Two groups reported new advances this week in one area of bio-nanotechnology, namely the quest to design protein units that can self-assemble into more complex structures.

One paper described the crystal structure of a self-assembling protein that was designed more than a decade ago, while the other described a general computational method for designing, with atomic-level precision, proteins that can self-assemble into symmetrical multipart structures. All three complexes were less than 15 nanometers in diameter.

Many natural proteins self-assemble, and those self-assembling proteins are what gets much of the work of the cell done. "Proteins have complicated functions, and the ability [to fulfill those functions] comes from their complex structure," Neil King told BioWorld Today. King is a researcher at the University of Washington and the first author on one of the papers, both of which were published in the June 1, 2012, issue of Science.

When large proteins self-assemble, their components, or subunits, are usually held together by many weak bonds. Predicting how two subunits will interact, however, has proven difficult – determining how each individual piece of the protein will fold from its amino acid structure is challenging enough before trying to predict how different subunits interact. (See BioWorld Today, Sept. 13, 2011.)

As a result, even when researchers have managed to design proteins that interact, those interactions have sometimes been in unexpected ways, and have led to proteins with unexpected shapes.

But scientists' technological ability to design interacting proteins, and to predict the shape their interactions will take, has been improving rapidly. King pointed to a number of recent reports, including using fusion proteins and adding metal ions to influence binding, that indicated progress in the area. "It's been a great year for protein design," he said.

His team's findings extend that winning streak: When they used X-ray crystallography to experimentally determine the shape of the protein they had designed, its actual structure agreed with their predictions "to within the width of an atom."

For the experiments now published in Science, King used the Rosetta software, a program that predicts how proteins will fold from their amino acid sequence. Rosetta spinoff foldit has been used to crowdsource protein folding problems. (See BioWorld Today, Aug. 13, 2010.)

King and his team programmed the software "to enable it to design large self-assembling proteins." They first designed a basic building block – King termed it "basically a very complex Lego piece" – by starting with a group of naturally occurring proteins that assemble into three-part structures, or trimers. By substituting some amino acids, they were able to make trimmers that aggregated into larger structures. 12 trimeric subunits could self-assemble into a tetrahedron, or three-sided pyramid, while 24 of them could bind together to make an octahedron, which looks like two four-sided pyramids stuck together at their bases.

The structures designed by King and his team, as well as the structure described in the second paper, are cage-like. King said one possible use for them would be as targeted drug-delivery vehicles. King compared such self-assembling protein shells to a viral envelope. "Without self-assembly, the [viral envelope] protein could not fulfill its function" as, basically, a targeted delivery vehicle for the viral genome.

King said such drug targeting currently is being attempted with approaches like liposomes – fat droplets that can transport fat-soluble drugs – or antibody-drug conjugates. "But I think that proteins, as materials, have a lot of advantages over those other approaches," such as the fact that they will naturally interact with other proteins with very high specificity.

The potential uses of self-assembling proteins could extend far beyond drug delivery. The subunit that King and his team described in their paper does not aggregate into larger than 24-piece assemblies. But theoretically, self-assembling nanobiomaterials could be turned into fibers and crystals that could make larger-scale structures. One possibility, King said, would be to make a crystal out of designed enzymes – a structure that "could have a very high density of active sites."