718
M. K. P. Jayatunga et al. / Bioorg. Med. Chem. Lett. 24 (2014) 717–724
pioneered the use of b-peptides to mimic
a-helical secondary
1-sided helix PPI 2-sided helix PPI 3-sided helix PPI
250
structure,47–49 with Schepartz later demonstrating inhibition of
the p53/MDM2 PPI.50 An example by Gellman showed an extended
helix which accurately mimicked the native sequence, and success-
fully inhibited the Bcl-2/Bim PPI.51
200
150
100
50
A fundamentally different strategy to mediate PPIs is to mimic
elements of protein secondary structure with entirely non-peptidic
constructs. The rational design of helix mimics aims to project
functional groups from a scaffold in order to reproduce the orien-
tation of the native side-chains, whilst avoiding the expression,
purification or chemical synthesis of peptidic elements. Such pep-
tidomimetics present an opportunity to design compounds which
have the same binding mode as the native protein, yet are meta-
bolically stable, can permeate membranes and are readily
absorbed.11 The structural diversity provided by chemical synthe-
sis allows a much greater array of groups, with varied steric and
electronic properties, to be projected than those provided by the
side chains of proteogenic amino acids. Majmudar et al. identified
phenol carboxylic acid natural products, sekikaic- and lobaric acid,
that inhibit the PPI between p300 and both MLL (Mixed-Lineage
Leukemia) and pKID (phosphorylated kinase-inducible domain of
CREB). These molecules bind at the same site as the helical region
of peptides, hinting that Nature has evolved its own helix
mimics.52
Early work focused on the mimicry of hydrophobic residues on
a single helical face, yet important targets often have amphiphilic
helices with hot-spot regions projected from multiple faces. Impor-
tant helical sequences with hot-spot residues in the i, i + 4 and i + 7
positions were frequently targeted since they are found on a com-
mon face (Fig. 1).14 Several classes of helix mimetic have been
developed, many inspired by the terphenyl scaffold introduced
by Orner et al. in 2001.53 The p53/MDM2 and Bcl-xl/Bak interac-
0
1990-1995
1996-2000
2001-2005
2006-2011
Figure 2. PDB structures with helices at the PPI interface over time. Subcategories
illustrate which proportion contained helices with one, two or three key binding
domains.15,16
Bcl-2/Bid interaction, resulting in decreased growth of human leu-
kemia xenografts in vivo.28 Strategies developed by Cabezas and
Arora involve replacement of the internal a-helix hydrogen bond-
ing network with a covalent linkage.29–32 A number of different
approaches have since been developed to introduce constraints
in an effort to promote helicity (Fig. 3).33,34 Arora et al. demon-
strated inhibition of the HIF-1a/p300 PPI with an olefin-stapled
hydrogen bond surrogate (HBS) where key residues are found at
the i, i + 2, i + 5, i + 6 and i + 10 positions.35 This strategy has the
benefit of retaining the i, i + 4 and i + 7 residues and avoiding the
use of synthetically challenging
a,a-di-substituted amino acids.
In another strategy, tethers have been used to staple long chains
and stabilize extended helical structures. An impressive example
is a 36 amino acid anti-viral sequence stabilized by two hydrocar-
bon bridges formed through ring-closing metathesis (RCM),
increasing potency of a peptide inhibitor, enfuvirtide, four-fold to
2.1 nM and at the same time improving its pharmacokinetics.36
Short peptide truncates are frequently unstructured, prompting
researchers to explore the helical propensities of more metaboli-
cally stable b-peptides.37,38 Variations of this strategy have been
employed to promote additional helicity, with many groups
exploiting conformationally constrained variants,39–42 and mixed
tions are two well-studied targets, with
a-helices projecting
crucial residues on a single face.54,55 Both interactions inhibit cel-
lular apoptosis: MDM2 binds and inactivates p53, the principal
architect of cell-death, thereby arresting apoptosis.56 Similarly,
Bak is a pro-apoptotic protein whose function is disrupted on bind-
ing of Bcl-xl (Fig. 4a).57 Both interactions are thought to be prom-
ising cancer targets since neoplastic cells often disable intracellular
apoptotic mechanisms. Yin et al. designed terphenyl helix mimet-
ics that inhibited the p53/HDM2 interaction in vitro with a Ki of
sequences of
a
/b-amino acids.43–46 Seebach and Gellman
182 nM where a native peptide truncate had a Ki of 3.51 l
M.54
(a)
(b)
Arora has highlighted the importance of binding to multiple
surfaces of a single helix in many interactions.15 The current
collection of single-sided helix mimetics are ill-equipped to simul-
taneously mimic hot-spot residues displayed on more than one
face, presenting a need for the development of more complex scaf-
folds (Fig. 4).58 Moreover, the restricted length of mimics of two
a
-helix turns is often insufficient to inhibit interactions of ex-
tended helices. The Bcl-2/Bid interaction is an example with key
residues orientated on two faces of the
-helix.59 Key hydrophobic
a
residues Ile82, Ile86 and Leu90 lie on the same face, with Asp95
projected on a different face to form important hydrogen bonds
to Bcl-2 arginine and asparagine residues (Fig. 4b). Enzymes E1
and E2, involved in the replication of papillomavirus, are another
(c)
(d)
important cancer target displaying a two-faced
a-helix binding
mode.60
Examples with hot-spot residues found on all three faces of an
-helix include the interaction of calmodulin with CaM kinase I.61
a
Peptide truncates show that a particularly dense array of residues
bound in the N-terminal domain of calmodulin is crucial, with the i
(Val), i + 1 (Arg), i + 2 (His), i + 3 (Met) and i + 4 (Arg) positions of
the peptide contributing strongly to binding (Fig. 4c).
Burgess has reasoned that since there is homology in the way
many secondary structural elements display residues, it is possible
to design ‘universal peptidomimetics’, able to reproduce the
Figure 3.
bonding, (b) stapled
a
-Helix tethering strategies; (a) canonical
a-helix with 13-membered H-
a-helix through residues on the i and i + 4 positions (22-ring
macrocycle), (c) Cabezas’ hydrazone HBS (13-ring macrocycle), (d) Arora’s olefin
HBS (13-ring macrocycle). HBS = hydrogen bond surrogate.