Rational Design of Peptide-Based Biosupramolecular SystemsRational Design of Peptide-Based...

Rational Design of Peptide-BasedBiosupramolecular Systems

Aimee L. Boyle and Derek N. WoolfsonUniversity of Bristol, Bristol, UK

1 Introduction 16392 Repetitive Extended Supramolecular Systems 16433 Discrete Supramolecular Systems 16504 Templated and Directed Assemblies 16555 Conclusions and Outlook 1661References 1662

1 INTRODUCTION

A highly attractive feature of peptides and proteins for usein what might be termed “biosupramolecular chemistry” istheir ability to self-assemble in water with exquisite speci-ficity and high affinities through noncovalent interactions.An aim of peptide design and engineering is to exploit thisto create new supramolecular assemblies from the bottomup.1, 2 Within this, one approach is to program simple pep-tide and protein building blocks (which might be referredto as “tectons”) to fold and self-associate in prescribedways. This de novo design route contrasts with top-downapproaches, which are concerned with protein engineeringand the modification of already existing and even assemblednatural proteins. The bottom-up approach has the potentialadvantage that it allows for full design and control overfolding, assembly, size, and shape of the targets, and soopens up a wide range of structural space. It also tests ourunderstanding of peptide and protein folding and assembly

directly. However, at present, it has clear disadvantagesover engineering natural systems, as we do not fully under-stand protein folding, and, therefore, the rational design ofprotein structure and function is in its infancy.3, 4

This chapter focuses on supramolecular assemblies thatare formed using a variety of de novo designed peptide-based tectons. A brief introduction to amino acids (thebuilding blocks of peptides and proteins) is given, followedby a discussion of the basic structures that polypeptidechains of amino acids can adopt. These structures formthe basis of the supramolecular assemblies that will bereviewed. The subsequent sections provide details of recentexamples of repetitive, effectively “infinite,” and discretesupramolecular peptide-based assemblies, and also a dis-cussion of their potential applications.

1.1 Amino acids

α-Amino acids are the fundamental building blocks ofpolypeptides. They consist of a central carbon atom (theα-carbon), which has both an amine and a carboxylicacid functionality attached. This central tetrahedral carbonadditionally bears a hydrogen atom and an “R” groupcommonly referred to as the side chain. In natural proteins,there are 20 different common side chains (Figure 1). It isthese that ultimately confer functionality onto polypeptides.The side chains can be classed according to their functionalgroups and are usually defined as being hydrophobic orpolar, though a small number might best be referred to as“special,” as they do not fit easily into these brackets, forexample, glycine, proline, and cysteine.

As the α-carbon has four different functional groupsattached, α-amino acids are chiral (with the exception of

Supramolecular Chemistry: From Molecules to Nanomaterials, Online 2012 John Wiley & Sons, Ltd.This article is 2012 John Wiley & Sons, Ltd.This article was published in the Supramolecular Chemistry: From Molecules to Nanomaterials in 2012 by John Wiley & Sons, Ltd.DOI: 10.1002/9780470661345.smc168

2 Supramolecular aspects of chemical biology

Glycine Alanine Serine Threonine Cysteine Proline

H OH OH

NH

HN

S

SH

CO2H

CH3

Valine Leucine Isoleucine Methionine Phenylalanine

O OO OOH

OH

OH

NH2

NH2

NH+

NH2+

NH3+ NH2

NH

HN

Aspartate Asparagine Glutamate Glutamine Histidine

Tryptophan Tyrosine Lysine Arginine

Figure 1 The 20 naturally occurring amino acid side chains.Where appropriate, the side chains are charged as they would beat pH 7 in water.

H R

CO2−+NH3

CO2−

+NH3

H

H

HN

R

RO

Peptide bond

+NH3

HN

HN

NH

O O

OR1 R3

R2 R4

(a) (b)

(c)

Figure 2 Amino acids and polypeptides. (a) Amino acid struc-ture; (b) Two amino acids joining and forming a peptide bond;(c) A linear polypeptide.

glycine, where the side chain is hydrogen). It is the L-isomerthat occurs naturally in ribosomally synthesized peptidesand proteins (Figure 2a), although examples of both natural(non-ribosomally synthesized) and designed structures thatcontain both D- and L-amino acids are known.

In polypeptide chains, amino acids are joined togetherthrough the condensation of carboxylic acid and aminegroups, leading to the formation of an amide or “peptide”bond (Figure 2b). A string of amino acids joined togetherin such a way is referred to as a polypeptide. Naturalpolypeptides are predominantly linear polymers and can

range from a few (Figure 2c) to thousands of amino acidsin length. All the information concerning the folding,assembly, and function of such polypeptides is usuallycontained within this primary sequence, which is the orderof amino acids read from the amino (N-) to the carboxy(C-) terminus. The trick in rational peptide and proteindesign is to capture this information in short, chemicallyaccessible (usually 30–100 amino acid) chains.

Contained within the folded tertiary structures of proteinsare areas of regular local folding, known as secondarystructures. The two most common secondary structures arethe α-helix and the β-strand, which form the basis of mostof the building blocks or tectons in peptide design.

1.2 α-Helices and β-strands

In the α-helix the main chain is wound up as a right-handed helix and the side chains point outwards fromthis central coil (Figure 3a). The structure is stabilized byhydrogen bonds parallel to the helix axis, formed betweenthe carbonyl oxygen of amino acid “i” and the amide protonof residue “i + 4.” α-Helices have 3.6 residues per turn,which means that side chains that are three or four residuesapart are brought together in space. This leads to one of thekey design principles in the supramolecular assembly of α-helices, namely the use of patterns of hydrophobic (H ) andpolar (P ) residues, spaced three or four residues apart, forexample, (HPPHPPP)n, along the chain to make so-called amphiphilic structures that can assemble in watervia association of their hydrophobic faces (Figure 3c).

(a)

(b)

(c) (d)

Figure 3 α-Helices, β-strands, and higher order structures.(a) An α-helix with the hydrogen-bonding pattern parallel tothe helix axis; (b) β-Strand with illustrative hydrogen bondsorthogonal to the direction of the chain; (c) α-Helices forminghigher order structures, the hydrophobic faces (shaded in darkred) coming together to sequester the hydrophobic residues fromthe solvent; (d) β-Sheet assembly is stabilized by the formationof intermolecular hydrogen bonds.


Design of biosupramolecular systems 3

β-Strands are the other common secondary structural ele-ment. Compared to the α-helix, the amino-acid residues ina β-strand are more extended, with the backbone hydrogen-bonding groups pointing orthogonal to the direction of thechain (Figure 3b) and side chains alternating between thetwo faces of the strand; effectively this gives the β-strand atwo-residue repeat. As a result, construction of amphipathicβ-strands requires alternating patterns of hydrophobic andpolar amino acids, for example, (HPHPHP)n.

Higher-order structures based on both α-helices andβ-strands are common, with amphipathic α-helices com-ing together to form either helical bundles (Figure 3c) orcoiled coils, in which hydrophobic faces are buried from thesolvent to form hydrophobic cores, while β-strands asso-ciate via intermolecular hydrogen bonds to form β-sheets(Figure 3d), which again can be amphipathic and associatein layers.

1.3 Helical bundles, coiled coils, and β-sheets

Helical bundles and coiled coils are common examples ofhigher-order assemblies of α-helices. The two structureslook similar in many respects: they consist of two ormore α-helices that associate in either a parallel or anantiparallel fashion. A key difference is that coiled coilsare based on a tight side-chain packing regime, knownas “knobs-into-holes” (KIH) packing.5 The basis of thispacking is that the hydrophobic residues that appear threeand four residues apart on one helix act as “knobs” andpack into diamond-shaped “holes” formed by four residuesfrom another helix (Figure 4a). This type of packing can beseen in crystal structures of coiled-coil peptides and proteinsand can be identified using the program SOCKET.6 As aresult of this regular, well-defined structural motif, clearsequence-to-structure relationships or rules that governcoiled-coil formation, association, and stability can beelucidated.7

Regularly repeated KIH packing leads to a repeatingsequence pattern known as the heptad repeat, in whichhydrophobic amino acids alternate every third and fourthresidue giving rise to a (HPPHPPP)n pattern. Thispattern is the defining characteristic of coiled-coil sequencesand is usually given the nomenclature abcdefg, with theH residues falling at positions a and d. As the numberof residues per turn in the α-helix is 3.6 and the averagespacing of the hydrophobic residues in the heptad repeat is3.5, an amphipathic structure is formed with all the H -type amino acids on one face of the helix (Figure 4b).Two or more of these helices will then come togethervia their hydrophobic faces to form a coiled coil, drivenby the hydrophobic effect in water. However, as 3.5 and3.6 do not match precisely, the helices wind around each

a

a

ed

a

a

e

e

b

bf f

c

c

g

g

d

d

(a)

(b)

(c) (d)

Figure 4 The α-helical coiled coil. (a) KIH packing—theresidues forming the “hole” are gray (with the positions in theheptad repeat labeled) and the “knob” is black; (b) Helical wheeldiagram for a parallel dimeric coiled coil, showing how thehydrophobic residues (at a and d) form the hydrophobic inter-face; (c) A top-down view of a dimeric coiled coil; (d) A sideview of the same, showing the left-handed supercoil.

other ensuring maximal contact between the hydrophobicresidues, resulting in a supercoil with a left-handed twist(Figure 4c and d). Note that variations on the heptadpattern are known and these lead to different helicalsupercoiling.8

The hydrophobic interactions, while being the drivingforce for coiled-coil formation, are not the only interactionsthat specify and stabilize these structures. Side chains foundat the e and g positions of neighboring helices are alsoclose in space and often have complementary charges,leading to electrostatic interactions across the hydrophobicinterface and further stabilization and specification of thestructures.

Extensive studies of coiled coils have elucidatedsequence-to-structure relationships that direct oligomeriza-tion, orientation, and partnering of the helices found inNature. These can be used with relative confidence innew designs. Generally, bulky and aromatic residues areexcluded from the a and d sites, and glycine and prolinetend not to be found or used in coiled coils as they breakα-helices. More specifically, a combination of isoleucineand asparagine at a positions and leucine at d definesdimeric coiled coils; designs with isoleucine at both a andd positions form trimers; and tetramers are directed byleucine at a positions and isoleucine at d.9 The e and g

positions show preferences for certain charged and polarresidues: for example, lysine, glutamic acid, and argininecan form electrostatic interactions across the helical inter-face, and in designed coiled-coil sequences these interac-tions can be exploited to direct the formation of coiledcoils with a specific orientation (parallel or antiparallel)and association (homomeric or heteromeric). The remain-ing b, c, and f positions are less restricted as to whichamino acids can be included and in design can often be



substituted to introduce function. For design purposes, itis usually sensible to include residues with a high α-helical propensity, such as alanine, leucine, lysine, andglutamic acid, but if, for example, the coiled coils needto be functionalized, residues such as cysteine can beincluded.

β-Strands are the other common secondary structuralelement. While there are fewer rules for β-strand forma-tion, certain sequence-to-structure relationships and fun-damental principles are clear.10, 11 For instance, it is wellestablished that, for natural proteins, amino acids with β-branched (e.g., valine, isoleucine, and threonine) or aro-matic (particularly, tyrosine, and phenylalanine) side chainsare among those most frequently observed. In addition,though this is not necessarily the case in natural proteins,12

for design purposes alternating patterns of hydrophobicand polar residues (HP )n can produce self-assemblingamphipathic strands.13, 14 Such sequences may then asso-ciate side-by-side through interstrand hydrogen bonds toform amphipathic β-sheets, in which the different side-chain types protrude above and below the plane (Figure 3band d). These sheets may associate further to form dis-crete β-sandwiches or indefinite β-tapes. The β-strands cancombine in parallel or antiparallel fashions, or form mixedsheets; although, possibly due to better hydrogen bondingor simply that consecutive strands tend to be joined byshort loops, antiparallel sheets are favored in proteins. Sidechain–side chain pairings across both antiparallel and par-allel β-sheets also show preferences,15–17 which have beensuccessfully employed in peptide design,13, 14 though fullimplementation of these in protein-structure design and pre-diction remains a challenge. Finally, it should be noted thatmost β-sheets are not flat, as the strands twist relative toone another, which has implications for understanding fun-damental interactions in β-sheets and exploiting them indesign.

To date, most β-structured designs have focused onantiparallel β-assemblies of either two-stranded β-hairpins,where contrary to the findings in natural proteins, trypto-phan residues are found to be stabilizing and so are fre-quently used,18, 19 or indefinite supramolecular assembliesof β-tapes and amyloid-like systems.13, 14 Designed β-sheetassemblies have recently been the focus of much research,as β-sheets and the structures formed from their associationreadily lead to amyloid-like fibrous and gelling systems,which in Nature are implicated in many disease states suchas Alzheimer’s and Parkinson’s.20 For these reasons, thereis great interest in exploring their formation and behaviorand also exploiting their potential as materials for use inbionanotechnology.

(a) (b)

Figure 5 Collagen building blocks. (a) A polyproline II typehelix; (b) A collagen triple helix assembled from three polyprolineII strands. (Reproduced from Ref. 1. American ChemicalSociety, 2008.)

1.4 Collagen

Collagen is a naturally occurring protein found in theconnective tissues of mammals, though examples are nowclear in bacteria and viruses.21 It has a triple-helicalstructure, but in contrast to α-helical coiled-coil structuresthe individual collagen polypeptides form polyproline II(PPII) type helices, which are left-handed and associate viainterchain hydrogen bonds to form right-handed supercoils(Figure 5). The sequences of fibrous collagens are definedby an (X–Y-Gly)n repeat. The glycine at every third residueis essential to allow tight packing of the three chains in thetriple helix. Proline and hydroxyproline are prevalent at theX and Y position, respectively, and most of the recent denovo designed collagen peptides use one (or both) of theseresidues at these positions. The superhelix can be eitherhomotypic, meaning all three chains are identical (knownas AAA homotrimers), or heterotypic. There are two typesof heterotypic collagen, one where all three chains aredifferent (an ABC heterotrimer), or a structure where twohelices are the same and the remaining helix has a differentprimary sequence (an AAB heterotrimer). Examples of allthree types of designer collagen will be discussed later inthis chapter.

1.5 Peptide amphiphiles

Peptide amphiphiles (PAs) are another example of alterna-tive tectons that can be used in the design of supramolecularsystems. These non-natural, hybrid structures comprise apeptide, which is usually polar, and a nonpolar aliphaticregion. They adopt structures similar to those formed



by lipids, which have polar head groups and nonpo-lar fatty acid tails: in both cases, the hydrophobic tailspack together sequestered away from solvent, while thepolar groups are solvent-exposed, leading to micelle-likestructures. For the PAs, different functionalities can beconjugated via the peptide moieties. For example, longalkyl carbon chains can be attached, as can syntheticpolymers such as polyethyleneglycol (PEG) or N-(2-hydroxypropyl)methacrylamide (HMPA). Similarly, amphi-pathic peptides with suitable arrangements of hydrophobicand polar side chains have been designed to assemblein a similar fashion.22, 23 The design of such amphiphileshas been the focus of several recent reviews; these coverin detail the different types of amphiphiles that can besynthesized and the resulting systems that can be pro-grammed.24, 25

2 REPETITIVE EXTENDEDSUPRAMOLECULAR SYSTEMS

With the structural basis of the commonly used buildingblocks for the de novo design of biosupramolecular sys-tems having been covered, specific examples of assembliesthat can be formed are discussed. This section reviews typesof repetitive and extended (effectively infinite) structures.These tend to contain many thousands of monomeric tec-tons self-assembled in one direction, hence the infinite (orat least indefinite) and repetitive description. Examples ofthese kinds of systems are predominantly fibrous assem-blies and fibers capable of gelling. Many groups have madefibrous structures from various straightforward peptidic tec-tons over the past 15 years, resulting in a variety of systemsof increasing sophistication. Owing to the large body of lit-erature, including many review articles,26–28 this section isbrief and covers archetypal examples of extended fibrousstructures.

2.1 Fibrous systems with β-structuredbuilding blocks

The earliest examples of designed fibrous systems are basedon β-structured tectons. The majority of these systems tendto adopt amyloid-like structures: that is, extended arraysof hydrogen-bonded β-strand forming sheets that pack intofibrils and then further assemble into fibers. In many cases,these fibers appear to be predisposed to form gels.

Some of the earliest work with β-structured tectons is byZhang and colleagues. They examine a variety of simplerepeating sequences and their fiber-forming abilities. Thefirst peptide, EAK16, is based on a natural sequence found

in the yeast protein zuotin.29 It forms stable membranesat high salt concentrations. Scanning electron microscopy(SEM) reveals interwoven filaments within the membranes,akin to amyloid fibrils. The group proposes that the fibersare formed by β-structured peptides interacting via bothhydrogen-bonding and ionic interactions to give arraysof staggered peptides that twist around each other toform filaments which then associate and organize intomembranes. More recent work demonstrates that EAK16and a related de novo designed peptide RAD16 support thegrowth of mammalian cells.30

A second example of gels based on designed amyloid-like structures is presented by Aggeli et al.31 The group hadpreviously studied natural β-structured peptides that gel inaqueous solution, and from these studies concluded that,along with the formation of intermolecular hydrogen bonds,attractive forces between side chains of adjacent strands,lateral recognition of the strands, and also the ability ofsolvent to adhere to the strands (to aid solubility) are allimportant criteria for gel formation. On the basis of thesecriteria, a design for an 11-residue β-strand-forming peptidewith the sequence QQRFQWQFEQQ was developed. Inaqueous solution, this assembles to form a gel consistingof “β-sheet polymer tapes.” The responses of the gel tochanges in pH and shear forces have been studied. Thegroup has expanded on this work and presented alternativedesigns that form similar “tape-like” structures as well asdesigns shown to form ribbons (essentially “double tapes”that form at higher peptide concentrations) and fibers.32

The group has also developed a useful, and potentiallymore broadly applicable, definitive statistical model forpeptide assembly, which demonstrates the self-limitingnature of such higher order structures due to a competitionbetween the free energy of attraction of the peptides forthe assemblies, and an energy penalty associated with theelastic deformation of the peptides as they are incorporatedinto fibrils.

The Schneider and Pochan groups have outlined severaldifferent controllable designs for fibrous hydrogels basedon β-hairpin tectons. One example exhibits thermallycontrolled and pH-dependent behavior33: MAX1, a peptideincorporating nine lysine residues is unfolded at pH 9 atroom temperature, but adopts the β-hairpin conformation,self-assembles, and hydrogelates as the temperature isincreased. The group has demonstrated that replacing lysineresidues with glutamic acid (and thus reducing the netpositive charge of the peptide) results in hydrogelationat lower pH values for a given temperature. This isbecause less thermal energy is required to overcome therepulsive electrostatic interactions that result from thelysine residues being on one face of the strand. Analternative heteromeric design from the group also exhibits



a temperature-induced transition from random coil to β-structure.34 At low temperatures, one of the peptides of thesystem, (strand-swapping peptide) SSP-1, is unstructured,but as the temperature is raised it adopts a β-hairpinconformation in which a β-strand extends from a β-turnregion. This β-strand is designed to domain-swap with asecond, similar β-strand domain (SSP-2), leading to theproduction of a strand-swapped dimer. This dimer then self-assembles into fibrils that gel (Figure 6). The group hasother designs of β-structured hydrogels, many of whichhave been explored with regard to their controllability andmaterials properties.35, 36 They have also demonstrated thepotential for applications such as cell growth and delivery,as well as showing that one design has inherent antibacterialactivity.37, 38

It is not just β-structured amyloid-like designs that arecapable of forming fibrous systems. A notable example,and one of the first to demonstrate fiber formation,

is Ghadiri’s extended nanotubes made from peptidescontaining alternating D- and L-amino acids, motivated bythe antibiotic Gramicidin S.39 The design consists of acyclic octapeptide with the sequence (D-AED-AN)2. Therationale is that the rings formed should be near flat andstack in an antiparallel fashion via β-sheet-like intermolec-ular hydrogen bonds. In this conformation, the side chainsall lie on the outside of the rings, due to the alternating D,L-configuration, and so give hollow tubes. Indeed, the pep-tides spontaneously assemble at acidic pH, to give fibrousnanotubes as viewed by transmission electron microscopy(TEM) (Figure 7), and the underlying β-sheet structureis confirmed by Fourier-transform infrared (FT-IR) spec-troscopy. The designs have since been modified to producefurther cylindrical assemblies,40 as well as antibacterialagents.41–43

Related to the above work on β-structured systems,Gazit, Ulijn, and others have produced β-tubes, fibers, and

Unfoldedpeptide

Foldedβ-hairpin

Heat

Strand swapped dimer

H-bon

ding

axis

2.5 nm

6 nm

Fibrilnanostructure

100 nm

(b)

(a)

Figure 6 Assembly of SSP-1 and SSP-2 peptides. (a) The proposed design and assembly mechanism detailing how fibril formation isachieved; (b) Transmission electron microscopy (TEM) image of the resulting fibrils. (Reproduced from Ref. 34. American ChemicalSociety, 2008.)



(a)

(b) (c)

Figure 7 TEM images of Ghadiri’s cyclic peptide nanotubes.(a) TEM image of many nanotubes (scale bar = 1 µm); (b) Asingle nanotube; (c) An enlarged image of the boxed region of(b) showing striations on the surface of the nanotube. (Reproducedfrom Ref. 39. Nature Publishing Group, 1993.)

tapes from shorter Fmoc-protected and related peptides.Such work has been covered comprehensively in recentreviews and is not discussed further here.44–46

2.2 Extended systems based on coiled coils

There are two broad kinds of extended systems based oncoiled-coil frameworks: those designed using homomericconstructs, and those that use heteromeric designs; thelatter carry advantages for control over assembly of thesystems.

(a)

(b)

Figure 8 TEM images of the α3 peptide. (a) Fibrils form inphosphate buffer containing 0.1 M KCl at pH 6; (b) With 1 M KCllarger fibers are observed (scale bars = 100 nm). (Reproducedwith permission from Ref. 47. The Japan Academy, 1997.)

2.2.1 Homomeric fibrous systems

The first example of a fiber-forming system constructedfrom coiled-coil building blocks is from Kojima et al.47

The design α3 consists of a three-heptad peptide with therepeating sequence LETLAKA at the abcdefg sites of thecoiled-coil heptad repeat, respectively. Previous analysisof the peptide by sedimentation equilibrium analyticalultracentrifugation (AUC) showed that it forms tetramerichelical bundles. However, changing the buffer renderslarger assemblies, which, when probed by TEM, are foundto be fibrils with widths of 5–10 nm and lengths ofseveral micrometers; the precise morphology alters withsalt (Figure 8). The mechanism of fiber formation is notdiscussed in detail, but it is proposed that the hydrophobiceffect provides the main driving force for fiber assembly.A more-recent paper describes the effect of reversing thesequence.48 This later design is more helical, has increasedstability, and forms longer fibers, though with the samewidth as those produced by α3. In this case, increased saltproduces irregular aggregates.

In 2001, Potekhin et al. presented an α-helical fiber-forming peptide (α-FFP) with the sequence QLARELat cdefga, followed by (QQLAREL)4 at bcdefga.49 Thepeptide is designed to assemble into pentamers, with therationale that fiber formation results from slippage of thisrepetitive sequence to give staggered ends that promotelongitudinal assembly (Figure 9a).



(a) (b)

0.1 µm

Figure 9 Fibers formed from the pentameric coiled-coil αFFP.(a) Schematic representation of how a pentameric coiled coilcan adopt a staggered assembly and form extended fibers.(Reproduced from Ref. 49. Elsevier, 2001.) (b) TEM imageof a modified fiber design. (Reproduced from Ref. 50. OxfordUniversity Press, 2003.)

Indeed, αFFP forms fibrils below pH 6 and hydrogelsat high peptide concentrations. The group has since mod-ified the initial design, replacing the charged glutamicacid residues with glutamine or serine so that the fib-rils form over a wider pH range (Figure 9b). The fib-rils have been functionalized with an N-terminal RGDsequence to promote integrin-dependent cell adhesion witha view to using the fibrous system as a scaffold for cellgrowth.50, 51

Using a similar, but more sophisticated and rationalapproach, where staggered assembly is promoted by spe-cific interactions between designed peptide sequences,Conticello et al. describe a construct intended to formfibrils from dimeric coiled-coil tectons.52 To achieve this,they introduce hydrogen-bonding polar residues into thehydrophobic core, and place charged residues to promotethe desired staggered assembly and prevent the formationof the more usual, naturally observed, blunt-ended dimersin which the ends of the helices are fully in register, orflush, with respect to each other. Circular dichroism (CD)spectroscopy confirms that the peptides do indeed adopt anα-helical structure, and TEM images reveal long fibers. Thegroup has built on this initial design and made other fiber-forming systems, which, for example, are pH responsive.53

A subtly different mode of fiber formation is proposedby Fairman et al.54 The designed peptide, CpA, is basedon the GCN4 leucine-zipper sequence, but has two alanineresidues inserted between two identical two-heptad coiled-coil sequences. This insert results in the hydrophobic seamswitching from one “face” of the helix to the other halfwaythrough the sequence (Figure 10a). This positional shiftdrives a contiguous, longitudinal (as opposed to blunt-ended) coiled-coil assembly. The free ends generated allowadditional peptides to add to the dimeric construct leadingto fiber formation (Figure 10b). As each coiled-coil half isshort, at only two heptads long, high salt concentration isneeded to induce the helical structure of the peptides andtrigger fiber formation. Ammonium sulfate is found to bethe most effective salt for inducing helical structure and forforming longer and more stable fibers (Figure 10c).

Designed helix-turn-helix peptides that form fibrils havealso been described.55 Two 18-residue α-helical segmentsare joined by a variety of natural turn sequences fromhuman apolipoprotein A–I to give a series of four pep-tides. All of these are α-helical and three of the designs(those incorporating a proline residue in the turn region)demonstrate fibril-forming ability. One design has been

121

21 1

21

2 2

(a)

(b) (c)

1 µm

Figure 10 Representations of the CpA peptide. (a) Molecular model of the CpA peptide showing the position of the hydrophobicface; (b) Schematic representation of how the monomeric tectons form staggered fibers; (c) Atomic force microscopy (AFM) image ofthe resulting fibers in ammonium sulfate. (Reproduced with permission from Ref. 54. National Academy of Sciences, 2005.)



16.0 Å 9.5 ÅC

CN

N

(Tippedhorizontallyand vertically)

Helix pitch 5.0 ÅRise per residue 1.55 Å

Polymerize

and twist

32.5 ÅProtofilament

diameter

Figure 11 Schematic of the two peptides employed by Lazaret al. in which α-helices stack perpendicular to the long fiberaxis. (Reproduced from Ref. 55. American Chemical Society,2005.)

investigated in detail, to reveal that the α-helical axesare perpendicular to the long axis of the fibril; this isthe first example of fibrils in which the α-helical tec-tons lie relative to the fiber axis in this way (Figure 11).Clearly, there is added complexity in moving to helix-turn-helix systems—for instance, turns are notoriously difficultto engineer in peptides and proteins—which makes suchmulti-helix-based design (or indeed any multisecondarystructure system) more challenging. Such challenges shouldbe addressed, however, if we are to succeed in engineeringmore ambitious, complex, and ultimately useful systems.56

Most recently, Hartgerink and coworkers have producedfibers from short, blunt-ended coiled coils.57 They useseveral subtly different three-heptad sequences. Assemblyis concentration-dependent: at high concentrations, bundlesof fibers are formed, leading to gelation (Figure 12a). Inaddition, fibril thickness can be controlled by varying aminoacids at the b, c, and f positions of the heptad repeat;when positively charged residues are incorporated at these

positions, fibrils with narrow diameters are observed, butin their absence fibers with much larger diameters form.A general mechanism is proposed for fibril formation,in which blunt-ended coiled-coil dimers spontaneouslyform; these then associate to offset pairs of coiled-coildimers above a critical minimum concentration; and thisassociation serves to nucleate fiber formation (Figure 12b).

2.2.2 Heteromeric fibrous systems

The next step in terms of complexity within coiled coil-based systems is to design two-component (or higher)fibrous systems. This has the added benefit of introducingcontrol into self-assembly: assembly ensues only when allthe necessary components are mixed.

Arguably, the best studied example of a two-componentheteromeric coiled-coil fiber-forming system comes fromour own lab. The self-assembling fibers (SAFs) consist oftwo different complementary sequences that are designedto assemble into an offset parallel coiled-coil dimer.58

This leads to a sticky-ended tecton for fiber assembly.This design is realized with two four-heptad peptides withsequences as follows:

KIPPKLP KIPPLKP EIPPLEP ENPPLEP (SAF-p1)KIPPLKP KNPPLKP EIPPLEP LIPPLEP (SAF-p2)

In these sequences, the hydrophobic residues at a andd promote parallel dimers, while the charge patterning ate and g is set for staggered assembly. However, the keydesign feature is the inclusion of offset asparagine residuesat a positions, which specifies staggered assembly as thetwo asparagines residues most likely pair to form a hydro-gen bond across the hydrophobic interface (Figure 13). Theremaining P sites are filled primarily with residues with ahigh α-helical propensity in order to maximize the helicalcontent of the peptides.

(4) Lateral aggregation into matured fibers

(1) Coil to dimeric helix transition(2) Concentration-dependent nucleation(3) Supramolecular polymerization into fibrils

a. b.

e.

c.

d.

(1) (2)

(3)

(4)

3.7 nm

≠

50 nm

(a) (b)

Figure 12 Fibrous assembly using short coiled coils from Hartgerink et al. (a) Cryo-TEM image providing evidence for fiberformation; (b) The mechanism of fiber formation. (Reproduced from Ref. 57. American Chemical Society, 2008.)



N

N

p1

p2

Figure 13 The SAF concept from the Woolfson group. The dif-ferent colored blocks represent positively and negatively chargedheptads and the “N” represents the positions of the offsetasparagine residues, which help specify the offset sticky-endedheterodimer.

The α-helical structure of the peptides is confirmed byCD spectroscopy and X-ray fiber diffraction, and TEM con-firms the presence of long, thickened fibers.58 One of thedrawbacks of this initial design is that the resulting fibersare relatively unstable and only form below room tempera-ture. This led to a second-generation design,59 where pos-itively charged arginine residues were incorporated at twoconsecutive c sites of SAF-p2—that is, outside the coiled-coil interface—to match and interact electrostatically withsimilarly placed aspartic acid residues in the original SAF-p1. The characterization of this design revealed more stableas well as better ordered and thickened fibers (Figure 14).The thickening behavior is an interesting emergent propertythat has been investigated and exploited by our group.60 Instructural terms, we have confirmed that the peptides assem-ble end-to-end in line with the long axis of the assembledfibers and that fiber thickening occurs with hexagonal pack-ing of coiled coils.61

More recently, we have determined the kinetic path-way for peptide folding,62 and the designs have also beenextensively modified and redesigned to generate examplesof kinked, waved,63 and branched64 fibers. SAF variantsthat gel and support cell growth have also been devel-oped. Gelation is achieved by replacing residues at the b, c,

and f positions with either alanine (to promote interfiberhydrophobic interactions) or glutamine (promoting inter-fiber hydrogen bonding). These changes produce thinner,more-flexible fibers that interact as designed to form per-colated networks and so physical hydrogels. Extensivebiophysical, rheological, and microscopic characterizationdemonstrate robust hydrogels, particularly for the Ala-basedpeptides, that support mammalian cell growth and diffenti-ation.65

2.3 Fibrous systems based on collagens

Fiber-forming systems can also be designed using tectonsbased on the Pro-Hyp-Gly sequence repeat of fibrouscollagens. These systems have only recently been widelyinvestigated and, therefore, are less well represented in theliterature than the above systems based on α-helical andβ-structured components.

One of the first examples is from Kotch and Raines.66

They use two short collagen-based fragments rich in glycineand proline that self-assemble into an AAB-type triplehelix. The three strands are joined by covalent disulfidebonds, which force the helices to assemble in a stag-gered fashion, promoting fibril formation. Work from Hart-gerink’s lab a year later shows that designed nonfibroushomotrimers, AAB heterotrimers, and, for the first time,an ABC heterotrimer can be assembled and stabilizedthrough noncovalent interactions alone.67 They use CDspectroscopy to demonstrate the correct folding of thesestructures and also to show high thermal stability. Theyconclude that the designed electrostatic interactions arethe primary driving force for triple-helix formation, andthat these interactions stabilize the structures efficiently.Similarly, Conticello et al. show that collagen-like fibrilswith intriguing ultrastructure—that is, the detailed struc-ture of the fibrils visible by electron microscopy, and abovethe underlying secondary and quaternary protein struc-ture—can be assembled from straightforward peptides.68

(a) (b)

2 µm 50 nm

Figure 14 TEM images of the second-generation SAFs. (a) Low-magnification image after maturation for 12 h; (b) High-magnificationimage showing regularly patterned striations on their surfaces. (Reproduced from Ref. 59. Wiley-VCH, 2006.)



(a)

0.5 µm

(b)

1 µm

Figure 15 TEM images of the designed collagen-like fibrilsfrom Krishna and Kiick. (a) TEM image of many collagenfibrils; (b) Higher magnification image of an individual fibril.(Reproduced from Ref. 69. American Chemical Society, 2009.)

Their design, CPII, self-assembles into a homomeric triplehelix and these helices are able to form collagen-likemicrofibers, which remarkably display D-periodic-like pat-terns, an intrinsic feature of native fibrous collagens and onewhich many designed collagen-like fiber-forming systemslack. It is likely that this patterning results from the designcontaining hydrophobic and hydrophilic portions, leading tostaggered assembly and the formation of regularly spacedelectrostatic interactions. A final recent example of impor-tance is from Krishna and Kiick.69 They show that syntheticcollagen triple helices without hydroxyproline can be sta-ble and can self-assemble into fibers as observed by TEM(Figure 15). In contrast to the more common GOP repeat,this design has GPP repeats at the N-terminus and near theC-terminus; a five amino acid sequence incorporating twocysteine residues at the C-terminus; and charged residues inthe central portion to facilitate the stabilizing electrostaticinteractions. Though their design contains cysteine residues,they show that self-assembly of the peptides is possibleunder both oxidizing and reducing conditions, so the for-mation of disulfide bonds is not essential to the stability ofthe self-assembled structures. This design is important, as itincorporates only natural unmodified amino acids, making

it amenable to recombinant expression, thus expanding itsfuture usefulness.

2.4 Peptide amphiphile fibrous systems

PA systems have perhaps the potential to be the mostdiverse of all the repeating fibrous assemblies due to thewide range of synthetic moieties that can be incorporatedinto the tectons. Owing to this variety, many examples ofamphiphilic systems have been explored but a select few arehighlighted here in order to present a taste of the potentialsystems that can be designed and synthesized, as well asthe potential applications of such systems.

Work presented by Hartgerink and Stupp initiated thisfield. One of their notable designs consists of a PA contain-ing a long alkyl tail and a peptidic head group functional-ized with a variety of different moieties.70 The PA formsfibers (Figure 16a) but also gels at high concentrations inacidic solutions, and cryo-TEM reveals long ordered fiberswhich are assumed to be the PAs assembling into cylin-drical micelles (Figure 16b). The assemblies are capableof nucleating hydroxyapaptite on their surfaces (due to thepresence of a phosphoserine residue incorporated into thepeptidic head group). The hydroxyapaptite mineralizes in amanner such that the hydroxyapaptite crystals align with thefiber axis, which is similar to the hierarchical organizationfound in bone (Figure 16c).

Hartgerink’s own lab has built on this approach toself-assembly, with several new PA systems, specificallyshowing that the inhibition of cancer cell proliferation canbe achieved by utilizing a designed PA system.71

This field is vast and over recent years has spawnedever more intricate and sophisticated designs. Even thoughthe area is well represented, it is envisaged that moredesigns will be presented in the coming years, especiallythose which lead to functional systems; for example,hydrogels that are capable of supporting cell growth anddifferentiation are currently the subject of intense research.

50 nm 200 nm 20 nm

(a) (b) (c)

Figure 16 Images of the fibers formed from PAs designed by Hartgerink and Stupp. (a) TEM image of the self-assembled fibersarranged into ribbon-like arrays; (b) Cryo-TEM image of the fibers; (c) TEM image of a fiber covered in mature hydroxyapaptitecrystals (shown by the red arrows). (Reproduced with permission from Ref. 70. American Association for the Advancement ofScience, 2001.)



3 DISCRETE SUPRAMOLECULARSYSTEMS

While effectively “indefinite” or “infinite” fibrous structureshave proven relatively straightforward to design, and arenow advancing toward functional systems; the successfuldesign of finite or discrete self-assembled peptidic systemsis proving more challenging. These structures are muchsmaller than the repetitive fibrous systems and containa finite and ideally defined number of tectons; hencethe term discrete. While there are still relatively fewexamples of such assemblies, several systems have beendesigned successfully and are presented in this section. Thedesign principles behind each system are covered, alongwith potential applications. As most of the examples arerelatively unrelated in terms of their underlying designprinciples and function, they are given chronologically.

3.1 Belt-and-braces peptides

An early example of a discrete self-assembled system isour own “belt-and-braces.”72 The design comprises threecoiled-coil peptides, one of which is twice the length of theothers, and acts as a template for their assembly; the “belt”peptide is six heptads long, and the two “brace” peptideseach have three heptads. The design is based on paralleldimeric assemblies, with the majority of the a sites in allthree peptides occupied by isoleucine residues and all ofthe d sites leucine. Homo-assemblies are discouraged bygiving the e and g sites on each peptide the same charge;the belt peptide has all glutamic acid residues at thesepositions, while the braces have lysines. To distinguish thetwo braces and direct them to different ends of the belt,one a site of the C-terminal brace is filled with asparagineto complement an asparagine at the a position in the belt.Together, these amino acid placements specify a ternarycomplex as demonstrated by biophysical and microscopicanalyses.

The brace peptides are terminated by cysteine at their“free ends” in the complex to facilitate coupling to goldnanoparticles, the aim being to direct nanoparticle assem-bly. The resulting brace peptide–nanogold constructs inter-act only when the belt peptide is added; precipitationof aggregated particles is observed, accompanied by acolor change, indicative of higher-order nanogold assem-bly. Assemblies are confirmed by TEM, which shows theassembly of both 2D and 3D networks of colloidal nanopar-ticles (Figure 17).

3.2 Formation of supramolecular assembliesusing leucine-zipper displaying dendrimers

Ghosh and coworkers demonstrate that discrete supramolec-ular assemblies can be formed from leucine-zipper peptidestethered to a dendrimer core.73 These give fibrous struc-tures when two complementary discrete dendrimer–peptideassemblies are combined; this is one of the first exampleswhere both discrete and repeated assemblies can be formedusing the same system (Figure 18a).

The group uses a poly(amido amine) (PAMAM) den-drimer core, functionalized with surface maleimide groups,chosen to react with cysteine-containing peptides. Thecoiled-coil peptides used are an acid–base pair, whichassemble to form coiled-coil tetramers. Both have cys-teine at the C-terminus to allow covalent attachment. Twomolecules are described: one dendrimer with the acidic (EZ)peptide attached, and the other with the basic (KZ) peptide.The dendrimer–EZ molecule (D–EZ) combines with fourequivalents of the free KZ peptide at pH 8.4 as judged byCD spectroscopy and AUC. Adding four equivalents of thefree EZ peptide to the D–KZ construct at low pH givessimilar results. Mixing the dendrimer–peptide assemblies(D–EZ and D–KZ) results in protofibrils and eventuallyfibers, as confirmed by TEM (Figure 18b and c).

In the following year, the group built on this work toshow that the dendrimer core forms similar assemblies

50 nm

(a) (b)

50 nm

Figure 17 TEM images of gold nanoparticle networks formed using the belt-and-braces peptides. (a) A 2D network; (b) a 3D network.(Reproduced from Ref. 72. American Chemical Society, 2003.)



D–EZ4

D–EZ4

D–KZ4

D–KZ4

D–EZ4/4 KZcomplex

D–KZ4/4 EZcomplex

KZ

EZ

pH 8.4

pH 5.6

pH 7.0

+

+

+

Self-assembled fiber(simplest model)

n

n

200 nm1000 nm

(a)

(b) (c)

Figure 18 Discrete and fibrous assemblies utilizing thedendrimer-peptide constructs of Ghosh et al. (a) Both discreteand extended systems can be formed from coiled-coil/dendrimerconstructs; TEM images of (b) protofibrils and (c) fibrils formedfrom the self-assembling constructs. (Reproduced from Ref. 73. American Chemical Society, 2004.)

with the natural coiled-coil peptides Fos and Jun at neutralpH—making the design more biocompatible and amenableto further derivitization.74

As Fos forms weaker homodimers,75 it is attached tothe dendrimer hub. Even so, the dendrimer–Fos construct(D0 –Fos4) is 85% helical by CD spectroscopy. AUCconfirmed that this folding is a result of intra- (as opposedto intermolecular) homodimer formation as no higherorder assemblies are observed; that is, despite its lowdimer content, Fos peptides interact because of their higheffective concentration when attached to the dendrimer. Aneightfold excess of Jun is required to form four Fos-Junheterodimers when starting from the D0 –Fos4 construct.With 4 equivalents of Jun, only 2 appear to incorporateinto the D0 –Fos4 assembly. It is suggested that the need forthe excess could be attributed to either the energy penaltyassociated with breaking intramolecular Fos homo-pairings

or for steric reasons. AUC analysis confirms that with aneightfold excess of the Jun peptide the D0 –Fos4/4PJuncomplex is formed, albeit along with other homodimericand monomeric Jun species. The group envisaged that thisframework could be used as a basis for receptor targetingor DNA binding.

3.3 Self-assembly of regular polyhedralnanoparticles

The previous examples involve one coiled coil directingthe assembly of a partner to form heterodimeric systems.The next case is slightly different in that two differentcoiled-coil sequences are contained in one building block. Acoiled-coil construct comprising a sequence shown to formhomopentamers is linked (via a two-glycine spacer) to ahomotrimeric sequence.76 It was anticipated that 15 of thesemonomeric tectons would self-assemble into a constructcontaining 3 pentamers and 5 trimers and that these unitswould self-assemble further to form polyhedral nanoparti-cles containing 60 monomeric units or 4 of the smaller self-assembled constructs (Figure 19a). Nanoparticle assemblyis probed using both TEM and AUC. The former showsthe appearance of regular monodisperse nanoparticles whenspecific folding regimes are used (Figure 19b), while AUCshows that assembly of discrete nanoparticles is highlyconcentration-dependent, with the desired assembly of 60monomeric units being formed though under very specificconditions.

It is proposed that such a system could be used as aplatform for antigen display, as it is highly symmetrical andrepetitive and resembles a viral capsid. Such assembliesare known as synthetic virus-like particles (SVLPs) (seebelow). Indeed, other papers have followed from thegroup demonstrating varying levels of success in displayingantigenic actin determinants, as well as being utilized in thedevelopment of prototypic vaccines for malaria and severeacute respiratory syndrome (SARS).77–79

3.4 Synthetic virus-like particles

Continuing the theme of self-assembled nanoparticles,Robinson and coworkers have implemented a slightly dif-ferent approach to form SVLPs. Their building block con-sists of a lipid conjugated to a coiled-coil peptide thatis designed to form trimers. The rationale is that thecoiled coils will self-assemble into parallel trimeric bun-dles, and the lipids will then drive further self-assemblyof these trimeric bundles into nanosized SVLPs withthe lipids buried and the C-termini of the coiled coilsexposed (Figure 20a and b). The presence of regularly



(a)

(b)

50 nm

Figure 19 Synthetic peptide nanoparticles. (a) Computer modelof the self-assembled nanoparticle incorporating 60 copies ofthe monomeric construct; (b) TEM image of the nanoparticles.(Reproduced from Ref. 76. Elsevier, 2006.)

sized nanoparticles was confirmed by TEM (Figure 20c).Antigens have been added to the C-terminal end of thepeptides to demonstrate the utility of the system to presentnovel immunogens. Further experiments confirm that theseSVLPs do indeed generate antigen-specific antibodies inanimal models.80

3.5 A self-assembling peptide polynanoreactor

Ryadnov describes a coiled-coil design that self-assemblesto form a polynanoreactor.81 The design consists of twocomplementary peptide supradendrimers (noncovalent pep-tidic dendrimers) designed to self-assemble and forma polynanoreactor with various sized cavities. Peptidesupradendrimer 1 (SD-1) is designed around a homodimeric

AntigenAntigens

Peptidechain

Peptidechains

Lipidtails

Synthetic virus-like particle(SVLP) + antigens

Lipidtails

Lipopeptidebuilding block

Oligomerizationof coiled coils

Self-assembly

Lipid core

(a)

(b) (c)

Figure 20 Nanosized SVLPs from Boato et al. (a) Schematicof SVLP assembly; (b) Computer-generated image of the SVLP;(c) TEM image of the SVLPs (scale bar = 100 nm). (Reproducedfrom Ref. 80. Wiley-VCH, 2007.)

sequence. However, polar charged residues are placedto encourage intermolecular electrostatic interactions andpromote associations between dimers, leading to self-assembled, ordered noncovalent networks (Figure 21a). Thecavities within the networks are too small to be of anypractical use. To address this, a second peptide supraden-drimer sequence (supradendrimer 2, SD-2) is designed toform heterodimers with SD-1. In addition, cysteine residueson the outside promote trimers of heterodimers, resulting ina starburst arrangement and self-assembly of a network withlarger cavities (Figure 21b).

A metal redox reaction is described to determinewhether the polynanoreactors are functional; the coiled-coilsequences incorporate cysteine residues so that encapsu-lated colloidal silver can be generated, and TEM confirmsthe presence of regularly dispersed nanoparticles.



1 1

1

I

I

II

II′

c

c

c

c

e

e

e

e

b

b

b

b

f

f

fc

f

f

g

g

g

g

d

d

d

a

aad

f

ae b

c

d g

2b

f

c

d

e a

d

a

a

g e

b

1 1

bf

dgc a

e

(a)

(b)

2

2

g

Figure 21 Designed coiled-coil polynanoreactors. (a) SD-1self-assembles to form a nanoreactor with small cavities; (b) SD-1and SD-2 together form a nanoreactor with both small and largercavities. (Reproduced from Ref. 81. Wiley-VCH, 2007.)

3.6 A reduced SNARE model for membranefusion

In an attempt to mimic membrane fusion events, the Kroslab has constructed a simplified peptide-based model ofthe SNARE (soluble NSF (N-ethylmaleimide sensitive fac-tor) attachment protein receptor) proteins intrinsic to this

process (Figure 22a and b).82 They describe two triblockcopolymers consisting of a lipid domain, a central PEGspacer, and either an acidic or a basic designed coiledcoil (Figure 22c). The concept is that the lipid domainshould incorporate into vesicles, leaving the coiled coilson the vesicle surface. Two vesicle populations, function-alized with either acidic or basic coiled coils, can thenbe mixed, allowing the coiled coils to heterodimerize andcause vesicle fusion as the vesicles are bought into closeproximity (Figure 22d). They test the design by performinga FRET (Forster resonance energy transfer) assay. A fluo-rescent complex is encapsulated in one vesicle populationand a nonfluorescent ligand in the other. If fusion occurs,the contents of the two vesicle populations will mix and anincrease in fluorescence will occur, which is exactly whatis observed in practice.

3.7 Shape and release control of peptidedecorated vesicles

A second example from the Kros lab is particularly interest-ing, as it uses a nonpeptidic scaffold to template the forma-tion of β-structured peptides.83 A β-cyclodextrin vesicle(CDV) is employed as a scaffold to template the assem-bly of an octapeptide, (LE)4, that is N-terminally modifiedwith adamantane. The adamantane moiety forms an inclu-sion complex on the surface of the vesicles. The group usesCD spectroscopy to show that the peptide is unstructured

(a) (b)

(c) (d)

DOPE

DOPE

E

K

PEG12

PEG12

1

23

Figure 22 Membrane fusion using model peptides designed by Marsden et al. (a) A schematic of how the model peptide constructsare designed to interact and form a heterodimer; (b) A representation of peptide interactions in the natural system; (c) A computerimage of the two triblock copolymer constructs; (d) A schematic of the fusion process with the model peptides. (Reproduced fromRef. 82. Wiley-VCH, 2009.)



(a) (b) (c)

Figure 23 Cryo-TEM images showing the stages in the peptide–vesicle interaction process from Kros et al. (a) The addition ofpeptide to CDVs at physiological pH does not affect vesicle morphology; (b) Fibers are seen 1 h after acidifying the solution; (c) Thedisappearance of fibers is observed when the pH is returned to physiological pH. (Scale bar in all images = 100 nm.) (Reproduced fromRef. 83. American Chemical Society, 2009.)

and remains so when added to CDVs at physiological pH(Figure 23a). However, upon acidification of the solutionthe peptides adopt a β-structure. This structural change isconferred to the CDVs and a transition from vesicles tofibers is observed by cryo-TEM (Figure 23b). These transi-tions are reversible upon altering the pH to 7.4 (Figure 23c).

The group also demonstrates, through encapsulation of afluorescent dye, that this system can be used as responsivecapsules, allowing pH-triggered release of encapsulatedcargo, opening possibilities for drug delivery.

3.8 Supramolecular replication of peptide andDNA patterned arrays

Stellacci and Stevens outline how ordered peptide arrayscan be printed onto a surface.84 They have implemented asoft stamping technique known as liquid supramolecularnanostamping (LiSuNS), which they used previously toreplicate DNA features onto surfaces.85 In this new work,they extend the scope of the technique to printing peptides.They use a thiol-modified version of the heterodimeric

x

x

y

1 2 3 4

zx

z

Prepolymer

Master Complementarypeptide Complementary

fluorescent peptide

Solid polymer

1) Coiling; 2) Printing; 3) Curing and decoiling; 4) Recoiling(a)

(b) (c)

30 µm 100 µmDN

A

Pep

tide

Figure 24 The LiSuNS process. (a) The printing process; (b) Fluorescence micrograph of the printed peptide array; (c) Fluorescencemicrograph showing the DNA and peptide array using false color overlay. (Reproduced from Ref. 84. Royal Society of Chemistry,2010.)



EK coiled coil introduced by Tripet and Hodges.86 The Kpeptide is incubated with a gold-printed silicon wafer andbinds to its surface. This peptide-functionalized wafer isknown as the “master” and is incubated with free E peptide,resulting in the formation of EK heterodimers at the surface.A poly(dimethyl siloxane) (PDMS) prepolymer solution ismixed with the master, cured, and removed, at which pointthe E peptide should be attached to the PDMS substrate(Figure 24a). To test this, they incubate the preparedPDMS surface with fluorescently labeled K peptide, andimaged by fluorescence microscopy (Figure 24b). In a finaldemonstration of the capabilities of the technique, the groupshows that it is possible to print a mixed peptide–DNAarray (Figure 24c).

4 TEMPLATED AND DIRECTEDASSEMBLIES

The work outlined in the section above demonstratesthat assembly of discrete supramolecular nanostructures isachievable through directed hydrophobic and electrostaticinteractions alone. While some of the examples employmetals, this is to illustrate that their assembled systemsfulfill a particular function; they are not an integral part ofthe designs. The examples in this section employ metals aspart of the assembly process. In many of the cases, assemblyof metal ions or nanoparticles is directed by the peptides;in others, it is the metal that influences organization of thepeptides. Examples of peptide assemblies constructed fromcoiled-coil, collagen-like, and β-structured tectons are allhighlighted.

4.1 Peptide and metal hybrid systems

The majority of examples of hybrid systems are based onβ-structured tectons and have been used as fibrous scaffoldsfor nanowires; assemblies that use other tectons have awider scope, and have been used to produce a range ofinterestingly shaped discrete objects.

4.1.1 Assemblies utilizing β-structured tectons

The Lindquist group was among the first to demonstrate thatamyloid-like fibrils could template nanowire formation.87

They use the so-called NM region of a yeast prion,which self-assembles into β-structured fibers, modifying thepeptide building blocks to contain a chemically accessiblecysteine residue for subsequent covalent attachment ofnanogold particles. TEM demonstrates that nanogold isevenly distributed along the fibers. To make the fibers

NMK184C fibers

Silver enhancement

Gold enhancement

Labeled fiber

Bare fiber500 nm

withcolloidal gold

(a) (b)

Figure 25 The formation of gold nanowires demonstrated byScheibel et al. (a) A schematic of the metal enhancement process;(b) AFM image of the gold-enhanced fiber. (Reproduced withpermission from Ref. 87. National Academy of Sciences,2003.)

conducting, they enhance the gold-decorated nanofibers(Figure 25a), resulting in fibers displaying closely packedmetal nanoparticles (Figure 25b).

Similarly, Reches and Gazit use a straightforwarddiphenylalanine building block as a basis for amyloid-likefibril formation and templating.88 In this case, the dipeptideforms regularly sized hollow nanotubes, visualized by SEM(Figure 26a and b); they propose that the aromatic dipheny-lalanine residues stack to form the higher order assemblies.To exploit the tubes, ionic silver is added to the peptidenanotubes, resulting in silver nanoparticles and, hence, sil-ver nanowires within the tubes after reduction to silvermetal. The nanotubes could then be degraded with protease,leaving only the assembled silver nanowires (Figure 26cand d). These are observed by TEM, and the chemical iden-tity of the nanowires confirmed by energy dispersive X-ray(EDX) analysis.

A more recent example demonstrates how amyloid fibrilscan be used to display cytochromes. A tandem repeat of afibril-forming SH3 domain is fused to a cytochrome via ashort peptidic linking sequence and it is demonstrated thatfibers (albeit with different morphologies to those formedwith the SH3 domains alone) are formed in the presenceof both apo- and holoproteins.89 Such a system has thepotential to be used as conducting nanowires, and alsoas a basis for understanding charge transfer in biologicalsystems.

4.1.2 Coiled-coil assemblies

Several papers detail how coiled coils can be used to directthe assembly of nanoparticles; in particular gold nanopar-ticles are widely used. For instance, following on fromthe aforementioned belt-and-braces system,72 Stevens et al.use an acid–base heterodimeric coiled coil to direct theassembly of different sized gold nanoparticles. They alsodemonstrate that control can be exerted over assembly



(a) (b)

(d)

(c)

Sodiumcitrate

Silver-filled nanotubeSilver nanowire

~20 nmProteinase K

Ag+Ag+ Ag+

Ag+

1 µm 100 nm200 nm

Figure 26 Formation of silver nanowires by Reches and Gazit. (a) TEM image of hollow nanotubes; (b) Magnified image of a singlenanotube; (c) TEM image of a nanowire after protein degradation; (d) The process of nanowire assembly. (Reproduced with permissionfrom Ref. 88. American Association for the Advancement of Science, 2003.)

and disassembly by altering pH.90 Studies with the gold-functionalized acid peptide at basic pH show a dispersepopulation of nanoparticles, whereas at acidic pH aggre-gation occurs, indicative of coiled-coil homodimer forma-tion. Mixing the acid and the base peptide–nanoparticleconjugates at neutral pH gives higher order nanoparticleassemblies. These assemblies are imaged using TEM, whichreveals a central 53 nm gold nanoparticle (attached to thebase peptide) surrounded by a layer of smaller 8.5 nm par-ticles (conjugated to the acid peptides) (Figure 27a). Thenanoparticles are presumably brought together by the for-mation of acid–base heterodimers. On lowering the pH, theassemblies appear more random, presumably because theacidic peptide–nanoparticle conjugates form homodimers(Figure 27b).

Wagner et al. describe something similar, but using aslightly different approach.91 They employ a de novo

100 nm

(a) (b)

100 nm

Figure 27 TEM images of the Au-functionalized nanoparti-cles designed by Stevens et al. (a) TEM image of the peptide-nanoparticle conjugates at pH 7.4; (b) The same conjugates atpH 4.5. (Reproduced from Ref. 90. Wiley-VCH, 2004.)

designed homodimeric coiled coil, VW05, with arginineresidues at the surface-exposed f positions of the heptadrepeat. Under basic conditions, the arginine side chainis protonated, facilitating the formation of electrostaticinteractions with mercaptoundecanoic acid (MUA) coatedgold nanoparticles. The group uses UV–vis spectroscopyto monitor the surface plasmon resonance (SPR) band ofthe Au–MUA nanoparticles. Coiled-coil driven assemblycauses red shift and broadening of the SPR band asthe interparticle distance decreases. By monitoring theassembly process in this way, it is found that not only doespH have an effect on assembly but peptide concentrationand incubation time do also. The group visualizes thenanoparticle assemblies by cryo-TEM and confirms that,at pH 9, ordered networks of nanoparticle assemblies areformed. Very thin fibers are also seen at high magnificationsbut not for the peptide alone, leading to the conclusionthat high local peptide concentrations at the surface ofthe nanoparticle triggers nucleation and fibril formation.Finally, cycling the pH between 9 and 12 can be used tocontrol the assembly of the system.

In a final example, Slocik et al. demonstrate that coiledcoils can be used to direct the assembly of both goldnanoshells (NSs) and quantum dots (QDs), to producetwo types of assembly: extended NS–NS assemblies anddiscrete NS–QD complexes.92 Two designed peptides, E5and K5, form an antiparallel heterodimer, and both peptidesare functionalized with a cysteine residue to allow forcovalent attachment. To form extended NS–NS assemblies,NSs are incubated with either the E5 or the K5 peptide.Confirmation of an interaction between the NS and thepeptide is obtained by FT-IR spectroscopy. Mixing thetwo conjugates (NS-E5 and NS-K5) results in higher order



Assembly

Disassembly

Reassembly

(a)

QD-coil

(f)

(b) (d)

(c) (e)

500 nm

500 nm

430 nm

(g) 100 nm

ωωω

Figure 28 Coiled-coil-mediated NS–NS and NS–QD assemblies demonstrated by Slocik et al. (a) Peptide NS–NS assembly anddisassembly; (b–e) TEM images for each stage of this process; (f) NS–QD assembly; (g) TEM image of the assembled NS–QDsystem. (Reproduced from Ref. 92. American Chemical Society, 2007.)

assemblies as shown by TEM (Figure 28a–c). The imagesobtained confirm that the NSs are indeed bought intoclose proximity, presumably through the formation of theheterodimeric coiled-coil dimer. However, the assemblieshave a tendency to aggregate and form extended structures.They demonstrate triggered dissociation of the assembledNS–NS complexes by irradiation with IR light. Thisprocess is reversible: once the temperature decreases, thecoiled coils recombine and the extended assemblies re-form (Figure 28d and e). To form discrete assemblies,the K5 peptide is incubated with cysteine-functionalizedQDs and mixed with the E5–NS conjugates (Figure 28f ).The QDs organize into uniform layers on the NS surface(Figure 28g). Unlike the NS–NS complexes, the NS–QDcomplexes cannot be photothermally dissembled.

4.1.3 Collagen-based metal–peptide assemblies

Collagen-based fibrous structures are the most recent addi-tions to the field of de novo designed peptidic biomaterials;so it is not surprising that very few examples of collagen–metal assemblies exist. However, Chmielewski and cowork-ers have begun to explore possibilities for such assemblies.Their first example outlines the design of a peptide that

assembles into an extended system of collagen triple helicesupon the addition of divalent metal ions.93 The designis based on a standard (POG)9 sequence, where “O” isthe modified amino acid hydroxyproline. The peptide isC-terminally modified with dihistidine and N-terminallymodified with nitrilotriacetic acid (NTA). CD spectroscopyshows the presence of the extended collagen PPII helix.Addition of Ni(II), Zn(II), Co(II), or Cu(II) to the peptideresults in the solution becoming turbid, and dynamic lightscattering (DLS) reveals large aggregates. The formationof these is probed further to show that the morphology ofthe species formed depends on the ratio of the peptide tothe metal. SEM reveals a wide range of species, rangingfrom florettes to open tubes, composed of layered sheetsand irregularly shaped flakes (Figure 29a–d). Incubationof the peptide with Cu(II) and Co(II) also results in theformation of spherical microflorettes. However, with Ni(II)the structures formed are much smaller, and atomic forcemicroscopy (AFM) analysis shows them to be composed ofinterconnected irregularly nanosized spheres. Assembly isreversible upon the addition of ethylenediaminetetraaceticacid (EDTA). If the collagen helices are functionalized onlyat one end, then higher order assemblies do not form.



(a) (b)

(c) (d)

Figure 29 SEM images indicating the range of structures that can be produced with the collagen-based designed peptides afterincubation with varying concentrations of ZnCl2. (a) 200 µM ZnCl2; (b) 400 µM ZnCl2; (c) 600 µM ZnCl2; (d) 800 µM ZnCl2. Scalebars = 5 µm. (Reproduced from Ref. 93. American Chemical Society, 2009.)

To illustrate potential functions of these assemblies, thegroup has modified the initial design and demonstrated itsability as a 3D cell culture scaffold. The same basic buildingblock is used, but with a bipyridyl (bipy) moiety incorpo-rated at a central position in the peptide.94 Incubation of thispeptide with Zn(II), Co(II), Cu(II), and Ni(II) again givesturbid solutions, and SEM reveals networks of cross-linkedstrands (Figure 30a–d). In this case, the metal-to-peptideratio has no effect on the morphology of the structuresformed, although there is some variation in the morpholo-gies of the fibrous networks depending on the metal used.Two metals can also be incorporated into the peptide frame-work: incubating the peptide with rhuthenium binding to thebipyridyl moiety and then subsequent addition of a diva-lent metal ion binding the NTA and His functionalities. Todemonstrate that the scaffold can be used to display differ-ent molecules, a second peptide that incorporates biotin iscoassembled with the bipyridyl-functionalized peptide andmetal to generate a 3D scaffold. Fluorescently labeled strep-tavadin can then be added, and fluorescence microscopyused to show that the biotin-labeled peptides have beenincorporated into the matrix (Figure 30e and f ).

To demonstrate that the scaffold can be used to supportcell growth, another peptide incorporating a fluorophore ismixed with the bipyridyl-functionalized peptide and metal.HeLa cells are added and incubated in a cell culturemedium. Fluorescence microscopy shows that the HeLacells are encapsulated by the 3D network, and the cells areshown to have similar viability to those grown on normalcell culture plates.

This demonstrates that peptidic collagen-based materialscan support cell growth and potentially be used as scaffoldsin regenerative medicine.

4.1.4 Assemblies using PAs

There are a few reported examples of nanoparticles drivingthe assembly of PA-based systems. One is from Li andStupp, who demonstrate a system where β-structured PAsform linear arrays with gold nanoparticles (Figure 31a).95

Two PAs, one modified with thymine in order to binddiaminopyridine (DAP) modified gold nanoparticles, areused, and upon mixing the two a gel forms. Addition ofDAP-modified nanoparticles leads to aggregated structures



Zn(II)

(a) (b)

(c) (d)

(e) (f)

Cu(II)

Co(II) Ni(II)

Figure 30 Fibrous scaffolds produced by Pires et al. through incubation of the designed peptides with metal. (a) With Zn(II); (b) WithCu(II); (c) With Co(II); (d) With Ni(II); (e) A bright-field image of the network when the biotin-labeled peptide is incorporated andbound to fluorescently labeled streptavidin; (f) A fluorescence microscopy image of (e). (Scale bars in images (a–d) = 5 µm and (e, f)= 200 µm.) (Reproduced from Ref. 94. Wiley-VCH, 2009.)

which are shown by TEM to be long linear arrays of goldnanoparticles forming along the fibers (Figure 31b).

4.2 Conformational state switching

In addition to driving the assembly of peptide-based sys-tems, metals can be employed to induce folding or tocause a change of conformational state within a system.Ogawa’s lab primarily investigates metal coordination tocoiled-coil peptides and presents several examples of met-als inducing folding or conformational changes of designed

peptides. One details a designed four-heptad coiled coilthat binds metal complexes.96 To facilitate metal binding,an f position incorporates 4-pyridyl alanine (Pal) and ad position contains cysteine, allowing for disulfide bondformation. The disulfide-bonded coiled-coil dimers reactwith a rhenium compound resulting in assemblies wherethe dimers act as bridging ligands between fac-[Re(CO)2]cores. Matrix-assisted laser desorption ionization (MALDI)mass spectrometry shows assemblies incorporating one,two, and three ruthenium units.

In a second example, the group uses a similar Pal-containing coiled coil to form noncovalent complexes.97



(a)

1 2

+

(b)

100 nm

Figure 31 Fibers formed from PAs and gold nanoparticles by Li and Stupp. (a) A schematic representation of the formation ofgold-coated amphiphilic fibrils; (b) TEM image showing the regular ordering of gold nanoparticles along the fibrils. (Reproduced fromRef. 95. Wiley-VCH, 2005.)

(a) (b)

76 nm75 nm

Figure 32 AFM images of coiled-coil-metal complexes fromTsurkan et al. AFM image of (a) nanospheres and (b) nanofibrils.(Reproduced from Ref. 97. American Chemical Society, 2007.)

Two coiled coils coordinate to a platinum center and thesemonomeric units spontaneously self-assemble as the coiledcoils form dimers. The morphology of these assembliesis probed by TEM and AFM to show predominantlynanospheres with dimensions of 30–50 nm; in addition,fibrils with lengths of several 100 nm and widths rangingfrom 4 to 10 nm are observed (Figure 32a and b).

A different example of metal coordination from the samegroup details the de novo design of a coiled coil based ona rubredoxin model; the Cys-X–X-Cys binding motif fromrubredoxin is incorporated into a coiled coil.98 Interestingly,the group observes that the peptide adopts a random-coilconformation when no metal is present. They assumedthat the peptide would spontaneously fold into a coiledcoil as the sequence follows the heptad repeat, but theyconclude that incorporation of two cysteine residues intothe hydrophobic core disrupts folding. Upon addition ofCdCl2, however, an α-helical structure is observed and thepeptide is shown to bind the metal in 2 : 1 ratio indicatingthat a metal-bridged coiled-coil dimer is formed.

A more recent paper gives design principles from bothtypes of example. A coiled coil containing two Pal residuescoordinates to a platinum center.99 A coiled-coil dimer isformed when metal is absent but, on reacting the peptidewith Pt(en)(NO3)2, a metal-bridged coiled-coil tetramer isobserved; this is proven by reversed-phase HPLC (high-performance liquid chromatography), SDS-PAGE (sodiumdodecyl sulfate polyacrylamide gel electrophoresis), andMALS (multiangle light scattering) analysis.

Tanaka’s lab has also produced assemblies comprisingcoiled coils and metals. They had previously designed aparallel trimer IZ,100 which is modified by incorporatinghistidines at a and d positions to produce IZ-3adH. In theabsence of metal, the peptide adopts a random-coil confor-mation, but on addition of Co(II), Ni(II), or Zn(II) formsα-helical parallel coiled-coil trimers. Further investigationof the Ni(II)–peptide complex shows that Ni(II) coordinatesto the histidine residues with an octahedral geometry.

In a later paper, they again use IZ, but change oneisoleucine to alanine and a second to cysteine. They showthe resulting peptide once again has a random-coil structurein the absence of metal, but upon addition of either Cd(II)or Hg(II) adopts a trimeric coiled-coil structure.101

Dublin and Conticello use the trimeric histidine-containing coiled-coil TZ1H, which is previously shownto form long fibers at pH values above the pKa of histi-dine, to bind metals.53, 102 They postulate that the histidineresidues could be arranged in a trigonal planar geometrythat allows metal binding. They add silver triflate to thepeptide and show that it changes conformation from ran-dom coil to α-helix. Further analysis by TEM shows that,in the absence of Ag(I) and at pH 5.6, no higher orderassemblies are observed but addition of Ag(I) results inlong fibers indicating metal coordination is driving fiberformation (Figure 33).

There are also examples from our own lab of peptidesdesigned to switch conformational state. One describes



100 nm

Figure 33 A TEM image of the fibers formed upon addition ofsilver triflate to the peptide designed by Dublin et al. (Reproducedfrom Ref. 102. American Chemical Society, 2008.)

Template-α, which is a designed coiled coil.103 To encour-age a conformational switch, residues at f positions werechanged from glutamine to threonine to produce Template-αT. Below 70 ◦C, the peptide is α-helical and abovethis temperature it forms β-structured amyloid-like fibrils.Cross-linking the peptides to achieve a β-hairpin-like con-formation increases amyloid fibril formation.

A later paper details the de novo design of peptidesthat switch between a coiled-coil and a helical-hairpinconformation.104 The parent peptide (coiled-coil switchpeptide) CSP-1 is a parallel dimeric coiled coil, butoxidation of cysteine residues in the peptide causes theswitch to a monomer. It remains α-helical, although theα-helical content is low. Increasing the loop length givesCSP-3, which is a peptide with a higher helical content.CSP-6, an anagram of CSP-3, is more helical still andcan be switched between the helical-hairpin and coiled-coilconformations.

A final example from our lab details how ZiCo, anotherde novo designed peptide, can switch conformational stateupon binding zinc.105 In the absence of zinc, the peptideadopts a coiled-coil structure, but upon addition of zinc,absorption bands indicative of β-sheet are observed by FT-IR spectroscopy. Zinc binding is reversible and occurs in a1 : 1 ratio.

Similarly, Ambroggio and Kuhlman demonstrate thede novo design of a peptide able to switch between a2Cys-2His zinc finger-like fold and a trimeric coiled coil(Figure 34).106 They use the sequences of the zinc-fingerZif268 and the coiled-coil region of hemagglutinin in acomputer algorithm designed to find the lowest energycombination of these sequences. Their design, Sw2, isfolded as a trimeric coiled coil in the absence of zinc. Uponthe addition of zinc, the peptide adopts a conformationsimilar to that of other zinc fingers. The peptide and metalbind in a 1 : 1 ratio, regardless of whether zinc or cobaltis used; however, binding is reversible only with cobalt,indicating zinc binds strongly and the off rates are slow.

N

NN

N

C

CC

−Zn(II)

+Zn(II)

Figure 34 Computer-generated model of the Sw2 sequencethreaded onto a coiled coil (left) and a zinc-finger (right)backbone. (Reproduced from Ref. 106. American ChemicalSociety, 2006.)

Lombardi et al. give a more complicated example ofa de novo designed peptide that self-assembles in thepresence of tetrahedrally coordinating metal ions.107 Theyuse a rubredoxin model as a starting point to generate anundecapeptide containing the Cys-X–X-Cys binding motif.In the absence of metals, the peptide is unfolded but, uponthe addition of a wide variety of divalent metal ions, thepeptide is shown to adopt a β-turn conformation. Theydemonstrate that the peptide binds the metal in a 2 : 1stoichiometry and that the peptides organize themselves insuch a way as to bind the metal with the desired tetrahedralcoordination geometry.

5 CONCLUSIONS AND OUTLOOK

It is evident from the large number of working examplesthat the design and engineering principles for effec-tively infinite, fibrous, peptide-based assemblies are nowwell established, and research in this area is burgeoning.This is particularly true for systems based on α-helical,β-structured, and PA tectons. Largely because of diffi-culties with peptide synthesis and slow folding kinetics,similar fibrous assemblies based on collagen-like build-ing blocks are less-well developed, but recent work onthese is extremely encouraging. Challenges that remain forsuch peptide-based assemblies in general include takingprecise control over assembly kinetics, and of fiber mor-phology, functionalization, and dynamics; again, we notegood progress on all of these fronts. Thus, in our view,polypeptide-based materials are now at a stage to competewith more traditional fibrous systems based on syntheticpolymers and small molecules: if not in respect of costand bulk preparation, at least in terms of rational design.



This improved ability to design peptide-based fibers fromthe bottom up should pave the way to real-life applicationsof such materials in bionanotechnology, particularly withregard to templating inorganic and other functional materi-als, and as soft biomimetic scaffolds in 3D cell culture andtissue engineering.

In contrast, the rational design of discrete supramolecularassemblies based on peptides is more elusive. True, thegeneration of simple discrete oligomers has been possiblefor some time using certain peptide-folding motifs—forexample, the α-helical coiled coil. However, with notableexceptions, the next steps to more complex multicomponentsystems of defined stoichiometry and topology are provingmore difficult. Again, the issue seems to be one of control:how can peptidic tectons be encouraged to assemble, butat the same time how can such associations be limited anddirected precisely to give discrete rather than infinite objects(i.e., fibers and aggregates)? Of course, this is exciting initself, and it is encouraging that ambitious de novo designsand engineered systems are already emerging. Moreover,the prospects for bottom-up assembly (tiling), drug delivery,and biosensing, to name but a few examples, with suchassemblies in hand provide strong motivation for exploringthis avenue of research.

Therefore, on both fronts—functional soft fibrous assem-blies and discrete nanoscale objects—the rational designof peptide-based systems in what might be termed morebroadly as biosupramolecular chemistry, is alive, kicking,and screaming to be nurtured, developed, and applied.

REFERENCES

1. E. H. C. Bromley, K. Channon, E. Moutevelis, andD. N. Woolfson, Acs. Chem. Biol., 2008, 3, 38.

2. K. Channon, E. H. C. Bromley, and D. N. Woolfson, Curr.Opin. Struct. Biol., 2008, 18, 491.

3. L. Jiang, E. A. Althoff, F. R. Clemente, et al. Science,2008, 319, 1387.

4. D. Rothlisberger, O. Khersonsky, A. M. Wollacott, et al.Nature, 2008, 453, 190.

5. F. H. C. Crick, Acta Crystallogr., 1953, 6, 689.

6. J. Walshaw and D. N. Woolfson, J. Mol. Biol., 2001, 307,1427.

7. D. N. Woolfson, Fibrous Proteins: Coiled-Coils, Collagenand Elastomers, Elsevier Academic Press Inc., San Diego,CA, 2005, vol. 70, p. 79.

8. M. R. Hicks, D. V. Holberton, C. Kowalczyk, and D. N.Woolfson, Fold. Des., 1997, 2, 149.

9. P. B. Harbury, T. Zhang, P. S. Kim, and T. Alber, Science,1993, 262, 1401.

10. C. K. Smith and L. Regan, Science, 1995, 270, 980.

11. C. K. Smith and L. Regan, Acc. Chem. Res., 1997, 30, 153.

12. B. M. Broome and M. H. Hecht, J. Mol. Biol., 2000, 296,961.

13. K. Rajagopal and J. P. Schneider, Curr. Opin. Struct. Biol.,2004, 14, 480.

14. R. M. Hughes and M. L. Waters, Curr. Opin. Struct. Biol.,2006, 16, 514.

15. M. A. Wouters and P. M. G. Curmi, Proteins, 1995, 22,119.

16. E. G. Hutchinson, R. B. Sessions, J. M. Thornton, andD. N. Woolfson, Protein Sci., 1998, 7, 2287.

17. H. M. Fooks, A. C. R. Martin, D. N. Woolfson, et al.J. Mol. Biol., 2006, 356, 32.

18. A. G. Cochran, N. J. Skelton, and M. A. Starovasnik, Proc.Natl. Acad. Sci. U.S.A., 2001, 98, 5578.

19. S. J. Russell, T. Blandl, N. J. Skelton, and A. G. Cochran,J. Am. Chem. Soc., 2003, 125, 388.

20. J. R. Silveira, G. J. Raymond, A. G. Hughson, et al.Nature, 2005, 437, 257.

21. M. Rasmussen, M. Jacobsson, and L. Bjorck, J. Biol.Chem., 2003, 278, 32313.

22. H. Rapaport, G. Moller, C. M. Knobler, et al. J. Am. Chem.Soc., 2002, 124, 9342.

23. C. Whitehouse, J. Y. Fang, A. Aggeli, et al. Angew. Chem.,Int. Ed. Engl., 2005, 44, 1965.

24. S. Cavalli, F. Albericio, and A. Kros, Chem. Soc. Rev.,2010, 39, 241.

25. H. G. Cui, M. J. Webber, and S. I. Stupp, Biopolymers,2010, 94, 1.

26. D. N. Woolfson, Biopolymers, 2010, 94, 118.

27. I. Cherny and E. Gazit, Angew. Chem., Int. Ed. Engl., 2008,47, 4062.

28. S. K. Gunasekar, J. S. Haghpanah, and J. K. Montclare,Polym. Adv. Technol., 2008, 19, 454.

29. S. G. Zhang, T. Holmes, C. Lockshin, and A. Rich, Proc.Natl. Acad. Sci. U.S.A., 1993, 90, 3334.

30. S. G. Zhang, T. C. Holmes, C. M. Dipersio, et al. Bioma-terials, 1995, 16, 1385.

31. A. Aggeli, M. Bell, N. Boden, et al. Nature, 1997, 386,259.

32. A. Aggeli, I. A. Nyrkova, M. Bell, et al. Proc. Natl. Acad.Sci. U.S.A., 2001, 98, 11857.

33. K. Rajagopal, M. S. Lamm, L. A. Haines-Butterick, et al.Biomacromolecules, 2009, 10, 2619.

34. R. P. Nagarkar, R. A. Hule, D. J. Pochan, and J. P. Schnei-der, J. Am. Chem. Soc., 2008, 130, 4466.

35. R. A. Hule, R. P. Nagarkar, A. Altunbas, et al. FaradayDiscuss., 2008, 139, 251.

36. R. V. Rughani, D. A. Salick, M. S. Lamm, et al. Bio-macromolecules, 2009, 10, 1295.

37. L. Haines-Butterick, K. Rajagopal, M. Branco, et al. Proc.Natl. Acad. Sci. U.S.A., 2007, 104, 7791.

38. D. A. Salick, J. K. Kretsinger, D. J. Pochan, and J. P.Schneider, J. Am. Chem. Soc., 2007, 129, 14793.



39. M. R. Ghadiri, J. R. Granja, R. A. Milligan, et al. Nature,1993, 366, 324.

40. D. T. Bong, T. D. Clark, J. R. Granja, and M. R. Ghadiri,Angew. Chem., Int. Ed. Engl., 2001, 40, 988.

41. L. Motiei, S. Rahimipour, D. A. Thayer, et al. Chem. Com-mun. (Camb.), 2009, 3693.

42. V. Dartois, J. Sanchez-Quesada, E. Cabezas, et al. Antimi-crob. Agents Chemother., 2005, 49, 3302.

43. S. Fernandez-Lopez, H. S. Kim, E. C. Choi, et al. Nature,2001, 412, 452.

44. E. Gazit, Chem. Soc. Rev., 2007, 36, 1263.

45. E. Gazit, Prion, 2007, 1, 32.

46. R. V. Ulijn and A. M. Smith, Chem. Soc. Rev., 2008, 37,664.

47. S. Kojima, Y. Kuriki, T. Yoshida, et al. Proc. Jpn. Acad.,Ser. B, Phys. Biol. Sci., 1997, 73, 7.

48. S. Kojima, Y. Kuriki, K. Yazaki, and K. Miura, Biochem.Biophys. Res. Commun., 2005, 331, 577.

49. S. A. Potekhin, T. N. Melnik, V. Popov, et al. Chem. Biol.,2001, 8, 1025.

50. T. N. Melnik, V. Villard, V. Vasiliev, et al. Protein Eng.,2003, 16, 1125.

51. V. Villard, O. Kalyuzhniy, O. Riccio, et al. J. Pept. Sci.,2006, 12, 206.

52. Y. Zimenkov, V. P. Conticello, L. Guo, and P. Thiya-garajan, Tetrahedron, 2004, 60, 7237.

53. Y. Zimenkov, S. N. Dublin, R. Ni, et al. J. Am. Chem. Soc.,2006, 128, 6770.

54. D. E. Wagner, C. L. Phillips, W. M. Ali, et al. Proc. Natl.Acad. Sci. U.S.A., 2005, 102, 12656.

55. K. L. Lazar, H. Miller-Auer, G. S. Getz, et al. Biochem-istry, 2005, 44, 12681.

56. E. H. Bromley, R. B. Sessions, A. R. Thomson, and D. N.Woolfson, J. Am. Chem. Soc., 2009, 131, 928.

57. H. Dong, S. E. Paramonov, and J. D. Hartgerink, J. Am.Chem. Soc., 2008, 130, 13691.

58. M. J. Pandya, G. M. Spooner, M. Sunde, et al. Biochem-istry, 2000, 39, 8728.

59. A. M. Smith, E. F. Banwell, W. R. Edwards, et al. Adv.Funct. Mater., 2006, 16, 1022.

60. D. Papapostolou, E. H. C. Bromley, C. Bano, and D. N.Woolfson, J. Am. Chem. Soc., 2008, 130, 5124.

61. D. Papapostolou, A. M. Smith, E. D. T. Atkins, et al. Proc.Natl. Acad. Sci. U.S.A., 2007, 104, 10853.

62. E. H. C. Bromley, K. J. Channon, P. J. S. King, et al.Biophys. J., 2010, 98, 1668.

63. M. G. Ryadnov and D. N. Woolfson, Nat. Mater., 2003, 2,329.

64. M. G. Ryadnov and D. N. Woolfson, Angew. Chem., Int.Ed. Engl., 2003, 42, 3021.

65. E. F. Banwell, E. S. Abelardo, D. J. Adams, et al. Nat.Mater., 2009, 8, 596.

66. F. W. Kotch and R. T. Raines, Proc. Natl. Acad. Sci.U.S.A., 2006, 103, 3028.

67. V. Gauba and J. D. Hartgerink, J. Am. Chem. Soc., 2007,129, 2683.

68. S. Rele, Y. H. Song, R. P. Apkarian, et al. J. Am. Chem.Soc., 2007, 129, 14780.

69. O. D. Krishna and K. L. Kiick, Biomacromolecules, 2009,10, 2626.

70. J. D. Hartgerink, E. Beniash, and S. I. Stupp, Science,2001, 294, 1684.

71. L. Aulisa, N. Forraz, C. McGuckin, and J. D. Hartgerink,Acta Biomater., 2009, 5, 842.

72. M. G. Ryadnov, B. Ceyhan, C. M. Niemeyer, and D. N.Woolfson, J. Am. Chem. Soc., 2003, 125, 9388.

73. M. Zhou, D. Bentley, and I. Ghosh, J. Am. Chem. Soc.,2004, 126, 734.

74. M. Zhou and I. Ghosh, Org. Lett., 2004, 6, 3561.

75. E. K. Oshea, R. Rutkowski, W. F. Stafford, and P. S. Kim,Science, 1989, 245, 646.

76. S. Raman, G. Machaidze, A. Lustig, et al. Nanomedicine,2006, 2, 95.

77. S. A. Kaba, C. Brando, Q. Guo, et al. J. Immunol., 2009,183, 7268.

78. T. Pimentel, Z. Yan, S. A. Jeffers, et al. Chem. Biol. DrugDes., 2009, 73, 53.

79. U. Schroeder, A. Graff, S. Buchmeier, et al. J. Mol. Biol.,2009, 386, 1368.

80. F. Boato, R. M. Thomas, A. Ghasparian, et al. Angew.Chem., Int. Ed. Engl., 2007, 46, 9015.

81. M. G. Ryadnov, Angew. Chem., Int. Ed. Engl., 2007, 46,969.

82. H. R. Marsden, N. A. Elbers, P. H. H. Bomans, et al.Angew. Chem., Int. Ed. Engl., 2009, 48, 2330.

83. F. Versluis, I. Tomatsu, S. Kehr, et al. J. Am. Chem. Soc.,2009, 131, 13186.

84. A. Laromaine, O. Akbulut, F. Stellacci, and M. M. Stevens,J. Mater. Chem., 2010, 20, 68.

85. A. A. Yu and F. Stellacci, Adv. Mater., 2007, 19, 4338.

86. B. Tripet, L. Yu, D. L. Bautista, et al. Protein Eng., 1996,9, 1029.

87. T. Scheibel, R. Parthasarathy, G. Sawicki, et al. Proc. Natl.Acad. Sci. U.S.A., 2003, 100, 4527.

88. M. Reches and E. Gazit, Science, 2003, 300, 625.

89. A. J. Baldwin, R. Bader, J. Christodoulou, et al. J. Am.Chem. Soc., 2006, 128, 2162.

90. M. M. Stevens, N. T. Flynn, C. Wang, et al. Adv. Mater.,2004, 16, 915.

91. S. C. Wagner, M. Roskamp, H. Colfen, et al. Org. Biomol.Chem., 2009, 7, 46.

92. J. M. Slocik, F. Tam, N. J. Halas, and R. R. Naik, NanoLett., 2007, 7, 1054.

93. M. M. Pires and J. Chmielewski, J. Am. Chem. Soc., 2009,131, 2706.

94. M. M. Pires, D. E. Przybyla, and J. Chmielewski, Angew.Chem., Int. Ed. Engl., 2009, 48, 7813.



95. L. S. Li and S. I. Stupp, Angew. Chem., Int. Ed. Engl.,2005, 44, 1833.

96. M. V. Tsurkan and M. Y. Ogawa, Chem. Commun., 2004,2092.

97. M. V. Tsurkan and M. Y. Ogawa, Biomacromolecules,2007, 8, 3908.

98. O. A. Kharenko and M. Y. Ogawa, J. Inorg. Biochem.,2004, 98, 1971.

99. M. V. Tsurkan and M. Y. Ogawa, Inorg. Chem., 2007, 46,6849.

100. K. Suzuki, H. Hiroaki, D. Kohda, and T. Tanaka, ProteinEng., 1998, 11, 1051.

101. X. Q. Li, K. Suzuki, K. Kanaori, et al. Protein Sci., 2000,9, 1327.

102. S. N. Dublin and V. P. Conticello, J. Am. Chem. Soc., 2008,130, 49.

103. B. Ciani, E. G. Hutchinson, R. B. Sessions, and D. N.Woolfson, J. Biol. Chem., 2002, 277, 10150.

104. M. J. Pandya, E. Cerasoli, A. Joseph, et al. J. Am. Chem.Soc., 2004, 126, 17016.

105. E. Cerasoli, B. K. Sharpe, and D. N. Woolfson, J. Am.Chem. Soc., 2005, 127, 15008.

106. X. I. Ambroggio and B. Kuhlman, J. Am. Chem. Soc., 2006,128, 1154.

107. A. Lombardi, D. Marasco, O. Maglio, et al. Proc. Natl.Acad. Sci. U.S.A., 2000, 97, 11922.


Date post:	17-Aug-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

Rational Design of Peptide-Based Biosupramolecular SystemsRational Design of Peptide-Based...

Documents