INSIGHTS | PERSPECTIVES
the deformation is nonreversible, because
the ionic aggregates have rearranged via ligand exchange to accommodate the strain,
and after the stress is removed, the restoring
force from the PEG-based network is insufficient to drive ligand exchange.
This dynamic ligand exchange requires
a high concentration of ionic aggregates
and mobile charges, which points toward
the importance of using catechol-Fe3+ assemblies to produce the desired properties.
As the authors note, water weakens these
ionic interactions and, unlike traditional
ionomers with hydrophobic matrices, the
PEG in these materials is hygroscopic.
Nonetheless, the principle has been demonstrated that adding a dense array of ionic
aggregates to an elastomeric network will
simultaneously improve a variety of mechanical properties.
Moving forward, there are opportunities to explore a vast array of chemistries
to tune the physical properties of the cross-linked network and the nature of the ionic
aggregates. Filippidi et al. have designed a
versatile epoxy-based chemistry that incorporates iron-catechol interactions into an
elastomeric network. This combination of a
permanent covalent network and transient
ionic aggregates produces stiffer, stronger,
and tougher networks. Variations on this
platform include substituting the covalent
network with any of the established elastomers and might even be amenable to long-lived cross-links based on DNA and protein
assemblies. Substitutions for the stress-responsive, transient ionic aggregates might
have more limited choices because of the
requirements of high number density, although functionalized nanoparticles that are
smaller than the mesh size of the network
might prove useful. j
1. E. Filippidi et al ., Science 358, 502 (2017).
2. D. E. Discher, P. Janmey, Y.-l. Wang, Science 310, 1139
3. L. R. Middleton, K. I. Winey, Annu. Rev. Chem. Biomol. Eng.
8, 499 (2017).
4. C. F. Buitrago et al ., Macromolecules 48, 1210 (2015).
5. A. Eisenberg, J.-S. Kim, Introduction to Ionomers ( Wiley,
New York, 1998).
6. K. Wakabayashi, R.A.Register, Macromolecules 39,1079
7. L. R. Middleton et al ., Macromolecules 48, 3713 (2015).
A peptide mimic of an antibody
Antibody miniaturization leads to a cyclic peptide that
neutralizes influenza viruses
By Timothy A. Whitehead1,2,3
Influenza virus is one of the largest public health concerns. There is no universal vaccine, and only a few small- molecule drugs are available for therapy and prevention of influenza virus infec- tion. The surface of this virus has two
major proteins: neuraminidase and hemagglutinin (HA). These proteins are targets
for antivirals. Although antivirals such as
oseltamivir exist that bind neuraminidase,
a major reason that there is no small-molecule drug against HA is because most of
its surface is highly variable, presenting a
moving target for drug development. A major breakthrough in developing universal
vaccines and broad therapies came almost
a decade ago from the identification of a
site of vulnerability on influenza HA revealed through analyses of human broadly
neutralizing antibodies (antibodies that are
able to bind and neutralize many different
influenza subtypes) (1). On page 496 of this
issue, Kadam et al. (2) present the design
and iterative optimization of a cyclic peptide that targets the vulnerable site on HA.
This is the first example of such a small
peptidic molecule that can be viewed as an
Antiviral therapy by use of monoclonal
antibodies involves intravenous (iv) injection on the order of 100 mg of antibody
and a recommended hospitalization stay.
By contrast, peptides are smaller than proteins and are attractive as therapeutics because they can be easily manufactured and
have greater potential for oral bioavailability than that of larger biologics (3). Thus,
an orally delivered peptide that is able to
replicate the exquisite potency and specificity of an antibody would be attractive as
an antiviral therapeutic or prophylactic.
Influenza A and influenza B cause sea-
sonal epidemics in humans; several influ-
enza A subtypes have pandemic potential.
The HA protein contains a stem region that
is reasonably conserved among both influ-
enza A subtypes and influenza B. Binding to
this stem site by antibodies can effectively
neutralize virus in vitro and in vivo (1).
The first clue that this HA stem site
could be targeted by smaller molecules
came from a study in which small protein
binders to this site were computationally
designed (see the figure) (4). These proteins bound multiple HA subtypes with
high affinity but had very poor stability in plasma and thus are not useful as
therapeutics. Kadam et al. extended the
miniaturization concept illustrated by the
computational design to a minimal 11-resi-
due cyclic peptide to produce a small anti-influenza cyclic peptide, P7.
The elegant iterative workflow used
by Kadam et al. is a master class in improving molecular recognition by using
structural information. First, they used a
peptide sequence borrowed from a rare
broadly neutralizing antibody that inhibits all HA subtypes of influenza A. Binding
of this antibody is mediated mainly by a
single hypervariable loop in the antibody
(5). Although there are decades-old reports
of deriving peptides from antibody loops
(6), these studies did not have the extensive structural information used by Kadam
et al. Simple substitutions of aliphatic
noncontacting residues to charged amino
acids were used to enhance the aqueous
solubility of the peptide. Second, the team
experimented with different cyclization
chemistries in order to constrain the peptide in the bound conformation. The rationale behind this strategy is that the bound
peptide conformation incurs an entropic
penalty upon binding; constraining the
peptide at or near the bound conformation
minimizes this penalty.
Lastly, Kadam et al. made extensive use
of non-natural amino acids to improve
the affinity and heterosubtypic breadth
of binding to HA. Most notably, the original antibody loop contained a positively
charged amino acid, arginine, that had
weak electrostatic interactions with a negative patch on the HA surface. The authors
replaced this arginine with the hydrophobic 5-phenyl norvaline (XA) to “reach”
into an adjacent aliphatic pocket on HA,
improving van der Waals contacts with a
concomitant increase in binding affinity.
XA differs from the natural amino acid
phenylalanine by the addition of a single
1Department of Chemical Engineering and Materials Science,
Michigan State University, East Lansing, MI, USA. 2Institute for
Quantitative Health Science and Engineering, Michigan State
University, East Lansing, MI, USA. 3Department of Biosystems
and Agricultural Engineering, Michigan State University, East
Lansing, MI, USA. Email: email@example.com
“There are many compelling
reasons to control the
mechanical properties of
networks and gels…”