If superior design is what you seek, look no further than your own front yard. Plants and photosynthetic microbes can convert sunlight and air into energy, with an energy efficiency that would make a Prius owner blush. In a less dramatic, but no less impressive sense, evolution has aligned the countless cellular machines that sustain a cell’s life into great harmony. Thus the appeal of mimicking nature for its design capabilities is obvious. And while evolution takes a few billion years to generate systems of immense complexity, its underlying principles are so simple that they can be harnessed and applied to a relatively short-term science experiment.

Charles Darwin’s theory of evolution, first published in 1859, was a very controversial idea. Aside from the well-known theological/philosophical controversies this new theory generated, it also forced biologists, who were fixated on descriptive sciences such as dissections and taxonomies, to consider how and why creatures had evolved to their present form. In this way, Darwin’s theory has influenced the entire field of biology since the Victorian era. Now, scientists are turning to the simple principles of evolution, spelled out over 150 years ago, to further their goal of designing better proteins with an approach called directed evolution.

Why build a better protein?

Before delving into the evolutionary principles, it’s helpful to understand exactly why scientists would want to design better proteins, and why using evolution to generate them is a promising approach. First, a quick primer on proteins. Proteins are molecules that accomplish nearly every task required in a cell – from communicating with other cells, to replicating a cell’s DNA, to converting sugars into energy. A catalytic protein, meaning one that increases the speed of chemical reactions without itself getting used up, is called an enzyme. All proteins are chains of single units called amino acids, and there are 20 different amino acids to choose from at each position along the chain. The arrangement or “sequence” of amino acids in the chain directly influences the protein’s form and function. And as a consequence of this, changes in a protein’s sequence through the forces of evolution can potentially influence how the protein behaves.

Proteins have a number of desirable characteristics – the ability to catalyze many reactions quickly and efficiently, the ability to act on only very specific compounds, and a relative ease of synthesis. These characteristics make it desirable to try to repurpose them for a range of tasks outside of their naturally evolved habitats. For example, proteins could be engineered to manufacture drugs or convert sunlight into biofuels. One exciting approach scientists have taken is known as rational design, which involves using our knowledge of biochemistry to make directed and specific sequence changes to proteins to repurpose them from their original biological role to a more scientifically or industrially useful one.

However, bridging the gap between the enormous potential of engineered proteins and the realization of a successful functioning protein has proved quite difficult, as explained in a previous SITN Flash article [1]. In short, scientists have a limited understanding of the link between a protein’s amino acid sequence and its ultimate function. Tinker with a random few amino acids, and you’re far more likely to get a non-functioning blob than the protein you set out to design. Enter, evolution.

The process of directed evolution

What are these simple evolutionary principles? As the science philosopher Daniel Dennett writes, “evolution will occur whenever and wherever three conditions are met: replication, variation (mutation), and differential fitness (competition)” [2]. Let’s unpack that quote by explaining how this works in nature, and then looking at the parallels in a directed evolution experiment. For an organism to remain in the gene pool, it must be able to reproduce. Variation is introduced with each new generation because DNA replication isn’t 100% accurate, and sexually reproducing animals have offspring that are hybrids of their parents’ genes. Therefore, every organism produces offspring that are slightly different from the parents. Finally, since each individual in a population is slightly different from all of its peers, it may have a higher or lower chance of reproducing based on the specific advantages or disadvantages its variations have granted it. The more fit individuals reproduce more, and the cycle continues.

In the lab, directed evolution looks a bit different. If scientists want to evolve a protein that can carry out an important step in the manufacture of a drug, they will start with a natural protein that accomplishes a similar task in a living cell. (Microbes and plants are replete with molecules that humans have found to be useful as drugs, and thus these creatures also carry the protein machinery to make said molecules.) To achieve variation in the selected protein, scientists create a so-called ‘library’ of many different versions of the gene that encodes the protein of interest. Using modern day synthetic technology, scientists can create thousands of slightly altered variants of the gene of interest, making random mutations throughout the gene or specifically at sites in the gene that they think will affect the function of the protein it encodes. They can also start with the naturally-occurring gene and replicate it in a way that introduces relatively high levels of random mutations.

Other methods also exist, but however it is generated, the library of proteins is then put in competition amongst itself. This can be accomplished in two main ways: using a screen or a selection. A screen is when the scientist tests each protein in the library in parallel, and checks to see if any of the proteins accomplish the desired new task to satisfaction. A selection is an experiment where only the successful proteins are retained, and the unsuccessful mutants are discarded and never considered. The benefit of a selection is that it allows the researcher to hone in on only the proteins of interest, instead of monitoring every single member of the library; the downside is that selections can be difficult to design. Once the successful mutants – the “hits” – from the screen or the selection have been identified, the scientists have a choice: they can begin the ‘replication’ phase of the experiment, wherein they generate a new library based on the successful mutants that they’ll rescreen or reselect for the optimal mutant, or, if they are satisfied with the function of the mutants, they can carry on with the next stage of their research. See Figure 1 for a summary of the parallels between natural and directed evolution.

Figure 1. Variation in nature exists due to mutations and sexual reproduction. In comparison, a scientist creates a library of protein variants for a directed evolution experiment. While many factors (for example, predation) contribute to the survival and reproduction of organisms in nature, proteins in the lab are selected based on their ability to carry out a chemical reaction in a screen or selection. Scientists further study the most successful protein, mimicking the process by which successful organisms further their proportion in the gene pool. Image by Hannah Somhegyi.

Success of directed evolution

Directed evolution, an idea conceived in the 70’s and brought to prominence in the 90’s, has already achieved many notable successes. A recent review lists several dozen such stories, and also gives more details about each step of a directed evolution experiment [3]. One notable example involves the manufacture of the diabetes drug sitagliptin (marketed under the name Januvia) [4], the 14th-highest selling pharmaceutical in 2013. Formerly, one step in the industrial synthesis of the drug required very high pressure, a heavy-metal catalyst, and numerous purification steps because the product was impure. Researchers at two pharmaceutical companies decided to investigate using a protein biocatalyst to carry out the reaction, which would make the manufacture cheaper, better yielding, and less environmentally destructive. They began by using rational design, described earlier, to repurpose an existing enzyme so that it recognized a new substrate, the drug precursor. Once the engineered protein accepted its substrate, they carried out several rounds of directed evolution to increase the speed and yield of the reaction. The resulting optimal biocatalyst could perform the same reaction as the previous chemical approach, but with a 53% increase in productivity, >10% more yield, 19% less waste, and without using any hazardous heavy metals or expensive high-pressure machinery.

Evolved proteins may be used to address important problems in many different fields. At UC Berkeley and many other institutions, scientists are trying to turn cells into biofuel factories by evolving them to metabolize cellulose [5]. Cellulose is a polymer of sugar molecules that forms a major portion of every plant, but it is inaccessible as an energy source because it is indigestible by most enzymes. As another example, researchers at MIT are attempting to evolve antibodies that bring drugs directly to cancer cells [6]. These and other exciting projects emphasize that the principles underlying directed evolution and its associated techniques will likely remain relevant well into the 21st century.

Jeff Bessen is a graduate student in the Chemistry and Chemical Biology PhD program at Harvard University.


[1] Kelsic, Eric. Evolving better proteins… with a little help from viruses. SITN Flash, Oct. 2, 2011. http://sitn.hms.harvard.edu/flash/2011/issue102/

[2] Dennet, Daniel in Encyclopedia of Evolution (ed. Pagel, M) E83-E92 (Oxford Univ. Press, New York, 2002).

[3] Sharma, Rohit, and Jasjeet Kaur (2006). Directed evolution: An approach to engineer enzymes. Critical Reviews in Biotechnology, 26, 165-199.

[4] Savile, Christopher K, et al (2010). Biocatalytic asymmetric synthesis of chiral amines from ketones applied to sitagliptin manufacture. Science, 329, 305-309.

[5] Better biofuels through evolution. Understanding Evolution Resource Library, UC Berkeley, April 2009. http://evolution.berkeley.edu/evolibrary/news/090401_biofuels

[6] Trafton, Anne. Explained: Directed evolution. MIT News, May 13, 2010. http://web.mit.edu/newsoffice/2010/explained-directed-evolution-0513.html

One thought on “Anything You Can Do, I Can Do Better: How and why scientists mimic nature to design new proteins

Leave a Reply

Your email address will not be published. Required fields are marked *