Dynamic peptide libraries for the discovery of supramolecular nanomaterials Charalampos G. Pappas 1,2 , Ramim Shafi 1 , Ivan R. Sasselli 2 , Henry Siccardi 1 , Tong Wang 3 , Vishal Narang 1 , Rinat Abzalimov 1 , Nadeesha Wijerathne 1,4,5 and Rein V. Ulijn 1,2,4,5 * Sequence-specific polymers, such as oligonucleotides and peptides, can be used as building blocks for functional supramolecular nanomaterials. The design and selection of suitable self-assembling sequences is, however, challenging because of the vast combinatorial space available. Here we report a methodology that allows the peptide sequence space to be searched for self-assembling structures. In this approach, unprotected homo- and heterodipeptides (including aromatic, aliphatic, polar and charged amino acids) are subjected to continuous enzymatic condensation, hydrolysis and sequence exchange to create a dynamic combinatorial peptide library. The free-energy change associated with the assembly process itself gives rise to selective amplification of self-assembling candidates. By changing the environmental conditions during the selection process, different sequences and consequent nanoscale morphologies are selected. T he functionality of living systems is based on sequence- specific polymers, and such materials can also be used to create artificial structures that have a range of potential applications 1–5 . Peptides and proteins are of particular interest because of their chemical versatility, biodegradability and biocom- patibility 6–10 . Remarkably, it has been shown that even the shortest peptides—consisting of two or three amino acids—can act as powerful self-assembly motifs 11–15 . Short peptide assemblies can give rise to gelation and emulsion stabilization, as well as to func- tions not normally associated with biological molecules, which include semiconductivity, piezoelectricity, visible light fluorescence and exceptional mechanical properties 15–17 . The supramolecular consequences of exchanging single amino acids in short peptide assemblies are significant. For example, although diphenylalanine forms the well-known peptide nano- tubes 11 , isoleucine–phenylalanine forms nanofibrous hydrogels 18 and tryptophan–phenylalanine assembles into nanospheres 16 . Changes in the environmental conditions can also dramatically impact morphology in short-peptide self-assembly, with non- aqueous solvents seen to influence the nanoscale morphology of alanine/valine dipeptide assemblies 14 . Given the combinatorial complexity that is present even in short peptide sequences, it is difficult to design self-assembly sequences from conventional governing principles. The desire to explore the biopolymer sequence space to identify functional sequences has led to the development of a range of com- binatorial screening methods available for the selection of candi- dates from vast sequence libraries 19–24 . These approaches typically focus on the identification of binders and are not suited to the dis- covery of supramolecular nanostructures. Consequently, the identi- fication of biologically based supramolecular structures has either relied on chance discoveries, or has been based on rational mimicry, simplification and adaptation of known biological sequences 8,25,26 . More recently, coarse-grained molecular dynamics simulations have been used to search the tripeptide sequence space for self-assembling peptides 13 . Dynamic combinatorial libraries—in which multiple com- ponents can reversibly combine and exchange and ultimately favour the structure with the lowest free energy 27–29 (as shown sche- matically in Fig. 1a)—are well suited for the discovery of new self- assembling structures. The approach has been used for the reversible exchange of preformed peptide modules 30–32 , but this does not allow for the diversity of peptides to be exploited in full, as the module- exchange approach does not exchange the amide bond that links amino acids together. The use of (enzymatic) amide hydrolysis/con- densation has been demonstrated, though this suffers from the unfavourable thermodynamics associated with peptide formation in water, with hydrolysis being predominantly observed 33 . We demonstrated previously how this bias can be overcome by taking advantage of supramolecular gelation, using the enzymatic exchange of peptides appended with large aromatic moieties to favour their fibrillar assembly and gelation 34 . Alternatively, oligo- merization can be achieved by pre-selection of precursors that are known to favour the formation of β-sheet structures 35,36 . Other (non-enzymatic) ways to overcome the bias for hydrolysis involve the use of reversible native chemical ligation 37 , although this approach is limited to the exchange of amide bonds at cysteine resi- dues, or to the use of sequential hydration/dehydration steps 38 , which relies on extreme temperatures that are not compatible with self-assembly. We have now developed searchable dynamic peptide libraries based on the sequence exchange of (mixtures of) unprotected pep- tides under user-defined conditions. This is achieved by first select- ing (mixtures of) dipeptides as the input (Fig. 1b, input dyads). On the addition of a relatively nonspecific endoprotease thermolysin, a dynamic exchange of peptide sequences is initiated (exchange and selection (Fig. 1b) by enzymatic condensation, hydrolysis and trans- acylation). The choice of an endoprotease (which does not cleave terminal amide bonds) prevents hydrolysis of the internal amide bond in dyads, and thus disfavours direct hydrolysis of the starting materials (shown in red in Fig. 1a). Over time, the peptides that give rise to the formation of self-assembling structures are amplified 1 Advanced Science Research Center (ASRC), City University of New York, 85 St Nicholas Terrace, New York 10031, USA. 2 WestCHEM/Department of Pure & Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow G1 1XL, UK. 3 Imaging Facility of CUNY ASRC, 85 St Nicholas Terrace, New York 10031, USA. 4 Department of Chemistry, Hunter College, City University of New York, 695 Park Avenue, New York 10065, USA. 5 The Graduate Center of the City University of New York, New York 10016, USA. *e-mail: rein.ulijn@asrc.cuny.edu ARTICLES PUBLISHED ONLINE: 3 OCTOBER 2016 | DOI: 10.1038/NNANO.2016.169 NATURE NANOTECHNOLOGY | VOL 11 | NOVEMBER 2016 | www.nature.com/naturenanotechnology 960 © 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.