Classical Challenges in the Physical Chemistry of Polymer Networks

Nov 23, 2016 - Department of Chemistry, Federal University of Santa Catarina, Florianópolis, Santa Catarina 88040-900, Brazil. ABSTRACT: Polymer ...
0 downloads 0 Views 6MB Size
Download PDF
Article pubs.acs.org/accounts

Classical Challenges in the Physical Chemistry of Polymer Networks and the Design of New Materials Rui Wang,§ Michelle K. Sing,† Reginald K. Avery,‡ Bruno S. Souza,⊥ Minkyu Kim,§,∥ and Bradley D. Olsen*,§ §

Department of Chemical Engineering, †Department of Materials Science and Engineering, and ‡Department of Biological Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139, United States ⊥ Department of Chemistry, Federal University of Santa Catarina, Florianópolis, Santa Catarina 88040-900, Brazil ABSTRACT: Polymer networks are widely used from commodity to biomedical materials. The space-spanning, net-like structure gives polymer networks their advantageous mechanical and dynamic properties, the most essential factor that governs their responses to external electrical, thermal, and chemical stimuli. Despite the ubiquity of applications and a century of active research on these materials, the way that chemistry and processing interact to yield the final structure and the material properties of polymer networks is not fully understood, which leads to a number of classical challenges in the physical chemistry of gels. Fundamentally, it is not yet possible to quantitatively predict the mechanical response of a polymer network based on its chemical design, limiting our ability to understand and characterize the nanostructure of gels and rationally design new materials. In this Account, we summarize our recent theoretical and experimental approaches to study the physical chemistry of polymer networks. First, our understanding of the impact of molecular defects on topology and elasticity of polymer networks is discussed. By systematically incorporating the effects of different orders of loop structure, we develop a kinetic graph theory and real elastic network theory that bridge the chemical design, the network topology, and the mechanical properties of the gel. These theories show good agreement with the recent experimental data without any fitting parameters. Next, associative polymer gel dynamics is discussed, focusing on our evolving understanding of the effect of transient bonds on the mechanical response. Using forced Rayleigh scattering (FRS), we are able to probe diffusivity across a wide range of length and time scales in gels. A superdiffusive region is observed in different associative network systems, which can be captured by a two-state kinetic model. Further, the effects of the architecture and chemistry of polymer chains on gel nanostructure are studied. By incorporating shear-thinning coiled-coil protein motifs into the midblock of a micelle-forming block copolymer, we are able to responsively adjust the gel toughness through controlling the nanostructure. Finally, we review the development of novel application-oriented materials that emerge from our enhanced understanding of gel physical chemistry, including injectable gel hemostats designed to treat internal wounds and engineered nucleoporin-like polypeptide (NLP) hydrogels that act as biologically selective filters. We believe that the fundamental physical chemistry questions articulated in this Account will provide inspiration to fully understand the design of polymer networks, a group of mysterious yet critically important materials.

■

INTRODUCTION Polymer networks are structures made by chemical or physical bonds between large molecules to yield space-spanning macromolecules with a percolating structure.1−3 From rubber tires to medical materials, superabsorbents, and ingredients in many foods, polymer networks and gels are some of the most widely used soft materials today. Beyond existing applications, they are being investigated as organic electronic materials, soft actuators, drug delivery devices, tissue engineering matrices, and smart or responsive materials.4,5 Despite their widespread use, understanding the physical chemistry of polymer gels remains an outstanding challenge. While the chemistries used to form gels can be quite simple, the way that chemistry and material processing interact to yield the final gel structure and material properties is not fully understood © XXXX American Chemical Society

even after a century of active research. In most solid materials, order is characterized using the language of crystallography; however, gels are typically disordered materials, and the defects found in networks are of a topological nature. While a language has been defined to characterize these topological defects, they are not easily measured.3 The nanoscale revolution, partly enabled by large advances in imaging methodologies, has also been slow to yield advances in our understanding of gels. Not only do gels lack regular structure, they typically contain large amounts of solvent that make them difficult to image (although there are notable exceptions with cryo-TEM6). Therefore, structure identification is performed through scattering and Received: September 7, 2016

A

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 1. Universal cyclic topology of polymer networks. (a) Schematic of polymer networks with different orders of loops (the loop order refers to the number of chains contained in the loop) formed via end-linking bifunctional polymers and trifunctional junctions. (b) Linear dependence of 1/n1 (n1 is the primary loop fraction) on the dimensionless variable cb3(M/m)3/2 (c is the polymer concentration, M is the molar mass of the polymer, and b and m are the Kuhn length and molar mass of the monomer, respectively) predicted by the kinetic graph theory in comparison with experimental data. (c) Loop diagram by plotting the secondary loop fraction, n2, and the tertiary loop fraction, n3, versus the primary loop fraction, n1. Adapted with permission from ref 21. Copyright 2016 American Physical Society.

coexist with ideal bridging connections in the same network12 (Figure 1a). These cyclic defects reduce gel stiffness, extensibility, and toughness; however, most of our fundamental knowledge about polymer gels is built upon loop-free tree-like structures.1−3 Quantifying cyclic defects remains an outstanding challenge that is crucial to many long-standing problems: predicting the mechanical response of polymer gels,13 testing the validity of affine and phantom network models,14 and understanding the effects of trapped chain entanglements.15 Quantifying cyclic defects is difficult because they are almost chemically and spectroscopically indistinguishable from ideal bridging connections; it is thus difficult to count the exact number of different orders of loops in the network. Previous work has relied on indirect methods including gel point suppression, rheology, and nuclear magnetic resonance (NMR) to estimate the number of loops.16,17 Recently, our collaborators, Johnson and co-workers, reported network disassembly spectrometry (NDS), the first experimental method for directly counting primary loops.18,19 Combining NDS with rheological measurements, our two laboratories have shown that the modulus of real networks deviates significantly from ideal network models.20 Therefore, new molecular theories that can predict the number of loops and use these loop densities to calculate mechanical properties are needed. We have developed a kinetic graph theory, based on the work of Stepto,12 which captures the dependence of the cyclic topology on the preparation condition of the polymer network as well as inherent relations between different orders of loop structures (the loop order refers to the number of chains contained in the loop as shown in Figure 1a).21 The primary loop fraction predicted by our theory quantitatively agrees with NDS loop measurements without any fitting parameters (Figure 1b). Our theory predicts a linear relation between the primary loop fraction and a dimensionless product of concentration with single chain pervaded volume, similar to the Langmuir adsorption isotherm. Regardless of concentration or polymer strand molar mass, all gels of a given junction functionality fall onto a single universal curve. The slopes of the lines exhibit strong odd−even alternation for different junction functionalities.22 There is one-to-one correspondence between the fractions of higher-order loops and the primary loop fraction

mechanical property measurements or imaging modes that do not preserve the in situ structure of the gel. These challenges have led to a number of classical unsolved problems in the physical chemistry of gels. At the most basic level, we do not understand how to predict the mechanical properties of a polymer network based on its chemical design. There are multiple competing theories (the phantom and affine network theories3) even for simple properties like the linear elastic modulus; however, the validity of each theory has not been effectively tested. For physical networks, variants of transient network theory7−9 have been developed to model simple associative polymer structures that effectively fit experimental data, but quantitative design based on the properties of individual associative bonds remains beyond reach. The added complexity of interpenetrating networks and double networks10 and nanostructured materials such as block copolymer gels or polymer-cross-linked micellar gels11 further motivates the need for substantial fundamental advances to enable rational design of polymer networks. This Account will summarize areas where our knowledge of these fundamental physical chemistry questions has been rapidly evolving, particularly highlighting the work of our own group. First, our understanding of molecular defects in gels will be discussed, including advances in both metrology and theory that can dramatically improve the accuracy for predicting network topology and translate this into property predictions for real, defect-rich networks. Next, associative polymer gel dynamics will be discussed, focusing on our evolving understanding of how the presence of transient bonds leads to changes in mechanical response. We will discuss understanding and characterizing structured gels and finally will review the development of materials that emerges from the enhanced understanding of gel physical chemistry.

■

EFFECTS OF MOLECULAR CONNECTIVITY ON THE TOPOLOGY AND ELASTICITY OF POLYMER GELS The chemistry to make cross-linked polymer networks successively connects polymer strands with multifunctional cross-linkers via chemical bonding or physical association. Topologically distant reactive groups cannot distinguish between intra- and intermolecular reaction; therefore, cyclic defects B

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 2. Elasticity of polymer networks predicted by RENT. (a) Schematic of the conversion of independent loops to a mixture of phantom strands with different effective length. (b) Elastic effectiveness of polymer strands in loops of different order l for trifunctional ( f = 3) and tetrafunctional ( f = 4) networks. (c) Elastic effectiveness as a function of topological distance m from primary and secondary loops. (d) Comparison of experimentally determined shear modulus G′/(kTv0) vs primary loop fractions, n1,3 for trifunctional networks and n1,4 for tetrafunctional networks, to that predicted by affine theory, phantom theory, RENT with only a correction for primary loops, and RENT accounting for both primary and higher-order loops. Adapted from ref 20. Reprinted with permission from AAAS.

Figure 3. (a) Kinetic phase map showing transition from monotonic to nonmonotonic flow curves as a function of increasing shear rate. (b) Representative probability distribution functions for dangling and bridged chains. Overlays show (top) dangling chain flux and (bottom) representative shear. Adapted with permission from ref 27. Copyright 2015 Royal Society of Chemistry.

(Figure 1c). Therefore, higher order loops cannot be controlled independently of primary loops, in stark contrast to the intuition that arbitrary combinations of loops can be synthesized. To connect network topology to gel elasticity, we also developed a real elastic network theory (RENT), which systematically incorporates the impacts of the loop defects into the phantom network theory.20 RENT assumes that different cyclic defects are independent of each other and captures the essential fact that loop junctions are less constrained than ideal nonloop junctions, resulting in longer effective phantom lengths of the loop strands (Figure 2a). RENT predicts that the elastic effectiveness of strands in small loops (primary loops and secondary loops) is significantly lower than ideal strands (Figure 2b), decreasing gel modulus. The loop effect in reducing the elastic effectiveness of strands decreases rapidly with increasing loop order or distance from the loop (Figure 2b,c). Applying RENT to calculate the effect of topological defects on gel modulus shows excellent agreement with the experimental results without any fitting parameters (Figure 2d). The accuracy of RENT is especially striking when compared to affine and phantom network theories, which do not incorporate defects. Combining RENT and the kinetic graph theory, it is now possible to calculate the mechanical response of polymer networks given knowledge of their preparation conditions,

providing a key step toward predictably designing new materials.23

■

DYNAMICS OF ASSOCIATIVE NETWORKS Many real networks also exhibit viscoelastic relaxation due to physically associative bonds capable of breaking and reforming under stress. Based on this observation, Green and Tobolsky laid the foundation for our understanding of transient networks over 50 years ago by developing a theory for the viscoelastic relaxation of rubber.24 Tanaka and Edwards25 first applied chemical kinetics and the Bell’s Law concept of force-activated bond breakage26 to telechelic associative polymer gels, defining the modern concept of transient network theory. Modifications of the classical theory have been developed in the unentangled regime,7−9 enabling qualitative prediction of rheological response in telechelic associative polymers. In our own lab, we have improved transient network theory by developing a theory that tracks the entire chain conformation distribution during deformation, newly including both looped and bridged chains.27 A set of coupled Smoluchowski equations models the time-dependent deformation of loops, bridges, and dangling chains, yielding predictions that are qualitatively consistent with many experimental systems. The theory shows nonmonotonic stress/strain curves, particularly at low values of C

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research the equilibrium constant Keq = kd/ka (where kd is the dissociation rate constant and ka is the association rate constant) and low values of ka (Figure 3a). In addition, the theory predicts that there is still a high degree of chain association even at high deformation due to chain tumbling (Figure 3b). Stress overshoot is predicted in unsteady flows, and multiple stress relaxation times are observed due to dumbbell chain relaxation and the end block disengagement time. While the average loop fraction decreases with increasing chain extension in gels, which undergo monotonic stress increases, the presence of nonmonotonicities at low values of ka and Keq corresponds to increasing loop fractions at intermediate shear rates to better accommodate gel deformation (Figure 3a). To understand linear polymer chains containing multiple stickers spaced along the backbone, Rubinstein and Semenov have formulated a sticky Rouse theory.28,29 The core insight is that in the unentangled regime, stickers slow the Rouse modes of the polymer chain above a critical length/time scale, that is, above a certain segment length the polymer chain “feels” the presence of the stickers. Diffusion provides an important method to probe the dynamic modes predicted by this theory, although there are few studies of diffusivity in associative polymers.30 Using forced Rayleigh scattering (FRS), we are able to probe diffusivity across a wide range of length and time scales (Figure 4a). Using a model protein hydrogel containing four associative coiled-coil groups evenly spaced along a linear, soluble polyelectrolyte backbone, we observed the Fickian behavior anticipated by the sticky Rouse theory at large length scales (Figure 4b).31 However, at smaller length scales still more than an order of magnitude larger than the radius of gyration of polymers, a superdiffusive region was observed (Figure 4c). We also observed this superdiffusive regime in four-arm PEG star polymers with metal-coordinating associative end-groups. The observation in gels with no similarity in polymer chemistry, associative group type, or chain topology suggests a universal underlying principle that is not captured in existing scaling theories. In particular, the end of the superdiffusive regime corresponds to a critical length scale (∼1 μm) much larger than any structural length scale previously postulated in polymer gels. Diffusion in these systems is accurately fit by a kinetic model assuming that molecules in the gel can exist in two states, one that is relatively fast and weakly disassociated and one that is relatively slow and strongly associated. These two states are allowed to interchange according to an effective first order reaction, resulting in a superdiffusive regime centered around a time scale corresponding to the detachment time to go from the slow to fast state. Fitting this model provides a powerful ability to extract dynamic parameters for exchange between molecular states in the gel itself (as opposed to dilute solution), and in some cases, these can be used to achieve time−temperature− concentration superposition (Figure 4b).31 However, these parameters are empirical, and work to understand their molecular origins is underway. We have also explored the dynamics of associative protein polymers when they become entangled, entering the sticky reptation32 regime above the entanglement molar mass at a given concentration.33 This regime is particularly interesting for the physical chemistry of polymers, as the transition from unentangled to entangled is well-known to yield important changes in the mechanical properties of polymers in glasses and crystals. In associative polymer gels, we have shown that entanglements have an equally profound effect.33 Chainextended associative protein gels exhibit engineering strains of

Figure 4. (a) Schematic of labeled (highlighted in red) and unlabeled P4 proteins used for tracer diffusion measurements in FRS. (b) Anomalous diffusion data in reduced parameter space for varying concentrations and temperatures where the red dashed line indicates the Fickian regime and the blue line indicates the limiting case, Keq = 0. (c) Prediction from twostate model that shows the two Fickian regimes (red dashed lines) and superdiffusive regime (green dashed line). Adapted with permission from ref 31. Copyright 2015 American Chemical Society.

3000% prior to fracture, demonstrating that chain entanglement leads to high extensibility (Figure 5). The Bao group at Stanford has recently confirmed this observation in a chemically distinct system.34

■

MOLECULAR ARCHITECTURE OF NANOSTRUCTURED GELS A full understanding of the microscopic origins of gel behavior cannot be obtained without also understanding gel structure. Engineering the molecular architecture of gels requires a thorough understanding not only of the factors that influence nanostructure formation but also of the effects of microstructural changes on gel behavior. Although imaging of gels remains D

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

responsively turn toughness off and on in a material through nanostructural transitions. Inspired by the design of Shull et al., which uses ionic cross-links in the midblock of a triblock copolymer gel,11 we incorporated associating coiled-coil protein motifs developed in the Tirrell lab39,40 into the midblock of a micelle-forming temperature responsive block copolymer.41 The presence of the thermoresponsive end-blocks resulted in an environmentally responsive material that shear-thins below the end block aggregation temperature and toughens above it via the formation of micellar aggregates. The presence of the micelles toughens the network beyond increasing the number of cross-links within the gel by forming a network of close-packed micelles that provides an additional mechanism for elasticity (Figure 6a,b).41 Increasing either the end block molar mass or the number of midblock associations increases the amount of responsive reinforcement.42 By controlling the design of these gels, it is possible to reach elastic moduli of up to 130 kPa (effectively reinforcing the network by a factor of 14 relative to their low temperature moduli) and increase the stress relaxation time of the network by a factor of up to 50 (Figure 6c). These property changes were achieved using only physical associations and can be reversibly activated and deactivated with changes in temperature. By slowing the kinetics of the associating end block by adjusting the hydrophobicity of a single amino acid repeat unit (switch glycine with alanine), it is possible to further increase the value of the reinforced moduli to approximately 260 kPa.43 Additionally, it supports work by Craig

Figure 5. Stress−strain curve for associative protein hydrogel with entanglements. At low deformation, the gel undergoes plastic draw, after which point it shows strong strain stiffening behavior and an ultimate extension of over 3000% engineering strain. Adapted with permission from ref 33. Copyright 2014 American Chemical Society.

difficult for most systems, a combination of X-ray and neutron scattering, rheology, and controlled synthesis to produce specific molecular comparisons provides insight into the effects of chain architecture and chemistry on gel nanostructure and macroscopic behavior. Gel toughness has typically been achieved by either using multiple discrete gel-forming materials to form multinetwork systems10,35,36 or via addition of additives to make nanocomposites.37,38 However, previous examples of these gels are not responsive. Many applications require a material that is tough and also malleable enough for flow or injection, two oftencompeting properties. Therefore, a key challenge is to

Figure 6. Responsive toughening via the formation of micellar aggregates (a) for proteins containing end-blocks shown in panel b. ELPs responsively toughen with increasing temperature as shown in panel c, which illustrates the effects of varying end block kinetics and linear architecture. Low temperature moduli were determined at 0 °C, while high temperature moduli were determined from the maximum value of G′ following temperature increase above the aggregation temperature. Panel a reprinted with permission from ref 42. Copyright 2013 Royal Society of Chemistry. Panel c reprinted with permission from ref 43. Copyright 2015 American Chemical Society. E

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 7. (a) Schematic of oxidative chain extension, (b) 3-D real-space distribution of protein-rich phase showing porous nanostructure of kinetically arrested ELP, and (c) comparison of oxidatively chain extended ELP to ELP without cysteine modification. Adapted with permission from ref 46. Copyright 2016 American Chemical Society.

Figure 8. Shear-thinning hydrogel development for hemostats. (a) Composition of the nanocomposite and image highlighting the nanocomposite injectability and stability. (b) Clotting is accelerated as more nanoplatelet is included in the nanocomposite. (c) When compared to commercial hemostats, the nanocomposite performs comparable to thrombin-containing hemostats. (d) In vivo rodent liver bleeding models showed improved survival when compared to no intervention. Adapted with permission from ref 53. Copyright 2014 American Chemical Society.

et al.44 and Creton et al.45 in demonstrating that strong, tough materials are often accompanied by slow chain dynamics. The impeded dynamics and hydrogen bonding that contribute to the increased stiffness of these alanine-substituted fusions also

lead to a property that has never been reported in an elastin-like polypeptide (ELP) before: the ability to form a hydrogel in an ELP homopolymer above the thermoresponsive transition temperature where ELPs normally precipitate.47 At sufficiently F

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research

Figure 9. Artificially engineered NLP hydrogels that can act as biologically selective filters. (a) Minimal consensus sequences extracted from natural nucleoporin Nsp1 are genetically polymerized to match the size of Nsp1. (b) P-NLP-P design for physical hydrogel formation. P is the monomer of pentameric coiled-coil protein.40 (c) A schematic of capillary assay filled with P-NLP-P hydrogel and the buffer containing the following molecules to test the gel selectivity: the target green molecule, which can bind to the nuclear transport receptor (blue circles), and the red molecule (control), which cannot. (d,e) A single amino acid change in the NLP sequence can impact selective permeability. Adapted with permission from ref 58. Copyright 2015 John Wiley & Sons.

generate a physically associating gelatin-based hydrogel (Figure 8a).53 The silicate nanoplatelet contributes to the gel both as a physical cross-linker that binds electrostatically to gelatin, forming a shear-thinning and self-healing hydrogel, and as the hemostatic component. When the nanocomposite was tested in contact with blood in vitro and in vivo, decreases in clotting time and blood loss were evident, comparable to commercial systems (Figure 8b−d). Rapid clotting at the nanocomposite interface reflects both the stability of the nanocomposite in contact with blood components and the importance of the clay nanoplatelets in accelerating clotting relative to gelatin or whole blood alone (control) (Figure 8b).53 The nanocomposite clots blood faster than many solid hemostats and is comparable to thrombin (liquid formulation; Figure 8c). The physically cross-linked hydrogel system is more suitable for internal bleeding than solid hemostats given the hydrogel’s injectability and space filling, and it is preferred over liquid-based systems because the gel can localize to the injury site. Treatment of liver bleeding wounds in vivo reflected the in vitro results, showing significant improvements in survival (Figure 8d) and blood loss in nanocomposite treated animals compared to no-intervention controls. No negative side-effects were noted after treatment and monitoring during the 28 day postop period, suggesting biocompatibility and sufficient wound healing in the presence of the nanocomposite. Future investigations into traumatic and surgical applications could leverage the simple shear-thinning delivery to treat internal sites of vascular injury with higher accuracy and simpler delivery. A second example of physical chemistry yielding emergent new properties is in bioinspired engineering of hydrogels that mimic natural systems. One fascinating system is the nuclear pore complex in the nuclear membrane of eukaryotic cells, which acts as a selective filter, allowing the passage of less than 0.1% of all proteins in the body. Selectivity is achieved with a protein matrix formed by a family of ∼30 proteins known as nucleoporins.54 To adapt this function into synthetic hydrogels, we were inspired by previous efforts on minimal consensus repeat design for proteins such as spider silk,55 elastin,56 and resilin, 57 where minimal sequences have replicated the

high concentrations (ca. 15% w/w), a porous, kinetically trapped nanostructured network (Figure 7b) forms, making these gels surprisingly stiff.48 The addition of pendant cysteines to the N and C termini of these ELPs enables oxidative chain extension (Figure 7a) to change them from a stiff yet brittle material to a stiff and extensible material (Figure 7c).46 This new, chainextended material remains porous but resembles a fractal percolating network.

■

BIOMATERIAL DESIGN USING POLYMER PHYSICAL CHEMISTRY By application of this fundamental knowledge of hydrogel topology, dynamics, and architecture, it is possible to rationally develop functional hydrogel systems with previously unachievable physical properties. Greater focus on physical chemistry when studying and designing hydrogels also results in a better understanding of the emergent mechanics and biological activity that arise in hydrogel systems. Therefore, biofunctional, bioinspired, and biomimetic hydrogels are ideal research areas to assess the utility of physical chemistry-based biomaterials development.49 Recent reinvestigations of traditional biomaterials for the impact of physical chemistry on their biological activity highlight the importance of physical chemistry on hydrogel performance.50,51 As examples of this approach, we will discuss new physically associating hydrogels as injectable hemostats and media for biological separations. Injectable gel hemostats were designed to treat incompressible and internal wounds by exploiting the physical chemistry of physically associated, shear-thinning hydrogels. For incompressible and internal wounds, standard methods used to control bleeding (gauze, external pressure) are insufficient to manage blood loss. These wounds require hemostatic materials to be rapidly and specifically delivered to the wound site and remain there, resisting mobilization due to circulatory pressure. While clay-based materials are known hemostats (e.g., zeolite or kaolinite in QuikClot bandages),52 they have been restricted to topical wounds due to their granular form. Together with the Khademhosseini group at Harvard, we have developed a shearthinning nanocomposite hydrogel, using clay nanoplatelets to G

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research Present Address

mechanical response of the full protein. In this vein, we designed artificially engineered nucleoporin-like polypeptides (NLPs) by repeating a simple consensus sequence of 19 amino acids extracted from nucleoporin Nsp1 (Figure 9a).58 Since the NLP consensus sequences alone do not form gels, pentameric coiledcoil40 domains, P, were genetically fused on either side of the NLP. The resultant P-NLP-P structure forms hydrogels in physiological buffers similar to natural nucleoporins (Figure 9b). These NLPs expressed at 20- to 70-fold higher yield than nucleoporins alone and reduced gelation times from hours to minutes. The NLP biosynthetic mimics are the first consensus repeat that goes beyond mechanics, displaying transport rates and protein selectivity similar to the native protein. A great advantage of the consensus repeat design is the ability to tune the transport properties by changing individual amino acids, leading to the development of sequence−structure−property relationships. For example, replacing Asp in 1NLP with Ser in 2NLP within the consensus sequence enhanced the selective accumulation of target biomolecules into the engineered hydrogels (Figure 9c,d,e). Additionally, the modularity in sequence design offers an optimum platform from which to systematically assess detailed mechanisms of the natural nuclear pore complex, still under debate.59,60 By incorporating other biological motifs, such as specific binding recognition sites, into the hydrogel, we envision the development of novel technologies for filtering and separation applications in defense, drug delivery, and food toxicology.

∥

M.K.: Department of Materials Science and Engineering, Department of Biomedical Engineering, The University of Arizona, 1235 James E. Rogers Way, Tucson, Arizona 85721.

Author Contributions

The manuscript was written through contributions of all authors. Funding

R.W. was supported by NSF Award DMR-1253306. M.K.S. and R.K.A. were supported by ARO Contract W911NF-07-D-0004. R.K.A. acknowledges fellowship support from the National Institutes of Health (NIH) Interdepartmental Biotechnology Training Program (NIH/NIGMS 5T32GM008334). B.S.S. thanks Program Science without Borders, No. 234283/2014-9 (CNPq scholarship). M.K. and B.D.O. were supported by DTRA Contract HDTRA1-13-1-0038. Notes

The authors declare no competing financial interest. Biographies Rui Wang received his B.S. (2005) and M.S. (2008) in Chemical Engineering from Zhejiang University, China. He received his Ph.D. in Chemical Engineering from Caltech in 2014. He is currently a postdoctoral researcher at MIT. Michelle Sing received her B.S.E. from Case Western Reserve University’s department of Polymer Science and Engineering in 2011. She is currently a graduate student at MIT.

■

Reginald K. Avery received his B.S. from University of Maryland, College Park, in the department of Bioengineering in 2012. He is currently a graduate student at MIT.

CONCLUSIONS AND OUTLOOK Although polymer gels and networks are old materials, their design and engineering continues to be an explosive area of growth due to a large variety of compelling applications. However, in a surprising number of areas, the design of these materials is largely empirical. Due to obstacles in the characterization of molecular networks, many fundamental challenges have gone unaddressed in the physical chemistry of these materials, leading to a huge scientific opportunity for the development and rational design of these materials. As a result of the great degree of activity in the polymer networks and gels community, new advances in both theory and experiment are being made that are solving the classical scientific challenges and increasingly enabling rational design of gels. This Account highlighted the progress from our own lab in the context of the broader field in the key areas of the impact of network topology on properties, the dynamics of transient networks, and controlling nanostructure, and consequently structure−property relationships in polymer gels. The increased understanding provided by these advances can lead to the development of novel application-oriented materials. Despite current successes, much work remains in this venerable and challenging field to fully understand the design of polymer networks, and it demands continued effort to tease out the secrets of these mysterious yet critically important materials.

■

Bruno S. Souza received his B.S. (2006), M.S. (2008), and Ph.D. (2012) in Chemistry at the Federal University of Santa Catarina, Brazil, where he is currently an Assistant Professor in the Department of Chemistry. Minkyu Kim received his B.S. (2004) in Mechanical Engineering from Kyung Hee University and his M.S. (2006) in Biomedical Engineering and Ph.D. (2011) in Mechanical Engineering and Materials Science from Duke University. He was a postdoctoral researcher at MIT from 2012 to 2016. He is currently an Assistant Professor at the University of Arizona in Materials Science and Engineering and Biomedical Engineering. Bradley Olsen received his S.B. in Chemical Engineering from MIT in 2003 and a Ph.D. in Chemical Engineering from UC Berkeley in 2007 where he was a Hertz Fellow. He was a postdoctoral researcher at Caltech from 2008 to 2009. He is currently an Associate Professor at MIT in Chemical Engineering.

■

REFERENCES

(1) Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca, NY, 1953. (2) de Gennes, P. G. Scaling Concepts in Polymer Physics, 1st ed.; Cornell University Press: Ithaca, NY, 1979. (3) Rubinstein, M.; Colby, R. H., Polymer Physics; Oxford University Press: New York, 2003. (4) Tanaka, Y.; Gong, J. P.; Osada, Y. Novel hydrogels with excellent mechanical performance. Prog. Polym. Sci. 2005, 30, 1−9. (5) Langer, R.; Tirrell, D. A. Designing materials for biology and medicine. Nature 2004, 428, 487−492. (6) Pochan, D. J.; Pakstis, L.; Ozbas, B.; Nowak, A. P.; Deming, T. J. SANS and Cryo-TEM study of self-assembled diblock copolypeptide hydrogels with rich nano- through microscale morphology. Macromolecules 2002, 35, 5358−5360.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Bradley D. Olsen: 0000-0002-7272-7140 H

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research (7) Annable, T.; Buscall, R.; Ettelaie, R.; Whittlestone, D. The Rheology of Solutions of Associating Polymers - Comparison of Experimental Behavior with Transient Network Theory. J. Rheol. 1993, 37, 695−726. (8) Tripathi, A.; Tam, K. C.; McKinley, G. H. Rheology and dynamics of associative polymers in shear and extension: Theory and experiments. Macromolecules 2006, 39, 1981−1999. (9) Vandenbrule, B. H. A. A.; Hoogerbrugge, P. J. Brownian Dynamics Simulation of Reversible Polymeric Networks. J. Non-Newtonian Fluid Mech. 1995, 60, 303−334. (10) Gong, J. P.; Katsuyama, Y.; Kurokawa, T.; Osada, Y. Doublenetwork hydrogels with extremely high mechanical strength. Adv. Mater. 2003, 15, 1155−1158. (11) Henderson, K. J.; Zhou, T. C.; Otim, K. J.; Shull, K. R. Ionically Cross-Linked Triblock Copolymer Hydrogels with High Strength. Macromolecules 2010, 43, 6193−6201. (12) Stanford, J. L.; Stepto, R. F. T.; Waywell, D. R. Rate Theory of Irreversible Linear Random Polymerization 0.2. Application to Intramolecular Reaction in a-a+B-B Type Polymerizations. J. Chem. Soc., Faraday Trans. 1 1975, 71, 1308−1326. (13) Osada, Y.; Gong, J. P. Soft and wet materials: Polymer gels. Adv. Mater. 1998, 10, 827−837. (14) Akagi, Y.; Gong, J. P.; Chung, U.; Sakai, T. Transition between Phantom and Affine Network Model Observed in Polymer Gels with Controlled Network Structure. Macromolecules 2013, 46, 1035−1040. (15) Duering, E. R.; Kremer, K.; Grest, G. S. Structure and Relaxation of End-Linked Polymer Networks. J. Chem. Phys. 1994, 101, 8169− 8192. (16) Stepto, R. F. T. Theoretical and Experimental Studies of Network Formation and Properties. Polymer 1979, 20, 1324−1326. (17) Saalwachter, K.; Gottlieb, M.; Liu, R. G.; Oppermann, W. Gelation as studied by proton multiple-quantum NMR. Macromolecules 2007, 40, 1555−1561. (18) Zhou, H. X.; Woo, J.; Cok, A. M.; Wang, M. Z.; Olsen, B. D.; Johnson, J. A. Counting primary loops in polymer gels. Proc. Natl. Acad. Sci. U. S. A. 2012, 109, 19119−19124. (19) Zhou, H. X.; Schon, E. M.; Wang, M. Z.; Glassman, M. J.; Liu, J.; Zhong, M. J.; Diaz Diaz, D.; Olsen, B. D.; Johnson, J. A. Crossover Experiments Applied to Network Formation Reactions: Improved Strategies for Counting Elastically Inactive Molecular Defects in PEG Gels and Hyperbranched Polymers. J. Am. Chem. Soc. 2014, 136, 9464− 9470. (20) Zhong, M. J.; Wang, R.; Kawamoto, K.; Olsen, B. D.; Johnson, J. A. Science 2016, 353, 1264−1268. (21) Wang, R.; Alexander-Katz, A.; Johnson, J. A.; Olsen, B. D. Universal Cyclic Topology in Polymer Networks. Phys. Rev. Lett. 2016, 116, 188302. (22) Wang, R.; Johnson, J. A.; Olsen, B. D. Manuscript in preparation. (23) Balazs, A. C. Polymer Chemistry: Wasted loops quantified. Nature 2013, 493, 172−173. (24) Green, M. S.; Tobolsky, A. V. A New Approach to the Theory of Relaxing Polymeric Media. J. Chem. Phys. 1946, 14, 80−92. (25) Tanaka, F.; Edwards, S. F. Viscoelastic Properties of Physically Cross-Linked Networks - Transient Network Theory. Macromolecules 1992, 25, 1516−1523. (26) Bell, G. I. Models for Specific Adhesion of Cells to Cells. Science 1978, 200, 618−627. (27) Sing, M. K.; Wang, Z. G.; McKinley, G. H.; Olsen, B. D. Celebrating Soft Matter’s 10th Anniversary: Chain configuration and rate-dependent mechanical properties in transient networks. Soft Matter 2015, 11, 2085−2096. (28) Rubinstein, M.; Semenov, A. N. Thermoreversible gelation in solutions of associating polymers. 2. Linear dynamics. Macromolecules 1998, 31, 1386−1397. (29) Semenov, A. N.; Rubinstein, M. Thermoreversible gelation in solutions of associative polymers. 1. Statics. Macromolecules 1998, 31, 1373−1385.

(30) Seiffert, S.; Oppermann, W. Diffusion of linear macromolecules and spherical particles in semidilute polymer solutions and polymer networks. Polymer 2008, 49, 4115−4126. (31) Tang, S. C.; Wang, M. Z.; Olsen, B. D. Anomalous Self-Diffusion and Sticky Rouse Dynamics in Associative Protein Hydrogels. J. Am. Chem. Soc. 2015, 137, 3946−3957. (32) Rubinstein, M.; Semenov, A. N. Dynamics of entangled solutions of associating polymers. Macromolecules 2001, 34, 1058−1068. (33) Tang, S. C.; Glassman, M. J.; Li, S. L.; Socrate, S.; Olsen, B. D. Oxidatively Responsive Chain Extension to Entangle Engineered Protein Hydrogels. Macromolecules 2014, 47, 791−799. (34) Li, C. H.; Wang, C.; Keplinger, C.; Zuo, J. L.; Jin, L.; Sun, Y.; Zheng, P.; Cao, Y.; Lissel, F.; Linder, C.; You, X. Z.; Bao, Z. A. A highly stretchable autonomous self-healing elastomer. Nat. Chem. 2016, 8, 618−625. (35) Sun, J. Y.; Zhao, X. H.; Illeperuma, W. R. K.; Chaudhuri, O.; Oh, K. H.; Mooney, D. J.; Vlassak, J. J.; Suo, Z. G. Highly stretchable and tough hydrogels. Nature 2012, 489, 133−136. (36) Ducrot, E.; Montes, H.; Creton, C. Structure of Tough Multiple Network Elastomers by Small Angle Neutron Scattering. Macromolecules 2015, 48, 7945−7952. (37) Gaharwar, A. K.; Dammu, S. A.; Canter, J. M.; Wu, C. J.; Schmidt, G. Highly Extensible, Tough, and Elastomeric Nanocomposite Hydrogels from Poly(ethylene glycol) and Hydroxyapatite Nanoparticles. Biomacromolecules 2011, 12, 1641−1650. (38) Liu, R. Q.; Liang, S. M.; Tang, X. Z.; Yan, D.; Li, X. F.; Yu, Z. Z. Tough and highly stretchable graphene oxide/polyacrylamide nanocomposite hydrogels. J. Mater. Chem. 2012, 22, 14160−14167. (39) Olsen, B. D.; Kornfield, J. A.; Tirrell, D. A. Yielding Behavior in Injectable Hydrogels from Telechelic Proteins. Macromolecules 2010, 43, 9094−9099. (40) Petka, W. A.; Harden, J. L.; McGrath, K. P.; Wirtz, D.; Tirrell, D. A. Reversible hydrogels from self-assembling artificial proteins. Science 1998, 281, 389−392. (41) Glassman, M. J.; Chan, J.; Olsen, B. D. Reinforcement of Shear Thinning Protein Hydrogels by Responsive Block Copolymer SelfAssembly. Adv. Funct. Mater. 2013, 23, 1182−1193. (42) Glassman, M. J.; Olsen, B. D. Structure and mechanical response of protein hydrogels reinforced by block copolymer self-assembly. Soft Matter 2013, 9, 6814−6823. (43) Glassman, M. J.; Olsen, B. D. End Block Design Modulates the Assembly and Mechanics of Thermoresponsive, Dual-Associative Protein Hydrogels. Macromolecules 2015, 48, 1832−1842. (44) Yount, W. C.; Loveless, D. M.; Craig, S. L. Strong means slow: Dynamic contributions to the bulk mechanical properties of supramolecular networks. Angew. Chem., Int. Ed. 2005, 44, 2746−2748. (45) Ducrot, E.; Chen, Y. L.; Bulters, M.; Sijbesma, R. P.; Creton, C. Toughening Elastomers with Sacrificial Bonds and Watching Them Break. Science 2014, 344, 186−189. (46) Glassman, M. J.; Avery, R. K.; Khademhosseini, A.; Olsen, B. D. Toughening of Thermoresponsive Arrested Networks of Elastin-Like Polypeptides To Engineer Cytocompatible Tissue Scaffolds. Biomacromolecules 2016, 17, 415−426. (47) Hassouneh, W.; Christensen, T.; Chilkoti, A. Elastin-like Polypeptides as a Purification Tag for Recombinant Proteins. Curr. Protoc. Protein Sci. 2010, 6.11.1−16. (48) Glassman, M. J.; Olsen, B. D. Arrested Phase Separation of Elastin-like Polypeptide Solutions Yields Stiff, Thermoresponsive Gels. Biomacromolecules 2015, 16, 3762−3773. (49) Hoffman, A. S. Hydrogels for biomedical applications. Adv. Drug Delivery Rev. 2002, 54, 3−12. (50) Brandl, F.; Sommer, F.; Goepferich, A. Rational design of hydrogels for tissue engineering: Impact of physical factors on cell behavior. Biomaterials 2007, 28, 134−146. (51) Chaudhuri, O.; Gu, L.; Klumpers, D.; Darnell, M.; Bencherif, S. A.; Weaver, J. C.; Huebsch, N.; Lee, H. P.; Lippens, E.; Duda, G. N.; Mooney, D. J. Hydrogels with tunable stress relaxation regulate stem cell fate and activity. Nat. Mater. 2016, 15, 326−334. I

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX

Article

Accounts of Chemical Research (52) Spotnitz, W. D.; Burks, S. Hemostats, sealants, and adhesives: components of the surgical toolbox. Transfusion 2008, 48, 1502−1516. (53) Gaharwar, A. K.; Avery, R. K.; Assmann, A.; Paul, A.; McKinley, G. H.; Khademhosseini, A.; Olsen, B. D. Shear-Thinning Nanocomposite Hydrogels for the Treatment of Hemorrhage. ACS Nano 2014, 8, 9833− 9842. (54) Grossman, E.; Medalia, O.; Zwerger, M. Functional architecture of the nuclear pore complex. Annu. Rev. Biophys. 2012, 41, 557−584. (55) Xu, L. L.; Rainey, J. K.; Meng, Q.; Liu, X. Q. Recombinant Minimalist Spider Wrapping Silk Proteins Capable of Native-Like Fiber Formation. PLoS One 2012, 7, e50227. (56) Zhang, Y. N.; Avery, R. K.; Vallmajo-Martin, Q.; Assmann, A.; Vegh, A.; Memic, A.; Olsen, B. D.; Annabi, N.; Khademhosseini, A. A Highly Elastic and Rapidly Crosslinkable Elastin-Like PolypeptideBased Hydrogel for Biomedical Applications. Adv. Funct. Mater. 2015, 25, 4814−4826. (57) Li, L.; Kiick, K. L. Resilin-Based Materials for Biomedical Applications. ACS Macro Lett. 2013, 2, 635−640. (58) Kim, M.; Chen, W. G.; Kang, J. W.; Glassman, M. J.; Ribbeck, K.; Olsen, B. D. Artificially Engineered Protein Hydrogels Adapted from the Nucleoporin Nsp1 for Selective Biomolecular Transport. Adv. Mater. 2015, 27, 4207−4212. (59) Moussavi-Baygi, R.; Mofrad, M. R. Rapid Brownian Motion Primes Ultrafast Reconstruction of Intrinsically Disordered Phe-Gly Repeats Inside the Nuclear Pore Complex. Sci. Rep. 2016, 6, 29991. (60) Kabachinski, G.; Schwartz, T. U. The nuclear pore complex structure and function at a glance. J. Cell Sci. 2015, 128, 423−429.

J

DOI: 10.1021/acs.accounts.6b00454 Acc. Chem. Res. XXXX, XXX, XXX−XXX