Machines Evolving to the Brink of Failure

While writing a post on molecular engineering for the Macromolecular Modeling Blog, I came across an EMBO Reports paper that provided new guidelines for protein engineering; It also illustrates a general principle that should be taken to heart by anyone thinking about molecular engineering from a biomolecular perspective:

Molecular machines tend to evolve toward the brink of failure, and it is wrong to blame their structural materials for this fragility.

The EMBO paper examines the tricks that enable natural protein molecules to avoid — usually — a dangerous failure mode: unfolding and sticking to one another to form an extended beta sheet. Many functional proteins contain bits of beta sheet (parallel polypeptide chains linked by hydrogen bonds), but runaway beta sheet formation is pathological. The amyloid fibers characteristic of Alzheimer’s disease form this way, so the process is of more than academic interest.

As regular readers know, protein can be a robust, high-performance polymer (and isn’t at all like meat). Nonetheless, proteins in biological molecular machines are commonly poised on the brink of unfolding, and quite unsuitable for engineering use. Why should anyone expect that artificial proteins would be better?

Quite simply because evolution doesn’t select for a property unless it offers some advantage, and evolutionary pressures for maintaining fold stability don’t begin to operate until a molecule evolves to the threshold of instability. If there were a large margin of stability, evolution would optimize some other property, or simply drift, until the margin became thin.

Recognizing this general principle persuaded me of the feasibility of protein engineering and of a biomolecular path to “the development of general capabilities for molecular manipulation”, leading to a range of capabilities that I later termed “nanotechnology”. The term caught on, and protein engineering took off, but the connection between the two was largely forgotten.

Amyloid: It’s not just a problem for brains

The EMBO reports paper is titled “Prevention of amyloid-like aggregation as a driving force of protein evolution”, and I find it interesting because points out that proteins have evolved to the brink, not just of transient unfolding, but of forming insoluble aggregates. This is not a property of a single molecule, of course:

As protein aggregation is a multimolecular process, its rate is highly dependent on protein concentration….Proteins typically expressed at high levels have a sequence with a low aggregation propensity, whereas polypeptide chains expressed in tiny amounts have a lower overall solubility….[D]iverse and complementary mechanisms to protect proteins from amyloid-like aggregation probably allow the fine regulation of their sequences, and maintain the delicate equilibrium between folding, aggregation and expression levels.

Thus, the paper describes another instance of evolution flirting with a failure mode.

In describing mechanisms that protect natural proteins from amyloid-like aggregation, the paper also describes guidelines for successful protein engineering. Just as designers can aim for stability greater than that of natural proteins (and have succeeded), they can also aim for greater resistance to aggregation. As usual, aiming for a margin of safety helps compensate for the margin of error in modeling.

As the authors say,

This nature-derived knowledge is extremely important as it can provide inspiration to rationally control aggregation when it is not desired, such as in pathology and biotechnology, as well as to promote it in a controlled manner when it is desired, as in the construction of new materials of biotechnological interest.

Unnaturally bold

Recognizing the tendency of proteins to evolve toward the brink of failure gives reason for designers and experimentalists to be bold in taking steps away from natural models: Artificial systems that exaggerate features learned from nature may lead to greater success. (Has this been pursued with vigor in engineering protein-protein complexes? An effective methodology for this will be a crucial development.)

Borrowing from human-developed technologies may help, too. Enzymatic catalysis on the natural model, for example, has been remarkably resistant to emulation (a really effective methodology arrived only recently). Although effective catalysis requires a delicate balance of binding and transition-state energies along a reaction pathway, chemists have achieved this in many families of organometallic complexes. Incorporating similar structures into the active sites of artificial enzymes would enable extensive tuning and customization, potentially increasing turnover rates and suppressing side-reactions. (When considering the potential range of applications, keep in mind that using enzymes in dry organic solvents is an established industrial technology.)

In general, designing systems “better” than those evolved in nature can be easy, when our criteria for “better” aren’t natural. With this recognition, a range of vague but widespread intuitions about the implications of biology for advanced molecular nanosystems melt way, or should.

---

Note: Even nature can outdo nature, in a pinch: human protein machinery rapidly coagulates (i.e., cooks) at 60 C, yet with similar machinery, strain 116 of Methanopyrus kandleri survives 2 hours at 130 C, and at 122 C, it actually grows.

---

Tagged as: bionanotechnology, molecular biology, Nanotechnology, protein engineering