Chapter 7 Protein Self-Assembly: From Programming Arrays to Bioinspired Materials Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 Quan Luo, Tiezheng Pan, Yao Liu, and Junqiu Liu* State Key Laboratory of Supramolecular Structure and Materials, College of Chemistry, Jilin University, 2699 Qianjin Street, Changchun 130012, P. R. China *E-mail: firstname.lastname@example.org Natural protein assemblies are intriguing soft materials that exhibit highly complex hierarchical structures and collective properties for many important biological functions, which make them difficult to mimic synthetically. This chapter summarizes the recent advances in the research field of protein assembly and highlights several innovative design strategies for precise manipulation of proteins into extended-, periodic arrays with desired morphologies. These artificially created protein nano/microstructures allow for further functionalization and serve as a versatile platform to create a wide variety of biomaterials by coupling their well-defined architectures with different functional groups, molecules, or nanoparticles for diverse applications. Mimicking the structures and properties of natural protein materials through hierarchical protein self-assembly may help to unravel the complexity and diversity of the protein aggregation process and develop a new generation of biomimetic materials. This will provide a valuable source of inspiration for future design of novel biomaterials and accelerate its development towards high precision, efficiency, and multifunctionality. Introduction Natural soft materials, such as cellular membranes, the cytoskeleton and surface shell/wall structures, are built in a mild environment via spontaneous © 2017 American Chemical Society Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 organization processes (1–3). They are comprised of different biomarcomolecules that form highly complex hierarchical architectures across multiple dimensions. Proteins are one of the most versatile constituents, consisting of 20 standard amino acids that can fold into a specific conformation to meet physiological demands (4). Protein molecules not only function well alone, but also assemble into exquisite superstructures for many important cellular functions (5). The driving forces for the self-assembly process are multiple, weak non-covalent interactions (6), which endow protein-based biomaterials with a large amount of biological/chemical versatility through dynamically reversible regulation of their collective properties for high performance and environmental adaptation (7). Inspired by the elegance of nature’s bottom-up fabrication, the development of artificially created protein assemblies presents an imminent challenge. This offers a great opportunity to develop novel functional biomaterials through mimicry and to gain valuable insights into biological assembly processes. Since supramolecular techniques have emerged as a powerful method for controlling molecular aggregation (8), the application of the principles of molecular recognition opens up tremendous possibilities for small organic molecules to design a wide variety of biomimetic nano/microstructures with promising properties such as processability, self-healing, recyclability, and stimuli-responsiveness (9, 10). However, the research field of direct manipulation of protein self-assemblies is still in its infancy due to the structural complexity, heterogeneity, and instability of protein molecules (11, 12). In addition, the extremely rich biodiversity leads to limited knowledge of their underlying assembly mechanism. Unlike small-molecule recognition, the high fidelity of protein association relies on a more extensive binding interface that contains complementary contacts to induce and stabilize the specific binding between two protein subunits (13). How to precisely control the relative orientation and multi-specificity of these complex building blocks is a key step in the construction of biomimetic nano/microstructure. The design protocol for dictating protein self-assembly is to employ genetic engineering or site-specific chemical modification to introduce supramolecular self-associating components at the protein interfaces, which attempt to control or interfere with protein-protein recognition processes so as to achieve the desired self-assembled architectures (14, 15). In addition, some experimentally determined oligomeric protein domains were utilized as a connector to create 1D (e.g. nanowires), 2D (e.g. the layered structures), and 3D (e.g. cage-like structures) periodic protein arrays through genetic fusion of them along the shared symmetry axes (16). The overall geometries of these protein assemblies can be predetermined by the intended relative orientation of neighboring pairs. The intent of this chapter is to describe the recent developments in the field of protein self-assembly. The strategies that drive protein recognition and aggregation to construct highly ordered hierarchical superstructures have been summarized. Furthermore, some typical examples are discussed with an emphasis on the design principle, self-assembly process, characterization method, selective modification, and surface functionalization. We hope these research details could provide guidance for the future design of bioinspired protein nano/micromaterials for a wide range of biomedical applications. 130 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 Proteins as a Building Block for Multidimensional Nanofabrication Naturally oligomeric protein domains provide a variety of basic building blocks for self-assembly into multidimensional biomaterials. Depending on the structural features, binding orientation and stoichiometric ratios, neighboring protein subunits hold together in the formation of particular symmetric superstructures to promote their biological functions (Figure 1) (17). For example, collagen can be assembled into cross-striated microfibrils to provide structural support for cells (18); S-layer proteins are known to form highly porous self-assembled meshworks to enhance mechanical and osmotic stabilization (19); viral capsids enable selective packaging of the genome into their tubular, globular, or polyhedral nanostructures (20). These structure-function correlations give important clues about the mechanisms behind how different protein components accomplish recognition and aggregation, as well as collective properties and functions under specific environmental conditions. Figure 1. Natural protein assemblies with multidimensional superstructures. (a) Linear collagen; (b) Planar S-layer (c,d) tubular and polyhedral capsids. Initially, great efforts have been made to reproduce some complex protein self-assembly processes in vitro for a better understanding of protein-protein interactions (PPIs) and their underlying assembly principles, which may be helpful for rational design and preparation of artificial protein assemblies. Typically, virus systems such as the tobacco mosaic virus (TMV), cowpea chlorotic mottle virus (CCMV), and bacteriophage φ6 have been studied to explore their nucleation and growth processes (21–23). These bioinspired nanotechnologies offer a facile and environment-friendly approach to develop large-scale protein biomaterials. In this design, well-characterized oligomeric protein domains or motifs can be employed to provide control over the relative orientation of the fusion partners. Symmetry-matching fusion, peptide-mediated assembly, and protein template-induced strategies are classified into this category. Following the above-mentioned studies, de novo design of new protein interaction surfaces made it possible to modulate the binding mode of existing protein assemblies and to dictate the nonself-associating protein monomers self-assembly into desired architectures (24). Several other strategies, such as metal-coordination-driven strategy, electrostatic-interaction-induced strategy, receptor-ligand interaction-directed strategy, host-guest recognition-driven 131 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 strategy, etc. allow the assembly design with greater structural complexity and diversity through site-directed mutagenesis or chemical modification of the relatively reactive groups on the protein surface to create completely new PPIs (14). Although many biomimetic nano/microstructures have been built successfully based on these strategies, how to duplicate the highly intricate morphologies and unique properties of natural systems is still one of the most difficult challenges. This will undoubtedly require a set of rational design rules and a combination of two or more innovative strategies to generate more complex protein assemblies with customized characteristics. In fact, the ability to create sophisticated nano/microstructure is only a first-step toward protein-based nanofabrication. The rapid development of this field provides a rich source from which to design versatile functionalities. A variety of specific groups, molecules, nanoparticles with chemical, electronic, or photonic properties can be precisely positioned on artificially created supramolecular protein scaffolds for further functionalization (25). Many promising applications were realized by coupling highly ordered protein superstructures with diverse functions of the modified components. On the other hand, the well-ordered protein nano/microstructures offer the opportunity to reinforce its innate capabilities via the favorable microenvironment formed by self-assembly (15). The available protein assemblies often exhibit significantly improved performance than that of the mixture of the monomeric forms, which is an alternative way to import new function into self-assembled protein materials with high efficiency. In the following sections, we will present the most significant examples that use different design strategies to create extended 1D, 2D and 3D protein arrays with increasing size and complexity. By mimicking the structure and function of natural protein assemblies, researchers have an unprecedented opportunity to unravel the mystery behind their inherent properties of robustness, strength, elasticity, adaptability, and multifunctionality. The challenge to translate these bioinspired design concepts into improvements in biomimetic model systems would bring practical benefits and be useful for application purposes. Protein Self-Assembly To Create 1D Array Fibrous protein materials (e.g. collagen, elastin, fibronectin, and laminin) are commonly characterized by superior mechanical properties, which makes them a promising biomimetic target for the design of biomaterials (26). Using nature’s design principles as an inspiration, a variety of polymeric protein assemblies have been generated by modular self-assembly of repeat protein units in a head-to-tail or end-to-end manner. Natural Self-Complementary Pairs To Induce Linear Protein Self-Assembly The direct utilization of the symmetry characteristic of naturally oligomeric protein domains or motifs is first considered as a reliable strategy for the 132 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 synthesis of protein nanowires. Rational alignment of a common symmetry axis between neighboring building blocks can determine the gross morphology of a protein assembly through highly specific, directional and strong interactions of the pre-existing protein interface. An early example that uses the genetically oligomeric fusion method to prepare protein filaments has been reported by the Yeates’ group (27). The relative orientation between dimeric M1 matrix protein and carboxylesterase can be controlled by a semi-rigid helical linker that holds the two protein domains together via genetic fusion according to a particular design rule that their symmetry axes do not intersect. As predicted, a self-complementary dimer was formed at each terminus of the designed fusion protein and the iterative dimerization resulted in the formation of large-scale linear protein homopolymer. Transmission electron microscopy (TEM) analysis showed the presence of protein filaments with a width of ~4 nm, which can further organize into higher-order bundles or network structures to mimic the self-assembly processes of long fibers found in Nature. Instead, some stable and reliable protein folding motifs can be fused to the N/C-terminus of target proteins to induce protein-protein association. As we know, secondary structures such as α-helices are the elementary units of natural protein domains, a few of which can be accurately predicted and designed to achieve the intended interaction specificity. The combinatorial design rules of amino acids in the peptide sequences allow for more diverse and minimized interactions as the driving force to drive protein self-assembly. Thus, genetic engineering of a short α-helical peptide into proteins opens up a new possibility to control nonself-associating protein dimerization via coiled-coil interactions. This strategy has been successfully applied to construct a series of heteroor homodimeric functional protein complexes (28, 29). Subsequent research has demonstrated that coiled-coil-mediated self-assembly was able to fabricate larger-scale linear polyprotein chains when the protein units were difunctionalized with two coiled-coil binding partners at both of its termini. For instance, Rief et al. prepared a protein-peptide fusion domain that contained Ig27 protein flanked by two 35-amino acid coiled-coil polypeptide (LZ10) (30). The strong binding between homodimeric LZ10 connected I27 domains to ensure the head-to-tail parallel orientation, and this interaction was further enhanced by incorporating a cysteine residue at a proper site in the LZ10 sequence to stabilize the coiled coil via disulfide bond formation. Size-exclusion chromatographic analysis confirmed that more than 30 coiled coil linked Ig27 monomers were assembled to form the polymeric protein chains. The mechanical properties of individual chains were examined by atomic force microscopy (AFM), in which the unzipping and overstretching transitions of a single coiled coil occurred at 10 pN and 25 pN respectively when exposed to a stretching force. These findings revealed the potential of coiled coil motifs in supramolecular polymerization of protein molecules for developing soft fibrous biomaterials with high elasticity and strength. Nature has also inspired the development of other self-complementary pairs for protein multimerization based on metal coordination, sequence-specific dsDNA recognition, intermolecular disulfide linkage, and receptor-ligand interaction. Incorporation of a certain metal-ligand pair on the surface of proteins 133 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 offers the advantage of strength and directionality to induce linear protein self-assembly. Bis-His is the most famous surface chelating motif that has the ability of selective metal binding under controlled environmental conditions. Liu et al. designed a C2-symmetric homodimeric enzyme, His(6)-tagged glutathione-S-transferase (GST), as the building block for the coordination-driven supramolecular polymerization (31). In the presence of Ni2+ ions, two His(6)-tag arms on the opposite side of GST provide high-affinity metal binding sites for supramolecular coordination that creates de novo PPIs to induce protein recognition in the formation of 1D protein nanowires (Figure 2a). Because of the microenvironment formed by enzyme arrays, a substrate may be more easily accessible to the active centers, which cause a slight increase in enzyme activities for metal-directed protein assemblies as compared with enzyme monomers. Moreover, the assembly and disassembly behavior of this system can be reversibly regulated by the pH-dependent dimerization of GST, which may help to inspire further design of “smart” protein materials in response to environmental changes. Figure 2. (a) Incorporation of His(6)-tag into GST for Ni2+-induced linear self-assembly; Adapted with permission from reference (31). Copyright 2012 Royal Society of Chemistry. (b) Coassembly of dimeric ENH protein and dsDNA into linear structures. Adapted with permission from reference (32). Copyright 2015 Nature Publishing Group. (c) Linear hemoprotein polymerization induced by heme-hemeprotein interactions. Adapted with permission from reference (36). Copyright 2007 American Chemical Society. 134 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 On the basis of protein-DNA recognition, linear co-assembling nanostructure was obtained by combining the two molecule types into one biomaterial (32). In this case, computational protein design has been used to optimize an energetically favorable interface on Drosophila Engrailed homeodomain (ENH) for C2 symmetrical homodimerization. The designed variant dualENH has a dsDNA binding site on its outward faces to recognize specific DNA sequences, and thus the growth direction of protein-DNA arrays mainly relies on the positioning of protein-binding sites on the dsDNA. By creating two protein-binding sites parallel to the horizontal axis of dsDNA, a nanowire with a width of 15 nm and a length of 300 nm was spontaneously formed by mixing the protein and dsDNA building blocks (Figure 2b). Atomic force microscopy (AFM) images showed the size and shape of highly ordered 1D protein-DNA architecture, the polymerization degree of which was estimated to be approximately 60 units based on the design model. Additionally, engineered DNA structures such as aptamers can be designed for protein recognition to achieve the linear arrangement. For example, Willner’s group synthesized a bisaptamer, a β-aptamer at the 3′ end and an α-aptamer at the 5′ end, for self-assembly into linear DNA-thrombin nanostructures via 1:2 binding of two different domains on thrombin (33). These hybrid biomaterials can integrate the physical/chemical nature of two components, which allows the materials design to develop towards greater functionality. Intermolecular disulfide bond formation offers a new opportunity for self-assembly of more robust 1D protein materials. Cysteine-modified protein variants are easy to connect linearly under oxidizing condition via a disulfide bond due to its desirable properties of high stability at low pH, long bond-distance and restricted dihedral angle compatible with local protein confirmation. Ballister and co-workers demonstrated the success of this design concept to construct a protein nanotube. The ring-shaped Hcp1 protein was engineered with cysteines on the top and bottom faces for covalent stabilization of ring-ring stacking in a head-to-tail fashion via the formation of disulfide linkages (34). Further studies revealed that the disulfide-stabilized Hcp1 nanotubes display a unique combination of two distinctive properties: covalent strength and redox-controlled assembly behavior. This feature is an obvious advantage in developing protein-based nanocapsules for controlled drug delivery when Hcp1 nanotubes were functionalized with polyamidoamine (PAMAM) dendrimer as the plug component via thiol-maleimide reaction. A similar strategy can be applied to another protein system, 11-mer cyclic thermostable TRAP (35). Two mutations V69C and E50C on opposite faces of TRAP ring result in supramolecular polymerization of them into long nanotubes and further self-assembly into micro-sized bundles. The nematic alignment of high-density TRAP nanotubes is similar to the liquid crystalline structure of microtubules. Insights into the self-assembly mechanism of this artificial soft material may be an important step to improve our understanding of complex microtubule dynamics in vitro. Protein receptor-ligand pairs have evolved to bind with very high affinity and selectivity, which provides a powerful tool to modulate protein self-assembly to form nanofibers. The strategy to control the head-to-tail assembly orientation can be realized by using a chemically linked ligand to induce protein association. In principle, the design accuracy and diversity is supposed to be guaranteed 135 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 by reliable interprotein interactions and distinct synthetic ligands. As a proof-of-concept, the strong hemoprotein-heme interactions was first explored by Hayashi’s group for supramolecular polymerization of hemeproteins (Figure 2c) (36). A single H63C mutation was introduced at the cytochrome b562 surface opposite to the heme pocket to react with an iodoacetamide-linked heme for preparing a covalently heme-attached hemoprotein. The fibrous protein arrays were achieved by successive heme-mediated interprotein interactions to obtain an appreciable polymerization degree of approximately 20 units, which exhibited an obvious linker length-dependent increase in size distribution. Following the feasibility of this strategy, the same group extended their investigations to another heme-binding protein, myoglobin (Mb) (37). In this case, it is not necessary to perform any chemical modification on protein surfaces to influence its structural and functional integrity. The synthetic heme dimers or heme-bis(biotin) dyads have proven to be effective in dictating disulfide-bond-linked Mb into polymeric chains without impairing their bioactivity (37, 38). This will lead to the future development of new functional biomaterials with electron-transfer, O2-binding, or enzymatic properties. Chemically Controlled Self-Assembly of Proteins into Linear Structures Besides the natural pairs mentioned above, chemically synthetic pairs have also attracted increasing attention in the construction of fibrous protein assemblies. Some macrocyclic molecules, such as cyclodextrins (CDs) and cucurbit[n]urils (CB[n]), contain central hydrophobic cavities ringed with multiple binding sites for selective inclusion of guest molecules in their interiors, analogous to the specific interactions in biological systems. Using the strategy based on these artificially created pairs, the high stability, reversibility, and responsiveness can be realized in a “head-to-tail” self-assembling protein system, which enables us to control material properties and functions by manipulating host-guest recognition-driven protein assembly. Brunsveld et al. pioneered the use of strong host-guest pairs (e.g. β-CD/ lithocholic acids (LAs), CB/tripeptide phenyalanine-glycine-glycine (FGG), and CB/methylviologen (MV)/naphthalene (Np)) for chemically induced protein dimerization (39–41). All the selected pairs were featured with a high equilibrium association constant (Kd > 1 x 10-6 M) to ensure the specific binding between two protein units. As evidence of functional protein modulation, their recent work demonstrated that enzymatic activity can be switched by selective and reversible dimerization of a split-luciferase heterocomplex using the CB/FGG pair (42). Beyond dimerization, the binding strength of CB/FGG pair allows for the design of 1D protein nanowires by N-terminal fusion of dimeric GST with FGG for a CB-mediated protein assembly (Figure 3a) (43). Furthermore, the fusion of a Ca2+-responsive recoverin into GST can further transform these biomaterials into a protein nanospring to mimic the elastic motions of natural muscle fiber (44). 136 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 Figure 3. (a) CB induced self-assembly of GST-FGG into nanowires based on host-guest interactions; Adapted with permission from ref (43). Copyright 2013 Wiley-VCH Verlag GmbH& Co. KGaA, Weinheim. (b) Positively charged QDs to drive the self-assembly of negatively charged SP1 into nanotubes via electrostatic interactions. Adapted with permission from ref (45). Copyright 2014 American Chemical Society. The chemically synthetic nanoparticles with charged surfaces offer a long-range driving force to stabilize protein interfaces for controlled arrangement of linear protein arrays. This strategy is easy to implement, but needs to carefully consider the steric geometric design and charge distribution on protein surfaces. Typically, the ring-like stable protein one (SP1) covered with a high density of negatively charged residues on the top and bottom faces was selected as the building block for electrostatic self-assembly (45). Only the positively charged quantum dots (QDs) with the size fitted for the central hole of SP1 nanoring is capable of inducing the self-assembly of proteins into highly ordered 1D nanowires via electrostatic interaction and spatial complement (Figure 3b). The design principle is also applicable to other positively charged globular nanoparticles, such as core-cross-linked micelles (CCMs) and fifth-generation poly(amino amine) (PAMAM) (46, 47). As visualized by TEM characterization, the nanoparticles were laid on both sides of each SP1 protein in a sandwich arrangement, which provided optimal scaffolds to incorporate different functional molecules (e.g. chromophores and synergistic GPx-SOD enzyme system) for extending the applications of protein nanowires for biosensors, catalysis, and pharmaceuticals. Furthermore, ethylenediamine (EDA) and carbodiimide (EDC)/N-hydroxyl sulfosuccinimidyl (sulfo-NHS) can serve as an electrostatic inducer and crosslinking agent for selective conjugation of its amino groups with the carboxyl groups of SP1 to promote the strength of protein assemblies (48). 137 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. The “zero-length” crosslink offers a distance constraint for assembly refinement to yield more stable and uniform protein nanowires, which might have greater value for future applications. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 Molecular-Level Control for Multi-Dimensional Protein Arrays Linear protein arrays are relatively simple assemblies, which lack the structural complexity for exploration of advanced biomaterials. Rational design and control of interprotein interactions by surface-attached functional groups at the molecular-level are able to change the order and orientation of the protein self-assembly process. The combination of high pre-existing symmetry of protein building blocks with external control further expands 1D nano/microsturcture to higher dimensional architectures. In some cases, the mechanisms that govern their growth sizes and morphologies can be quantitatively analyzed and then applied to guide the facile creation of periodic 2D and 3D protein arrays. This progress may help address fundamental principles about the structural diversity and shed light on some important biological assembly processes. Protein Self-Assembly To Create Cyclic and Tubular Arrays Cyclic protein assemblies are key intermediates in the modulation of microtubule nucleation phase by providing spatial and temporal control over the initiation of its growth. The molecular details are currently being revealed by studying supramolecular protein polymerization and macrocyclization in vitro. Wagner et al. synthesized a dimeric methotrexate (bisMTX) to induce the self-assembly of peptide-linked dihydrofolate reductase (DHFR) into highly stable ring-like structures through its intrinsic MTX-binding ability (49). Dynamic light-scattering (DLS) and TEM data supported the formation of protein nanorings, whose sizes are tunable between 8 and 20 nm by varying the length and composition of the peptide linker (Figure 4a). Also, the strength of bivalent ligand-mediated PPI is a critical factor to regulate energetics and geometry of protein self-assembly for ring formation. This design principle was validated by Meijer’ group using SEC and quadrupole time-of-flight (Q-TOF) mass analysis when ribonuclease S-protein and S-peptides were chemically connected via a flexible oligo(ethylene glycol) (EG) linker to construct cyclic protein oligomers based on their strong and directional noncovalent interaction (50). In fact, there existed an equilibrium between cyclic and linear products, which was thermodynamically controlled by the ring-chain competition mechanism under a critical concentration. Therefore, the related design criteria can be established to meet the requirements for future fabrication of either cyclic or linear protein chains in a predictable way. Furthermore, the formation of antibody nanorings were highly favored by fusion of an anti-CD3 single-chain variable fragment (scFv) to DHFR, which leads to an opportunity for developing targeted cellular delivery system through a third arm containing bis-MTX functionalized with toxic proteins or oligonucleotides for therapeutic purposes (51). 138 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 Figure 4. Supramolecular protein polymerization and macrocyclization. (a) DHFR nanoring formed by BisMTX-induced self-assembly; Adapted with permission from ref (49). Copyright 2006 American Chemical Society. (b,c) Cooperative self-assembly of GST and SBA into cyclic and tubular structures. Adapted with permission from ref (52) and (54). Copyright 2013 and 2016 American Chemical Society. Another type of cyclic protein polymerization was accomplished by cooperative interactions. One representative work was reported by Liu et al., who utilized the metal-coordination and protein-protein interactions to control over the assembly orientation based on rational steric protein design (52). The V-shaped configuration of two bis-His chelating sites on the engineered GST dimers endeavors to maintain a proper molecular curvature for macrocyclization, which was further reinforced and stabilized by the cooperative effect of multiple noncovalent interactions to determine the final architectures (Figure 4b). The design protocol was demonstrated to be feasible for another protein system, M13 bacteriophage (53). Two genetic modifications at each end of filamentous M13 viruses are able to create a protein building block with anti-streptavidin peptide and His(6)-tag. M13 virus-based ring structures were formed by cooperative protein self-assembly using a heterobifunctional linker that contains tetrameric streptavidin (STV) and Ni(II)-nitrilotriacetic acid complex (NiNTA) for selective 139 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 interconnection of dual-end modified viruses via peptide/STV interactions and metal coordination. Inorganic materials such as ZnS, CdS, or FePt with optical, electronic, or magnetic properties can be crystallized into these virus-based ring scaffolds for designing functional devices when the nucleating peptides were displayed on the major coat protein VIII (pVIII). Very recently, this strategy has also proven to be successful in the construction of large-scale artificial protein microtubules (54). The synthetic hetero-ligand with both sugar and rhodamine B (RhB) moieties, provided synergistic dual driving forces, protein-sugar interactions and π-π stacking, to induce the self-assembly of tetrameric soybean agglutinin (SBA) into periodic helical arrays that wind together to form a 26 nm-wide hollow tubular structure (Figure 4c). Circular dichroism (CD) spectroscopy and cryo-electron microscopy (cryo-EM) analyses indicated a pseudo-1D microtube growth mechanism that the helical protofilament might be a crucial intermediate for microtube formation. Protein Self-Assembly To Create Planar or Branched Network The structural extension can be increased by exploring new self-assembling components and more complex designs. Multi-arm or branched building blocks were designed to introduce symmetry into protein assembly to generate 2D planar or branched network by accurate control of molecular stoichiometries and orientation. As an early example, tetrameric L-rhamnulose-1-phosphate aldolase (RhuA) was modified with two cysteines in each subunit to prepare a four-arm junction (bR) that connects eight biotins on its surface via thiol-disulfide exchange reaction (55). The quadratic network was produced by self-assembly of two biologically unrelated proteins, C4-symmetric bR and D2-symmetric STV, based on the strong binding affinity between biotin and streptavidin. In fact, the design can be simplified to use the symmetry-matching fusion method to create distinct protein “bricks” with multiple connection points through genetic fusion of oligomeric protein subunits with matching rotational symmetries. Noble et al. employed this principle to design planar protein lattices using appropriate point group symmetry combination by manipulating the geometry and flexibility of the linker within fusion proteins (Figure 5a) (56). These well-ordered 2D protein arrays can be further fused with functional protein components to capture nanoparticles and thus provides a stable framework to manufacture biomaterials with diverse properties. On the other hand, the design and synthesis of multivalent ligands with geometric control also makes a significant contribution to the structural diversity of protein assemblies. By adding a novel heme triad as a pivot molecule, supramolecular linear hemoprotein assemblies mentioned above converted fully into 2D protein networks (Figure 5b) (57). One possible mechanism is that a “Y-shaped” trident node was first formed and then grew along its terminal direction via heme-heme pocket interactions to yield highly branched chain structures. This strategy greatly reduces the need for geometrically specific protein building blocks and shows great potential for fabrication of more complicated structures. 140 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 Figure 5. (a) Planar protein lattices created by self-assembly of ALAD-STV/Streptag I fusion proteins with matching rotational symmetry; Adapted with permission from ref (56). Copyright 2011 Nature Publishing Group. (b) The formation of the hemoprotein network assembly via heme-heme pocket interactions using a heme triad as a pivot; Adapted with permission from ref (57). 2012 Royal Society of Chemistry. (c) Linear and S-shaped thrombin patterns on self-assembled DNA tile scaffolds via biotin-STV interactions. Adapted with permission from ref (60). Copyright 2007 American Chemical Society. A more general and powerful approach to produce large 2D protein assemblies is the utilization of naturally evolved or artificially synthetic nano/microstructures as templates to organize proteins. S-layers are highly porous protein meshwork on cell surface. The unique self-assembly feature enables S-layer proteins to form mono- or double layers in solution for high-density display of foreign proteins to produce patterned films. According to this principle, a variety of functional proteins such as streptavidin, antibodies, glycoprotein, fluorescent protein, and enzymes can be fused to S-layer proteins at the N- or C-terminal positions and then self-assembled into geometrically well-defined protein layers. More importantly, the periodic arrangement of functional proteins combined with their biological activities offer a unique way to design high-performance biomaterials such as biosensors, vaccines, biomarkers, biocatalysts, and biodetectors. For 141 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 instance, the formation of S-layer nanostructures difunctionalized with methyl parathion hydrolase (MPH) and anthrax-specific antibodies not only increased the enzymatic stability, but also achieved higher anthrax-detection sensitivity than that of monomeric fusion proteins (58). This finding is regarded as a powerful evidence for coupling the intrinsic biological functions and superstructures of protein assemblies with significantly improved properties in early serological diagnosis of anthrax disease. Figure 6. (a) Self-assembly of WA20-foldon fusion proteins into different polyhedral using “Nanohedra” design rule; Adapted with permission from ref (61). Copyright 2015 American Chemical Society. (b) Self-assembly of KDPGal-FkpA fusion protein into a cube-shaped cage through symmetry manipulation; Adapted with permission from ref (62). Copyright 2014 Nature Publishing Group. (c) Computational design of self-assembled protein cages with tetrahedral or octahedral point group symmetries; Adapted with permission from ref (64). Copyright 2014 Nature Publishing Group. (d) Metal-directed self-assembly of proteins into 1D nanotubes, 2D and 3D planar arrays via thermodynamic control. Adapted with permission from ref (66). Copyright 2012 Nature Publishing Group. 142 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 The self-assembly of more diverse 2D protein arrays have been achieved by using abundant DNA nanostructures as templates due to their precise positional control. Many crossover DNA tiles, such as double-crossover molecules (DX), triple crossover molecules (TX), 2-, 4-, and 6-helix bundles, and cross-shaped tiles with sufficient stability and rigidity, are capable of forming large superstructures according to the number, orientation, and layout of the crossover pairs, which allow their use in the immobilization of target proteins into desired architectures. The LaBean group designed a biotin-labeled 4 × 4 DNA tile with sticky-ended cohesions for self-assembly of uniform 2D nanogrids with periodic square cavities (59). A single flat protein layer was formed by templated self-assembly of streptavidin onto the tethering sites of DNA nanogrids via strong streptavidin-biotin binding. Furthermore, self-assembly of DNA tiles can be reprogrammed by varying the sticky ends to control the final DNA lattice morphologies, leading to more sophisticated protein arrays for some practical applications. Moving forward, using a set of four DX molecules, Yan and co-workers were able to create the desired geometrical protein pattern on rectangular-shaped DNA nanoarrays (60). In this case, two protein-binding aptamer sequences were incorporated into different tiles at specific positions of DNA nanostructures for capturing growth factor and thrombin in the construction of linear and S-shaped protein patterns (Figure 5c). Taking advantage of the sequence specificity and the resulting spatial addressability of DNA nanostructures, this design principle could be further extended to achieve any assembled architectures by precise arrangement of proteins at predetermined locations. 3D Self-Assembly Design for Greater Structural Complexity and Diversity Although 2D protein arrays represent a significant progress toward controllable protein assembly, the development of complex 3D structures to reach higher levels of control and functionality remains a considerable challenge. Not only the symmetry factor, but also the design rules, manipulation accuracy, the stability of structural units, and thermodynamic control should be considered for spatial and hierarchical arrangement in protein assembly processes. Naturally oligomeric proteins always exhibit finite point group symmetries. These unique structural features combined with specific design rules provide a higher-order control for assembly design to develop a series of polyhedral structures. Arai et al. have recently discovered that the application of “nanohedra” design rule to two oligomeric protein domains may result in four distinctive nanostructures by creating a fusion protein via a relatively rigid peptide linker (61). Highly stable symmetric oligomers, dimeric rod-like WA20 protein and trimeric β-propeller-like foldon domain, were selected as the stable framework and 3-arm branched junction, respectively. Meanwhile, two- and three-fold symmetry constraints were imposed in the design of WA20-foldon fusion protein for self-assembly into barrel-like, tetrahedron-like, triangular prism-like, and cube-like structures (Figure 6a), the former two of which have been determined by small-angle X-ray scattering (SAXS) analysis. By careful manipulation of the symmetry axes of two different oligomeric proteins using genetic fusion method, a cubic 143 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 protein cage with the largest size reported so far (an outer diameter of 225 Å and an inner diameter of 132 Å) was generated on the basis of trimeric 2-keto-3-deoxy-6-phosphogalactonate (KDPGal) aldolase and dimeric N-terminal domain of FkpA protein (Figure 6b) (62). Native mass spectrometry (MS) data further revealed that the conformational flexibility of protein building blocks may affect the self-assembly process, leading to two additional self-assembling architectures, 12-subunit tetrahedron and 18-subunit triangular prism. These hollow porous nanostructures can be extensively studied for macromolecule encapsulation, assisted crystallization, and X-ray analysis. Accurate manipulation of protein interfaces allows greater structural complexity for protein assemblies. Computational methods contribute to atomic level understanding of the binding specificity between individual protein subunits, which can assist protein engineering by de novo designing high affinity interfaces and finding suitable binding geometries in the construction of higher-order protein assemblies. The prediction of “hot spot” residues for amino acid sequence design were demonstrated to be most effective in the creation of an energetically favorable interface for driving the self-assembly of proteins into cage-like protein nanostructures. Recently, Baker and co-workers reported two remarkable examples of computational design guided protein assembly to mimic virus capsid structures. RosettaDesign calculations was performed to find low-energy, symmetric PPIs for precise definition of the relative orientations of protein building blocks. In order to construct tetrahedral (T) and octahedral (O) cages, a C3-symmetric protein trimer with 3-fold rotational symmetry axis was chosen as the building block, a set of which tightly associated with each other along the 3-fold axes of the target architectures to achieve the desired point group symmetries (63). Using a similar strategy, heteromeric self-assembly of two distinct protein trimers or a protein trimer and a protein dimer into either T33 or T32 symmetric cage can be realized by alignment of the 3-fold or 2-fold axis of each building block with the corresponding axes of dual tetrahedral architecture (Figure 6c) (64). Looking forward, developing new theoretical tools and algorithms to guide experimental design could greatly expand the diversity of biomimetic nano/microstructures beyond nature’s materials. Other factors that complicate 3D self-assembly design were also evaluated. It is noteworthy that large arrays failed to assemble due to the inherent flexibility of building blocks. The design of more rigid self-assembling components that contain multi-arm junction motifs offers both stable framework and increased valence for assembly of complex, multi-dimensional architectures. Mao et al. synthesized three different types of DNA strands to assemble into symmetric polyhedron for immobilization of proteins onto its surface (65). The long DNA strands (L) coupled with medium strands (M) and short peripheral strands (S) successfully created DNA duplexes with sufficient stability and rigidity in the formation of polyhedral edges and joints. Introduction of a biotin moiety into strand S constituted several trivalent sites for multi-site protein display with strong binding affinity and freedom control. The engineered protein/DNA binary material showed the virus-like structure that may find application in the design of vaccines to elicit immunity against viruses. Furthermore, the dimensionality of protein assemblies was proposed to be affected by thermodynamic/kinetic 144 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. control. In general, 2D and 3D structures were under thermodynamic control while the structures of lower dimension were under kinetic control. Tezcan et al. demonstrated that Zn2+-mediated crystalline protein arrays were irreversibly regulated by external stimuli, in which 1D protein nanotubes as an intermediate was formed by kinetically fast nucleation and can be progressively converted into thermodynamically stable 2D and 3D planar arrays under low metal concentration and pH (Figure 6d) (66). This feature makes supramolecular coordination protein assemblies a versatile platform to engineer functional biomaterials under highly controllable conditions. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 Conclusions and Future Prospects Molecular-level control of protein self-assembly into 1D, 2D, and 3D nano/microstructures represents an important step toward developing protein-based functional materials. Although current research is still far from reproducing the functionality of natural systems, significant progress has been made in the rapidly growing field during the past couple of decades. First, several strategies, including biotechnological strategies, chemical strategies, and their combinations, have been established to dictate protein self-assembly into highly ordered architectures, which provide the engineering foundation for the pursuit of versatile functionalities on an artificially created platform. Secondly, de novo design of new interaction surfaces to drive protein-protein association increases the level of controlled complexity and hierarchy, leading to the growth of structural diversity in fabricating more sophisticated protein architectures. Thirdly, the underlying mechanism that governs the sizes, shapes, and properties of protein assemblies are beginning to be explored. Understanding the key factors involved in the formation of specific self-assembling morphologies may facilitate the future design of novel protein assemblies in a predictable way. The next stage will need to place more emphasis on application development. Custom-designed protein materials is concerned with both the structural design and selective functionalization on periodic protein arrays to achieve precisely addressed functionalities. The recent successes in some self-assembling protein systems indicate a great potential for improving the performance of functional materials, ranging from biocatalysis, biodetection, vaccine design to targeted and controlled drug delivery. The remarkable growth in this field will continue to be fueled by the development of interdisciplinary collaborative efforts between nanoscience, materials science, and chemical biology, which are expected to bring new breakthroughs in engineering bioinspired materials. Acknowledgments We acknowledge the financial support from the National Natural Science Foundation of China (nos. 21234004, 21420102007, 21574056, 91527302, 21474038, and 21004028), the Chang Jiang Scholars Program of China, and the Science Development Program of Jilin Province (nos. 20140101047JC and 20160520005JH). 145 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. References 1. 2. 3. 4. 5. 6. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. Monnard, P. A.; Deamer, D. W. Anat. Rec. 2002, 268, 196–207. Fletcher, D. A.; Mullins, R. D. Nature 2010, 463, 485–492. Shockman, G. D.; Barrett, J. F. Annu. Rev. Microbiol. 1983, 37, 501–527. Petsko, G. A.; Ringe, D. Protein Structure and Function; Lawrence, E., Robertson, M., Eds.; New Sciences Press: London, 2004. Marsh, J. A.; Teichmann, S. A. Annu. Rev. Biochem. 2015, 84, 551–575. Perkins, J. R.; Diboun, I.; Dessailly, B. H.; Lees, J. G.; Orengo, C. Structure 2010, 18, 1233–1243. Pawson, T.; Nash, P. Science 2003, 300, 445–452. Lehn, J. M. Supramolecular Chemistry. Concepts and Perspectives; WileyVCH: Weinheim, 1995. Lehn, J. M. Proc. Natl. Acad. Sci. U. S. A. 2002, 99, 4763–4768. De Greef, T. F.; Smulders, M. M.; Wolffs, M.; Schenning, A. P.; Sijbesma, R. P.; Meijer, E. W. Chem. Rev. 2009, 109, 5687–5754. Schulenburg, C.; Hilvert, D. Top. Curr. Chem. 2013, 337, 41–67. Andersen, N. H. Protein Structure, Stability, and Folding. Methods in Molecular Biology; Murphy, K. P., Ed.; Humana Press: Totowa, NJ, 2001. Bahadur, R. P.; Zacharias, M. Cell Mol. Life Sci. 2008, 65, 1059–1072. Luo, Q.; Hou, C.; Bai, Y.; Wang, R.; Liu, J. Chem. Rev. 2016, 116, 13571–13632. Bai, Y.; Luo, Q.; Liu, J. Chem. Soc. Rev. 2016, 45, 2756–2767. Lai, Y. T.; King, N. P.; Yeates, T. O. Trends Cell Biol. 2012, 22, 653–661. Jones, E. Y.; Yeates, T. O. Curr. Opin. Struct. Biol. 2012, 22, 197–199. Gautieri, A.; Vesentini, S.; Redaelli, A.; Buehler, M. J. Nano Lett. 2011, 11, 757–766. Sleytr, U. B.; Beveridge, T. J. Trends Microbiol. 1999, 7, 253–260. Sun, S.; Rao, V. B.; Rossmann, M. G. Curr. Opin. Struct. Biol. 2010, 2, 114–120. Fraenkel-Conrat, H.; Williams, R. C. Proc. Natl. Acad. Sci. U. S. A. 1955, 41, 690–698. Bancroft, J. B.; Hiebert, E. Virology 1967, 32, 354–356. Poranen, M. M.; Paatero, A. O.; Tuma, R.; Bamford, D. H. Mol. Cell 2001, 7, 845–854. Božič, S.; Doles, T.; Gradišar, H.; Jerala, R. Curr. Opin. Chem. Biol. 2013, 17, 940–945. Witus, L. S.; Francis, M. B. Acc. Chem. Res. 2011, 44, 774–783. Mouw, J. K.; Ou, G.; Weaver, V. M. Nat. Rev. Mol. Cell Biol. 2014, 15, 771–785. Padilla, J. E.; Colovos, C.; Yeates, T. O. Proc. Natl. Acad. Sci. U. S. A. 2001, 98, 2217–2221. De Crescenzo, G.; Pham, P. L.; Durocher, Y.; Chao, H.; O’ConnorMcCourt, M. D. J. Biol. Chem. 2004, 279, 26013–26018. Shekhawat, S. S.; Porter, J. R.; Sriprasad, A.; Ghosh, I. J. Am. Chem. Soc. 2009, 131, 15284–15290. 146 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 30. Dietz, H.; Bornschlögl, T.; Heym, R.; König, F.; Rief, M. New J. Phys. 2007, 9, 424–432. 31. Zhang, W.; Luo, Q.; Miao, L.; Hou, C.; Bai, Y.; Dong, Z.; Xu, J.; Liu, J. Nanoscale 2012, 4, 5847–5851. 32. Mou, Y.; Yu, J. Y.; Wannier, T. M.; Guo, C. L.; Mayo, S. L. Nature 2015, 525, 230–233. 33. Weizmann, Y.; Braunschweig, A. B.; Wilner, O. I.; Cheglakov, Z.; Willner, I. Chem. Commun. 2008, 28, 4888–4890. 34. Ballister, E. R.; Lai, A. H.; Zuckermann, R. N.; Cheng, Y.; Mougous, J. D. Proc. Natl. Acad. Sci. U. S. A. 2008, 105, 3733–3738. 35. Miranda, F.; Iwasaki, K.; Akashi, S.; Sumitomo, K.; Kobayashi, M.; Yamashita, I.; Tame, J.; Heddle, J. Small 2009, 5, 2077–2084. 36. Kitagishi, H.; Oohora, K.; Yamaguchi, H.; Sato, H.; Matsuo, T.; Harada, A.; Hayashi, T. J. Am. Chem. Soc. 2007, 129, 10326–10327. 37. Onoda, A1; Takahashi, A; Oohora, K; Onuma, Y; Hayashi, T. Chem. Biodivers. 2012, 9, 1684–1692. 38. Oohora, K1; Burazerovic, S; Onoda, A; Wilson, YM; Ward, TR; Hayashi, T. Angew. Chem., Int. Ed. Engl. 2012, 51, 3818–3821. 39. Uhlenheuer, D. A.; Wasserberg, D.; Nguyen, H.; Zhang, L.; Blum, C.; Subramaniam, V.; Brunsveld, L. Chem. - Eur. J. 2009, 15, 8779–8790. 40. Nguyen, H. D.; Dang, D. T.; van Dongen, J. L. J.; Brunsveld, L. Angew. Chem., Int. Ed. 2010, 49, 895–898. 41. Uhlenheuer, D. A.; Young, J. F.; Nguyen, H. D.; Scheepstra, M.; Brunsveld, L. Chem. Commun. 2011, 47, 6798–6800. 42. Bosmans, R. P.; Briels, J. M.; Milroy, L. G.; de Greef, T. F.; Merkx, M.; Brunsveld, L. Angew. Chem., Int. Ed. 2016, 55, 8899–8903. 43. Hou, C.; Li, J.; Zhao, L.; Zhang, W.; Luo, Q.; Dong, Z.; Xu, J.; Liu, J. Angew. Chem., Int. Ed. 2013, 52, 5590–5593. 44. Si, C.; Li, J.; Luo, Q.; Hou, C.; Pan, T.; Li, H.; Liu, J. Chem. Commun. 2016, 52, 2924–2927. 45. Miao, L.; Han, J.; Zhang, H.; Zhao, L.; Si, C.; Zhang, X.; Hou, C.; Luo, Q.; Xu, J.; Liu, J. ACS Nano 2014, 8, 3743–3751. 46. Sun, H.; Zhang, X.; Miao, L.; Zhao, L.; Luo, Q.; Xu, J.; Liu, J. ACS Nano 2016, 10, 421–428. 47. Sun, H.; Miao, L.; Li, J.; Fu, S.; An, G.; Si, C.; Dong, Z.; Luo, Q.; Yu, S.; Xu, J.; Liu, J. ACS Nano 2015, 9, 5461–5469. 48. Miao, L.; Fan, Q.; Zhao, L.; Qiao, Q.; Zhang, X.; Hou, C.; Xu, J.; Luo, Q.; Liu, J. Chem. Commun. 2016, 52, 4092–4095. 49. Carlson, J. C.; Jena, S. S.; Flenniken, M.; Chou, T. F.; Siegel, R. A.; Wagner, C. R. J. Am. Chem. Soc. 2006, 128, 7630–7638. 50. Bastings, M. M. C.; de Greef, T. F. A.; van Dongen, J. L. J.; Merkx, M.; Meijer, E. W. Chem. Sci. 2010, 1, 79–88. 51. Gangar, A.; Fegan, A.; Kumarapperuma, S.; Wagner, C. J. Am. Chem. Soc. 2012, 134, 2895–2897. 52. Bai, Y.; Luo, Q.; Zhang, W.; Miao, L.; Xu, J.; Li, H.; Liu, J. J. Am. Chem. Soc. 2013, 135, 10966–10969. 53. Nam, K.; Peelle, B.; Lee, S.-W.; Belcher, A. Nano Lett. 2004, 4, 23–27. 147 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017. Downloaded by UNIV OF FLORIDA on October 26, 2017 | http://pubs.acs.org Publication Date (Web): October 25, 2017 | doi: 10.1021/bk-2017-1252.ch007 54. Yang, G.; Zhang, X.; Kochovski, Z.; Zhang, Y.; Dai, B.; Sakai, F.; Jiang, L.; Lu, Y.; Ballauff, M.; Li, X.; Liu, C.; Chen, G.; Jiang, M. J. Am. Chem. Soc. 2016, 138, 1932–1937. 55. Ringler, P.; Schulz, G. E. Science 2003, 302, 106–109. 56. Sinclair, J. C.; Davies, K. M.; Vénien-Bryan, C.; Noble, M. E. Nat. Nanotechnol. 2011, 6, 558–562. 57. Oohora, K.; Onoda, A.; Hayashi, T. Chem. Commun. 2012, 48, 11714–11726. 58. Wang, X. Y.; Wang, D. B.; Zhang, Z. P.; Bi, L. J.; Zhang, J. B.; Ding, W.; Zhang, X. E. Small 2015, 11, 5826–5832. 59. Yan, H.; Park, S. H.; Finkelstein, G.; Reif, J. H.; LaBean, T. H. Science 2003, 301, 1882–1884. 60. Chhabra, R.; Sharma, J.; Ke, Y.; Liu, Y.; Rinker, S.; Lindsay, S.; Yan, H. J. Am. Chem. Soc. 2007, 129, 10304–10305. 61. Kobayashi, N.; Yanase, K.; Sato, T.; Unzai, S.; Hecht, M. H.; Arai, R. J. Am. Chem. Soc. 2015, 137, 11285–11293. 62. Lai, Y. T.; Reading, E.; Hura, G. L.; Tsai, K. L.; Laganowsky, A.; Asturias, F. J.; Tainer, J. A.; Robinson, C. V.; Yeates, T. O. Nat. Chem. 2014, 6, 1065–1071. 63. King, N. P.; Sheffler, W.; Sawaya, M. R.; Vollmar, B. S.; Sumida, J. P.; André, I.; Gonen, T.; Yeates, T. O.; Baker, D. Science 2012, 336, 1171–1174. 64. King, N. P.; Bale, J. B.; Sheffler, W.; McNamara, D. E.; Gonen, S.; Gonen, T.; Yeates, T. O.; Baker, D. Nature 2014, 510, 103–108. 65. Zhang, C.; Tian, C.; Guo, F.; Liu, Z.; Jiang, W.; Mao, C. Angew. Chem., Int. Ed. 2012, 51, 3382–3385. 66. Brodin, J.; Ambroggio, X.; Tang, C.; Parent, K.; Baker, T.; Tezcan, F. A. Nat. Chem. 2012, 4, 375–382. 148 Ito et al.; Advances in Bioinspired and Biomedical Materials Volume 1 ACS Symposium Series; American Chemical Society: Washington, DC, 2017.