Structure of a 16-nm Cage Designed by Using Protein Oligomers

See allHide authors and affiliations

Science  01 Jun 2012:
Vol. 336, Issue 6085, pp. 1129
DOI: 10.1126/science.1219351


Designing protein molecules that will assemble into various kinds of ordered materials represents an important challenge in nanotechnology. We report the crystal structure of a 12-subunit protein cage that self-assembles by design to form a tetrahedral structure roughly 16 nanometers in diameter. The strategy of fusing together oligomeric protein domains can be generalized to produce other kinds of cages or extended materials.

Considerable effort in bionanotechnology is aimed at designing protein molecules that will self-assemble into higher-order structures. Design targets include finite structures, such as three-dimensional cages, and materials that extend by growth in one, two, or three dimensions to form filaments, layers, and crystals. Here, we report the atomic structure of a three-dimensional protein cage with a tetrahedral shape, which self-assembles from 12 identical copies of a large (50 kD) designed protein subunit (Fig. 1).

The design reported here has its basis in principles of symmetry, which have played an important role in various strategies for designing protein assemblies. A number of recent studies have reported the creation of symmetric, well-ordered assemblies composed of a few protein molecules (14). Designing ordered assemblies of larger size and complexity has been a greater challenge.

Several years ago a symmetry-based idea was put forward for designing a wide range of large, complex materials by combining two distinct oligomeric protein domains, each conferring a symmetric mode of association such as dimerization or trimerization, into a single protein molecule or subunit (5) (Fig. 1). The importance of controlling the relative orientation of the two combined components and their respective symmetry elements was emphasized, and one specific strategy for exerting the necessary geometric control was tested; the two oligomerization domains were connected by a continuous, semirigid α helix. The approach was successful in producing filaments from one designed protein and somewhat polymorphic, cagelike structures from another (6). Recent studies have expanded on the idea by introducing strategic variations for how to connect separate oligomerization motifs (7). So far, however, crystal structures that could help validate symmetric fusion strategies have not been reported.

Fig. 1

Atomic structure of a designed protein cage. Geometric models (top) illustrate the design principle of fusing two naturally oligomeric protein domains in a specified geometric arrangement (5). Twelve designed subunits self-assemble to form a tetrahedral cage. (Bottom) The natural protein oligomers (left), and a comparison of a monomer of the designed fusion protein model with its actual crystal structure (middle). The crystal structure of the self-assembling cage (right) confirms the tetrahedral geometry of the designed assembly yet shows deviations from perfect symmetry. The center sphere (50 Å diameter) is shown to emphasize the open nature of the cage.

The protein designed in the present work is a geometrically controlled fusion of two oligomerization domains, a naturally trimeric protein (bromoperoxidase) and a naturally dimeric protein (M1 virus matrix protein), connected by an α-helical linker intended to orient the symmetry axes of the two components so they intersect at an angle half the well-known tetrahedral value of 109.5° (Fig. 1, top). The design is a variation on one of the constructions described earlier (5), which assembled to form structures of heterogeneous size rather than a specific 12-subunit cage as intended (6). Success in obtaining a well-ordered cage with the current design relied critically on making two amino acid changes to the earlier design. After those amino acid changes, homogeneous 12-subunit assemblies were obtained nearly exclusively in solution (6), and an x-ray crystal structure was obtained at 3.0 Å resolution (Fig. 1).

The cage is about 160 Å at its largest diameter. Within each subunit, the separate dimeric and trimeric oligomerization domains are connected by a semirigid linker that retains an α-helical conformation, as designed. However, the helical linkers in the 12 subunits exhibit variable amounts of bending and torsion. This, combined with disturbances in some of the M1 matrix protein dimeric interfaces (8), causes the assembled cage to deviate from perfect tetrahedral geometry by about 8 Å overall (6).

The atomic structure reported here validates the symmetry-based fusion approach for designing ordered protein materials of various types, originally dubbed “nanohedra” (5). The results also emphasize the consequences of flexibility between the linked components and the need for attention to this element during the design stage. Symmetry-based strategies promise to advance the design of new biomaterials, from cages to three-dimensional crystals, potentially rivaling designs already achieved by using nucleic acids instead of protein molecules as the assembly blocks (9).

Supplementary Materials

Materials and Methods

Figs. S1 to S4

Table S1

References (1017)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
  2. Acknowledgments: The authors thank M. Sawaya and N. King for helpful advice. This work was supported by NSF grant MCB-0843065. T.O.Y. holds a patent on self-assembling proteins: US6756039. Structural data are deposited in the Protein Data Bank (codes 3vdx and 4d9j).

Stay Connected to Science

Navigate This Article