How protein materials balance strength, robustness, and adaptability

Abstract

Proteins form the basis of a wide range of biological materials such as hair, skin, bone, spider silk, or cells, which play an important role in providing key functions to biological systems. The focus of this article is to discuss how protein materials are capable of balancing multiple, seemingly incompatible properties such as strength, robustness, and adaptability. To illustrate this, we review bottom-up materiomics studies focused on the mechanical behavior of protein materials at multiple scales, from nano to macro. We focus on alpha-helix based intermediate filament proteins as a model system to explain why the utilization of hierarchical structural features is vital to their ability to combine strength, robustness, and adaptability. Experimental studies demonstrating the activation of angiogenesis, the growth of new blood vessels, are presented as an example of how adaptability of structure in biological tissue is achieved through changes in gene expression that result in an altered material structure. We analyze the concepts in light of the universality and diversity of the structural makeup of protein materials and discuss the findings in the context of potential fundamental evolutionary principles that control their nanoscale structure. We conclude with a discussion of multiscale science in biology and de novo materials design.

Figure 1. Hierarchical structure of three example biological protein materials, intermediate filaments (panel A), collagenous tissues (panel B), and amyloids (panel C).

Figure adapted from Buehler and Yung (2009) and Ackbarow et al. (2009c).

Figure 2. Size effect associated with clusters of H-bonds (Ackbarow et al. 2007; Keten and Buehler 2008a,b).

Panel A: shear strength and robustness of clusters of H-bonds as a function of the size of the strand, showing a peak maximum strength of ≈200 MPa at a critical cluster size of 3–4 H-bonds. By utilizing this size effect, the fundamental limitation of H-bonds, being mechanically weak, can be overcome (Keten and Buehler 2008a; Buehler et al. 2008). The analysis further shows that the robustness increases continuously with the number of H-bonds. Panel B: number of H-bonds in common protein motifs. The comparison with the number of 3–4 H-bonds for optimal mechanical performance suggests that most natural protein motifs fall into this range.

Figure 3. Strength-robustness relation for alpha-helical protein filaments [results adapted from Keten and Buehler (2008a), Ackbarow and Buehler (2009a), and Qin et al. (2009a)].

Panel A shows the geometry of a single alpha-helix, composed of 3–4 H-bonds per turn. We study how the performance in the strength-robustness domain changes if several alpha-helices are assembled in different hierarchical patterns, as shown schematically in the plot (for eight alpha-helices). Panel B shows the results for eight subelements in the protein filament arranged in all possible hierarchical patterns. The definition of subelements and their arrangement are those shown in panel A. To present the results, we use the following nomenclature {b_N,b_N−1,…,b₂,b₁} to uniquely describe the various hierarchical structures. The values of b_i in this expression thereby define the number of elements that are found in parallel with each other at a particular hierarchical level, from the largest to the smallest elements. A single alpha-helix is characterized by {1}, a bundle of two alpha-helices {2} resembles a coiled-coil structure (CC2), and a bundle of four alpha-helices {4} resembles a fourfold coiled-coil structure (CC4; see inset in plot). The {8} structure represents a single bundle of eight alpha-helices in parallel; the {2,4} structure represents a fiber composed of two bundles of four alpha-helices; the {2,2,2} structure represents a fiber composed of two bundles of two bundles of two alpha-helices each; the {4,2} structure represents a fiber composed of four bundles of two alpha-helices. Panel C shows results for 16,384 subelements in the protein filament. An analysis of the distribution of structures and their performance shows that most data points (>98%) in panel d fall onto the banana-curve. Only less than 2% of all structures lead to high strength and high robustness. This analysis shows how high-performance materials can be made out of relatively weak constituents such as alpha-helices that are bonded by mechanically inferior H-bonds (Keten and Buehler 2008a; Qin et al. 2009a).

Figure 4. Hierarchical structure of a simplistic model of the intermediate filament protein network in cells [figure adapted from Ackbarow et al. (2009c)].

Seven levels of hierarchies are considered, from intrabackbone hydrogen bond (H0), alpha-helical turns (H1), and filaments of alpha-helices (H2) to the representative unit cell (H3) of protein networks (H4) that form the cell nucleus (defects in the network highlighted) (H5) of eukaryotic cells (H6). Even though this is a simple model system, it enables us to illustrate the major points associated with the deformation mechanics of hierarchical biological materials throughout multiple scales. The structure at each level is adapted to provide a suitable mechanical response and plays a key role in defining the overall mechanical behavior. Unfolding of alpha-helix turns (H1) proceeds via breaking of strong clusters of 3–4 H-bonds (H0). The large deformation of alpha-helix filaments (with maximum strains of 100–200%) (H2) is enabled by the serial arrangements of many alpha-helical turns (H1). The severe stiffening of the filaments is enabled by alpha-to-beta-sheet transitions and backbone stretching, followed by interprotein sliding at the filament level (H2), is a direct consequence of the structure of coiled alpha-helical proteins (it is noted, that in the simple model system and case study reviewed here interprotein beta-sheet formation and sliding is not considered, as only a single alpha-helical protein filament is modeled; however, recent studies of realistic intermediate filaments showed that the above mentioned mechanisms indeed occur (Qin et al., 2009b). The lattice structure (H3) is the key to facilitate large strain gradients in the protein network, enabling gigantic strain gradients at virtually no energetic cost at the network level (H4). This behavior is crucial for the flaw-tolerant behavior of the nuclear envelope level (H5), which is relevant to provide robust structural support to cells under large deformation (H6).

Figure 5. Deformation field of the protein network [plot adapted from Ackbarow et al. (2009c)].

The transformation of the crack shape can be recognized from the plots. Panel A shows an overview over different deformation stages. Panel B shows the stress field close to the crack tip, illustrating how the transformation from a horizontal crack to a vertical crack reduces the concentration of stresses at the tip of the crack. The cascaded activation of mechanisms at multiple levels is a remarkable behavior ubiquitously found in biological materials that renders them capable to withstand extreme deformation and large loads.

Figure 6. Examining the mechanobiological mechanism of angiogenesis (Yung et al., 2009a).

HUVECs and HASMCs were used as model cell types to study the angiogenic process. Panel A: strain microdevice used to apply cyclic strain to cell cultures (conferred via straining PDMS wells, the culture substrate, as shown in the inlay, (Yung et al., 2009b). Panel B: a model system to investigate angiogenic sprouting, demonstrated using HUVECs seeded onto microcarriers and embedded into fibrin gels, is induced to form tubelike extensions in response to microenvironmental cues. Images show cultures both under static conditions (left, no strain) and under application of cyclic strain (right). Image documentation of sprouts are observed after 5 days of culture using confocal microscopy, where endothelial CD31 membrane receptors are immunostained with FITC (green) and the nucleus with DAPI (blue). The images qualitatively show that application of cyclic strain significantly enhances sprout formation and suggests that mechanical cues alone are capable of triggering the formation of nascent blood vessels. Panel C: temporal secretion profile of angiogenic factors secreted by HUVECs and HASMCs. In response to cyclic strain, Ang-2 and PDGF are both upregulated, where the Ang-2 peak occurs at day 1 and the PDGF peak at day 2 in a temporal fashion relevant to their physiologic role in angiogenesis Panel D: potential mechanism of strain regulated angiogenesis representing a coupled mechanical-biochemical process. The unregulated secretion of Ang-2 results in the formation of HUVEC sprouts. The subsequent secretion of PDGF resulted in a chemotactic recruitment effect on HASMCs, necessary to stabilize nascent blood vessels in order to form the characteristic double layer comprised of HUVECs and HASMCs in small blood vessel tissues. Under a lack of cyclic strain, both the initial Ang-2 secretion and PDGF secretion are reduced, resulting in reduced angiogenic activity. This study showed that autocrine signaling via activation of Ang-2 may be the mechanistic pathway by which ECs transduce mechanical signals to process angiogenic responses. Furthremore, cyclic strain modulates the intercellular communication between EC and SMC by upregulating chemotactic paracrine factors secreted by ECs to recruit SMCs.

Figure 7. Universality and diversity of the structural makeup of biological protein materials, as discussed in Buehler and Yung (2009).

The illustration shows that universal and diverse protocols are distributed heterogeneously across different hierarchical levels in the material. The inlay shows a visualization of the topoisomerase protein, whose biological role is to cut strands of the DNA double helix. This example illustrates how universal motifs define the overall functional properties of this protein while the entire protein structure represents diversity.

Figure 8. Conventional materials science paradigm (panel A) and hierarchical materials science paradigm applied to biological systems, referred to as materiomics (panel B).

The variables Hi refer to hierarchy levels i=0…N, and Ri refer to material property requirements at hierarchy levels i=0…N (see Fig. 4 for an example of a hierarchical structure with labels for different levels). The figure was adapted from Buehler and Yung (2009). A crucial issue in studying the material properties of biological systems is the existence of feedback loops, for example realized through mechanotransduction mechanisms. As illustrated in the example of angiogenesis (see Fig. 6), environmental cues (in our study, mechanical strain) are sensed and result in changes to gene regulation (realized through a cascade of biochemical signals), which in turn change the structural makeup of tissues at multiple levels (in our study, the formation of new blood vessels through sprouting).