Understanding Evolution

And the Nobel goes to...evolution!

November, 2018

Left to right: 2018 Nobel winners George Smith, Frances Arnold, and Gregory Winter

Last month, the Royal Swedish Academy of Science announced that this year's Nobel Prize in Chemistry would go to Frances Arnold (currently at the California Institute of Technology), George Smith (University of Missouri), and Gregory Winter (MRC Laboratory of Molecular Biology, Cambridge, UK) for innovations that are being used to fine-tune manufacturing processes to reduce environmental harm, produce new renewable fuels, and build pharmaceuticals that harness the power of the body's own immune system to fight disease. Their inspiration? Evolution by natural selection.

Where's the evolution?

Arnold, Smith, and Winter pioneered using the principle of natural selection in a laboratory environment (a technique called directed evolution) to build new molecules that perform a particular job. In Arnold's case, her breakthrough research demonstrated how to use the technique to tweak an enzyme that normally works only in water so that it worked in a different liquid. Smith and Winter's work was deployed to build antibodies that attach to particular proteins or cells that cause disease.

To understand why directed evolution is so innovative, it will help to consider two other possible ways of building new molecular structures: natural selection and rational design. Rational design is what engineers mainly use to build large structures like buildings and bridges. The designer determines what functions the structure will need to serve (e.g., supporting 4000 cars and withstanding a magnitude 7 earthquake), and uses an understanding of physics and materials science to logically reason out a design that will do that. However, building molecules that serve particular functions is, in some ways, more difficult than building bridges that do. Our understanding of the rules of molecular design is inhibited by an inability to directly observe interactions at such a small scale and by the challenge of predicting exactly how the laws of physics and chemistry will play out among the hundreds or thousands of individual amino acids that might form a protein molecule. This makes it hard to solve molecular challenges using rational design.

Natural selection, on the other hand, requires no designer. The natural environment presents organisms with a set of challenges and constraints. Random mutations provide organisms with genetic variation that might (or might not) help solve those problems. Individuals carrying mutations that happen to make them better at dealing with the environment's challenges are more likely to produce offspring and pass those genetic variants on. As mutations continue to occur on the already favored genetic background and undergo selection by the environment, these solutions are fine-tuned. Natural selection produces molecules that are excellent at performing their jobs: hemoglobin effectively carries oxygen throughout our bodies, amylases break down carbohydrates into forms we can absorb and get energy from, actin and myosin form filaments that interact with one another to make our muscles work...the list goes on and on. However, natural selection works well only because population sizes are often very large and selection can occur over millions and millions of generations, enabling the testing of a myriad of variants that might solve any given challenge. And of course, natural selection only generates solutions to problems posed by the natural environment. It does not work to improve industrial manufacturing processes, make our cars run more efficiently, or cure diseases of the elderly.

Arnold, Smith, and Winter, however, figured out how to harness the principles of natural selection to work for us. Arnold's breakthrough work elegantly demonstrates how directed evolution tweaks the steps of undirected natural selection to make the process work in a laboratory, in a reasonable length of time, to solve a problem of our choosing:

  1. Pick a good molecule to start with. Natural selection starts with whatever molecules happen to be at hand. Directed evolution works best if scientists apply their knowledge to build or select a starting molecule that is closest to being able to perform the desired job. In Arnold's case, she wanted to wind up with an enzyme that breaks down proteins in an organic solvent, so she started with a gene for an enzyme (subtilisin) already known to do the same job in plain water.
  2. Introduce random variation, but not too random! Natural selection operates on genetic variation that is truly random: mutations may occur anywhere in the gene and happen without respect to the effect that mutation will have. However, screening that many variants, most of which will have no effect or make the molecule worse at doing its job, is not practical in a lab. Instead scientists use prior research and their knowledge of chemistry to determine what parts of the molecule play an important role in how it functions and introduce random variation only into the parts of the DNA that encode that part of the molecule. Arnold generated subtilisin DNA sequences that varied randomly only near chemically active sites.
  3. Figure out an efficient way to screen or select variants for their ability to solve the problem. In natural selection, screening occurs (ahem) naturally as organisms do or do not reproduce. Failing to reproduce effectively screens the genetic variants that organism carries out of the population. In directed evolution, scientists need a way to test the function of each individual variant they produced in step 2. Often this involves inserting all the randomly tweaked genes into miniscule, fast-reproducing organisms like bacteria and seeing how those critters perform. Arnold inserted her subtilisin DNA variants into bacteria, let the bacteria grow on petri dishes containing a protein, and identified the bacteria best able to break that protein down based on the size of the cleared area surrounding the clonal colony.
  4. Go step-wise with selection. In natural selection, the environment determines how harsh or lenient selection is. However, with directed evolution, scientists can make this determination. By starting selection with a mild version of the problem one hopes to solve, scientists can increase the chances of identifying mutations that partially solve the problem and provide a basis for future rounds of selection. For Arnold, this meant starting off with petri dishes that contained a relatively small percentage of the organic solvent — even though her ultimate goal was creating an enzyme that could perform well in the presence of much more solvent .
  5. Wash, rinse, repeat. Natural selection continues indefinitely, so long as a species remains extant: every generation, new genetic variation is introduced, and every generation, some of those variants may be favored, while others may be weeded out. In directed evolution, scientists select the gene variants that performed best at the selected task (usually by measuring how well bacteria or other critters carrying the genes perform), introduce more random variation into those variants and repeat the cycle of the selection until the gene solves the problem or a dead end is reached. With just three rounds of mutation and screening, Arnold produced an enzyme that breaks down proteins in an organic solvent 256 times better than the enzyme from the original gene did!

Greg Winter used a similar approach (along with work by George Smith) to iteratively vary and then select antibodies, proteins produced by the immune system to help defend the body from invaders. His company used this technique to produce the first drug based on human antibodies, adalimumab, which is used to treat conditions such as rheumatoid arthritis. Similar techniques are being used to develop antibodies that fight cancer, neutralize anthrax, and slow Alzheimer's disease. Meanwhile, Frances Arnold continues her work with directed evolution. She's currently working on enzymes to produce biofuels, environmentally-friendly plastics, and maybe eventually, products built from carbon absorbed from the atmosphere.

Scientists who harness the power of evolution are using it to produce remarkable, life- and planet-changing innovations. But that should come as no surprise to students of the biological world, where we can observe the impressive products of plain old undirected evolution everywhere: animals uniquely suited to their diverse environments, microbes that perform feats of survival and chemistry, and of course, plants that use sunlight to absorb carbon dioxide from the atmosphere and produce both oxygen and food – testaments to the myriad of problems that can be solved with nothing more than heredity, variation, selection, and time. We human problem solvers have finally caught on.

Primary literature

  • Chen, K., and Arnold, F. H. (1993). Tuning the activity of an enzyme for unusual environments: sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide. Proceedings of the National Academy of Sciences USA. 90: 5618-5622. read it
  • Clackson, T., Hoogenboom, H. R., Griffiths, A. D, and Winter, G. (1991). Making antibody fragments using phage display libraries. Nature. 352: 624-628. read it
  • Romero, P. A., and Arnold, F. H. (2009). Exploring protein fitness landscapes by directed evolution. Nature Reviews Molecular Cell Biology. 10: 866-876. read it

Discussion and extension questions

  1. Why is it challenging to use rational design to build new molecules that serve particular needs?
  2. Review the concept of natural selection. Use the four steps described on that page to explain how natural selection might improve hemoglobin's ability to carry oxygen incrementally over the course of many generations.
  3. What are the similarities and differences between the genetic variation that natural selection acts on and the genetic variation that is incorporated into directed evolution?
  4. What are the similarities and differences between the selection step involved in natural selection and the screening process that is incorporated into directed evolution?
  5. Advanced: What advantages might directed evolution have over rational design when it comes to building new molecules for particular purposes? What advantages might rational design have over directed evolution?


  • Chen, K., and Arnold, F. H. (1991). Enzyme engineering for nonaqueous solvents: random mutagenesis to enhance activity of subtilisin E in polar organic media. Biotechnology. 9: 1073-1077.
  • Chen, K., and Arnold, F. H. (1993). Tuning the activity of an enzyme for unusual environments: sequential random mutagenesis of subtilisin E for catalysis in dimethylformamide. Proceedings of the National Academy of Sciences USA. 90: 5618-5622.
  • Gibney, E., Van Noorden, R., Ledford, H., Castelvecchi, D., and Warren, M. (2018). 'Test-tube' evolution wins Chemistry Nobel Prize. Nature. Retrieved November 1, 2018 from https://www.nature.com/articles/d41586-018-06753-y


View this article online at:

Understanding Evolution © 2021 by The University of California Museum of Paleontology, Berkeley, and the Regents of the University of California