In the best marriages, each partner brings attributes to the union that makes it both stronger and greater than the sum of its parts. In a paper in Nature, Zhang et al.1 have achieved such a feat in the field of materials science. They have wed protein crystals to polymeric gels to create a self-healing hybrid material that can undergo multiple cycles of hyperexpansion and contraction without losing its ability to return to its original crystalline state. Such dynamic assemblies hold great promise for applications such as sensors, separators and actuators (devices that convert energy into movement), but have thus far been difficult to produce.
The ordered component of the wedded couple is ferritin, a protein that is ubiquitous in nature. Ferritin stores a mineralized form of iron called ferrihydrite in its hollow core, and releases iron(ii) ions (Fe2+) when they are needed by an organism. Each ferritin molecule is made up of 24 subunits that self-assemble to form a nearly spherical, cage-like structure.
Zhang and colleagues used a variant of ferritin, which had been engineered as previously reported2, to enable amino-acid side chains on two ferritin molecules to bind to calcium ions, forming bridges between the molecules. Each ferritin molecule connects in this way to 12 of its neighbours, enabling the protein to grow as cubic crystals that have long-range order extending to tens of micrometres.
The authors wanted to combine the crystal with a flexible component, and chose a polymeric hydrogel — a water-absorbing network of crosslinked polymers — as a suitable candidate. To bring the two components together, the authors first soaked the crystals in a solution of small hydrogel precursors; this allowed the precursors to diffuse throughout the water-filled void space of the ferritin lattice. They then polymerized the precursors by transferring the crystals to a solution of polymerization initiators that was highly saline to prevent undue crystal swelling by water infiltration. A continuous and elastic hydrogel network formed in less than two minutes in the void space, which was infiltrated soon after by the salt solution.
When the resulting composites were placed in water, they expanded isotropically (equally in all directions) to about 200% of their original size within minutes. The crystals did not lose their faceted morphology, or release ferritin to the bulk medium, for at least 50 minutes. However, continued expansion eventually led to the loss of detectable edges of the crystals.
More strikingly, when Zhang and colleagues transferred partially expanded crystals to a concentrated solution of either sodium chloride or potassium chloride, and then incubated them in a concentrated solution of calcium chloride, the crystals contracted through dehydration, recovering both their original size and their lattice structure (Fig. 1). This expansion–contraction process could be repeated at least eight times without a noticeable change in ferritin ordering. In fact, the authors observed that X-ray structures of ferritin–hydrogel hybrids that had undergone a single expansion–contraction cycle were of higher resolution than could be achieved using conventionally prepared ferritin crystals that lacked the hydrogel. This suggests that the lattices of such hybrids are more precisely ordered than those of conventionally produced ferritin. Polymer infusion might thus be a useful approach to improve the quality of other protein structures, or to access alternative structural states of proteins.
Several additional observations are worth highlighting. First, Zhang et al. found that not all hydrogels are equally good at supporting the expansion–contraction process, because extensive electrostatic and hydrogen-bonding interactions are required between polymer side chains and ferritin molecules for order to be recovered. Second, the reversibility of the process is not perfect: calcium ions help the ferritin lattice to ‘snap back’ into place, both by screening undesirable electrostatic interactions and by bridging ferritin molecules, but the structure of about half of these bridges changes after an expansion–contraction cycle. Third, the polymer-infused crystals often cracked when subjected to abrupt expansion or contraction. However, these fractures were swiftly repaired because the polymeric network dynamically interacted with, and rearranged around, the ferritin molecules, even though the hydrogel was not engineered to have intrinsic self-healing properties3.
Finally, the authors showed that neither crystals loaded with ferrihydrite, nor crystals made using fluorescently labelled ferritin, were impaired in their ability to undergo isotropic expansion–contraction when infused with hydrogel. They also demonstrated that the expansion and contraction of such crystals was unaffected when a shell of ferritin was grown on them and the assembly was infused with hydrogel. By contrast, when a ferritin shell was grown on a core of fluorescently labelled ferritin crystals whose molecules had been chemically crosslinked to prevent expansion, and hydrogel was incorporated into the core–shell structure, the shell shattered on treatment of the crystals with water, owing to the generation of a lattice mismatch between the core and shell as the shell expanded.
The internal cavity of ferritin has been used as a nanoscale reaction vessel in which to produce mineralized nanoparticles that have optical, catalytic or magnetic activity4, and packing nanoparticles into ordered assemblies leads to the emergence of collective behaviours that have found applications in opto-electronics, medicine and sensing5. Zhang and colleagues’ findings might therefore provide a useful method for fine-tuning such collective phenomena by enabling controllable and reversible structural ordering in 3D nanoparticle arrays. Success will hinge on the identity and properties of the particles produced in the cavities of ferritin molecules, on the separation distances and ordering that the protein structure and crystal lattice impose on these nanoparticles, and on whether the kinetics of polymer expansion and contraction can be accelerated or otherwise precisely controlled.
Beyond ferritin, the use of other natural protein cages4 — and, more excitingly, of synthetic protein cages designed from scratch6,7 — should provide the versatility needed to control nanoparticle separation distances and lattice systems. Furthermore, the development of creative polymer chemistry will aid efforts to modify the kinetics of expansion and contraction.
Nature 557, 38-39 (2018)