of the H
1
Alkylated DHP-2-one 11. To a solution of NaH (60 w%, 126
mg, 3.1 mmol) in dry DMF (20 mL) at 0 °C was added dropwise
a solution of DHP-2-one 5a (1.00 g, 2.9 mmol) in dry DMF (10
mL). After stirring for 75 min. at 0 °C, tert-butyl bromoacetate
(0.51 mL, 3.4 mmol) was added. The mixture was stirred and slowly
warmed to room temperature overnight, after which the reaction
was concentrated in vacuo. The residue was dissolved in EtOAc
(200 mL), washed with H2O (2 × 150 mL) and brine (100 mL),
and dried over MgSO4. The crude product was purified by column
chromatography (c-hexane/EtOAc 9:1 f 8:2) to afford 11 (1.26 g,
2.7 mmol, 93%) as a white foam. 1H (250 MHz, CDCl3): δ (ppm)
experiments described above indicate that amide protons H1 of
both model compounds 9 and 10 are involved in intramolecular
hydrogen bonds. However, no intramolecular hydrogen bond
is present between the two peptide chains. This is in agreement
with the X-ray crystal structure and therefore these Freidinger-
type peptidomimetics most likely do not fulfill the criteria of a
ꢀ-turn. In fact, the various NMR studies all suggest a conforma-
tion very similar to that in the X-ray crystal structure.
7.36-7.14 (m, 15H), 5.54 (d, J ) 4.5 Hz, 1H), 4.40 (d, JAB
)
Conclusion
17.3 Hz 1H), 4.32 (d, J ) 4.5 Hz, 1H), 3.96 (d, JAB ) 17.3 Hz
1H), 1.42 (s, 9H). 13C NMR (63 MHz, CDCl3): δ (ppm) 167.3 (C),
165.0 (C), 162.6 (C), 142.2 (C), 135.5 (C), 135.4 (C), 134.5 (C),
129.5 (2 CH), 129.4 (CH), 128.9 (3 CH), 128.5 (4 CH), 128.3 (CH),
128.1 (2 CH) 126.8 (2 CH), 111.2 (CH), 82.5 (C), 70.4 (C), 50.9
(CH), 47.8 (CH2), 28.1 (3 CH3). IR (KBr): 2977 (w), 2931 (w),
2360 (w), 2134 (m), 1741 (s), 1699 (s), 1372 (s), 1230 (s), 1153
(s), 759 (m), 687 (s). HRMS (EI, 70 eV): calcd for C30H28N2O3
(M+) 464.2094, found 464.2090.
Although the MM-SYBYL-AM1 analysis indicated realistic
type IV ꢀ-turn conformations for our model systems, the
spectroscopic data only confirm that the DHP-2-one scaffold
does induce a turn, albeit not a true ꢀ-turn conformation.
However, the rigidification introduced by the ring, the presence
of the intramolecular hydrogen bond between Ni+1 and COi+1
,
and the bulky phenyl substituents decrease the rotational
freedom of the peptide chains to such an extent that this Phe-
Gly dipeptide isoster can ideally be applied for the development
of conformationally constrained peptidomimetics.4h
Ugi Product 12a. According to General Procedure I, reaction
between alkylated DHP-2-one 11 (100 mg, 0.22 mmol), paraform-
aldehyde (10 mg, 0.33 mmol), benzylamine (36 µL, 0.33 mmol),
and propionic acid (18 µL, 0.24 mmol) in MeOH (500 µL) afforded
after 22 h 12a (143 mg, 0.22 mmol, 100%) as a white foam. Column
chromatography was performed with c-hexane/EtOAc 7:3 f 6:4.
1H (400 MHz, DMSO, 403 K): δ (ppm) 7.80 (bs, 1H), 7.67-7.63
(m, 2H), 7.44-7.40 (m, 2H), 7.38-7.20 (m, 12H), 7.07-7.02 (m,
2H), 7.01-6.97 (m, 2H), 5.68 (d, J ) 7.2 Hz, 1H), 5.21 (d, J )
7.2 Hz, 1H), 4.28 (d, JAB ) 15.7 Hz, 1H), 4.14 (d, JAB ) 15.7 Hz,
1H), 4.04 (d, JAB ) 17.1 Hz, 1H), 3.91 (d, JAB ) 17.1 Hz, 1H),
3.65 (s, 2H), 2.18-1.97 (m, 2H), 1.37 (s, 9 H), 0.95 (t, J ) 7.4
Hz, 3H). 13C NMR (101 MHz, DMSO, 403K): δ (ppm) 173.0 (C),
168.9 (C), 166.1 (C), 165.8 (C), 138.9 (C), 138.6 (C), 137.1 (C),
136.6 (C), 133.9 (C), 128.2 (2 CH), 127.9 (CH), 127.7 (2 CH),
127.7 (2 CH), 127.6 (2 CH), 126.9 (CH), 126.8 (4 CH), 126.6 (CH),
126.5 (4 CH), 126.3 (CH), 111.8 (CH), 80.9 (C), 63.1 (C), 49.4
(CH2), 49.2 (CH2), 46.6 (CH2), 41.9 (CH), 27.0 (3 CH3), 24.6 (CH2),
8.3 (CH3). IR (neat): 3362 (w), 2978 (w), 1742 (m), 1663 (s), 1495
(m), 1366 (m), 1225 (m), 1152 (s), 762 (m), 696 (s), 590 (m), 519
(m). HRMS (EI, 70 eV): calcd for C41H43N3O5 (M+) 657.3197,
found 657.3211.
Pentapeptide Mimic 19a. According to General Procedure II,
reaction between alkylated DHP-2-one 18 (100 mg, 0.18 mmol),
paraformaldehyde (8 mg, 0.27 mmol), benzylamine (30 µL, 0.27
mmol), and (Z)-Gly-OH (41 mg, 0.20 mmol) in MeOH/CH2Cl2 5:1
(1.2 mL)29 afforded after 26 h 19a (117 mg, 0.13 mmol, 72%) as
a white foam. Column chromatography was performed with
c-hexane/EtOAc 6:4 f 1:1. 1H (400 MHz, DMSO, 403 K): δ (ppm)
7.84 (s, 1H), 7.74-7.66 (m, 3H), 7.56-7.53 (m, 2H), 7.38-7.22
(m, 22H), 7.08-7.03 (m, 2H), 6.99-6.96 (m, 2H), 6.57 (bs, 1H),
In addition to the introduction of constraints in the peptide
backbone, the turn-like structure induced by the DHP-2-one
scaffold results in a constrained peptide that shows an excellent
preorganization for macrocyclization. Currently we are inves-
tigating the use of this dipeptide isoster in the synthesis of
constrained macrocyclic peptides. The incorporation of con-
strained elements in cyclic peptides has proven a useful method
for further rigidification of the peptide, which often results in
an improved receptor affinity and/or selectivity.27 Also, these
constrained derivatives proved to be important for SAR studies.
An especially attractive feature of the described synthetic
approach is the highly modular character. In contrast to many
existing Freidinger-type turn mimics, our MCR-alkylation-MCR
strategy provides rapid access to tetra- and pentapeptidic turn
mimics with possibilities for structural variation. It allows the
rapid generation of diversely substituted dihydropyridones, via
an initial MCR, to which a broad range of peptidic bromo-
acetamides can be attached by simple amide alkylation. Finally,
a second peptide moiety is readily introduced by second MCR,
such as the Ugi reaction, allowing additional diversification. This
unique combination of complexity and diversity generation make
this method ideally suitable for lead discovery and optimization.
Experimental Section
General Procedure I for the Ugi Reaction on Alkylated
DHP-2-ones. To a solution of alkylated DHP-2-one 11 (0.22 mmol)
in MeOH were subsequently added an aldehyde (0.33 mmol), an
(28) Partial cleavage of the DMB group took place under acidic conditions,
and therefore 1 equiv of acid was used in the reactions using 2,4-dimethoxy-
benzylamine.
(27) Cheng, R. P.; Suich, D. J.; Cheng, H.; Roder, H.; DeGrado, W. F. J. Am.
Chem. Soc. 2005, 123, 12710–12711, and references therein.
(29) CH2Cl2 was added to increase the solubility of the dihydropyridone
starting material, since it is very poorly soluble in MeOH.
J. Org. Chem. Vol. 74, No. 2, 2009 667

Scheffelaar et al.
5.60 (d, J ) 7.1 Hz, 1H), 5.18 (d, J ) 7.1 Hz, 1H), 5.15 (s, 2H),
5.07 (s, 2H), 4.25 (d, JAB ) 15.7 Hz, 1H), 4.18 (d, JAB ) 15.7 Hz,
1H), 4.18 (d, JAB ) 19.3 Hz, 1H), 3.90 (d, J ) 5.8 Hz, 2H), 3.87
(d, JAB ) 19.3 Hz, 1H), 3.84-3.61 (m, 2H), 3.73 (d, JAB ) 17.2
Hz, 1H), 3.64 (d, JAB ) 17.2 Hz, 1H). 13C NMR (101 MHz, DMSO,
403 K): δ (ppm) 168.7 (2 C), 168.5 (C), 166.8 (C), 165.4 (C), 155.4
(C), 139.2 (C), 138.9 (C), 137.2 (C), 136.6 (C), 136.0 (C), 135.4
(C), 134.1 (C), 128.5-126.4 (30 CH), 111.2 (CH), 65.3 (CH2),
65.0 (CH2), 63.3 (C), 49.3 (CH2), 48.5 (CH2), 46.5 (CH2), 42.1
(CH), 41.5 (CH2), 40.3 (CH2). IR (neat): 3347 (w), 3063 (w), 2945
(w), 1719 (m), 1651 (s), 1495 (m), 1445 (m), 1387 (m), 1341 (m),
1244 (m), 1186 (s), 1179 (s), 1030 (w), 959 (w), 737 (m), 696 (s),
590 (w). HRMS (FAB): calcd for C53H50N5O8 (MH+) 884.3654,
found 884.3637.
Pentapeptide Mimic 19d. To a solution of 19b (90 mg, 0.10
mmol) in CH2Cl2 (0.5 mL) was added TFA (0.5 mL). The mixture
was stirred for 45 min at room temperature followed by removal
of the solvent in vacuo and co-evaporation with toluene. The crude
product was purified by column chromatography (EtOAc/c-hexane
7:3 f 8:2) to obtain 19d (55 mg, 0.07 mmol, 70%) as a white
foam.1H (400 MHz, CDCl3): δ (ppm) 7.68 (bs, 1H), 7.63 (d, J )
7.2 Hz, 2H), 7.40-7.18 (m, 21H), 7.02 (d, J ) 6.8 Hz, 2H), 6.59
(bs, 1H), 6.34 (bs, 1H), 5.71 (bs, 1H), 5.65 (d, J ) 7.2 H, 1H),
5.20 (d, J ) 6.8 Hz, 1H), 5.14 (s, 2H), 5.10 (s, 2H), 4.01-3.87
(m, 4H), 3.80-3.57 (m, 4H). 13C NMR (101 MHz, CDCl3): δ (ppm)
170.4 (C), 169.5 (C), 169.0 (C), 167.4 (C), 167.0 (C), 156.6 (C),
140.7 (C), 138.4 (C), 136.7 (C), 136.3 (C), 135.3 (C), 134.4 (C),
129.1 (CH), 128.9 (4 CH), 128.8 (4 CH), 128.7 (3 CH), 128.5 (3
CH), 128.3 (3 CH), 128.2 (2 CH), 128.0 (3 CH), 127.3 (2 CH),
112.4 (CH), 67.3 (2 CH2), 64.4 (C), 48.2 (CH2), 44.3 (CH2), 43.0
(CH2), 42.2 (CH), 41.3 (CH2). IR (neat): 3337 (w), 3063 (w), 1719
(m), 1655 (s), 1651 (s), 1497 (s), 1389 (m), 1240 (m), 1215 (m),
1190 (m), 1036 (w), 959 (w), 754 (s), 698 (s). HRMS (FAB): calcd
for C46H44N5O8 (MH+) 794.3184, found 794.3192.
Acknowledgment. This work was performed with financial
support from the Dutch Science Foundation (NWO, VICI grant).
Dr. H. Peeters (University of Amsterdam, The Netherlands) is
kindly acknowledged for conducting HRMS measurements.
Supporting Information Available: Detailed experimental
methods and characterization data (1H and 13C NMR, IR,
HRMS) for all new compounds, details of the modeling studies,
crystallographic data for 10 in CIF format, and data of the VT
and DMSO titration NMR experiments. This material is
available free of charge via the Internet at http://pubs.acs.org.
JO802052J
668 J. Org. Chem. Vol. 74, No. 2, 2009