"Taming" proteins in a test tube: directed evolution of enzymes

The continuation of life requires energy, and the release, storage and utilization of energy all need to be achieved through chemical reactions, which rely on a special type of protein - enzymes. They are synthesized by living cells, catalyze various biochemical reactions inside and outside the body, and have been widely used in the production of food, medicine, feed and other materials.

Today, this chemical engine of life activities is undergoing a silent " domestication revolution ."

Enzyme.

The picture is copyrighted by the gallery. Reprinting and using it may lead to copyright disputes.

Just like agronomists improve crop varieties, scientists simulate the mechanism of natural selection to modify enzymes in a targeted manner: using gene mutations to produce functional variations, and then retaining the best individuals through artificial screening, thus overcoming the many shortcomings of natural enzymes, such as easy inactivation, poor stability, and possible side effects. So can we "domesticate" these proteins like we domesticate crops?

As early as 1859, British biologist Charles Darwin explained the biological principles of human domestication of crops in his masterpiece "The Origin of Species". In the natural reproduction process, organisms occasionally spontaneously produce random mutations, so even different individuals of the same species have differences. In the production process, people will intentionally retain the individuals that best meet their requirements, let them reproduce more offspring, and continue to select and retain individuals that better meet production needs among the offspring, and achieve crop domestication in a long selection process.

From this, we can summarize the essential elements of the "domestication" process: random variation and selection based on demand . In 1952, scientists had revealed through multiple experiments that the genetic material of cellular organisms is DNA, which directs the synthesis of proteins in cells. The essence of biological variation is the change of DNA in cells. In the process of domesticating crops, variation mainly comes from gene recombination during sexual reproduction and random mutations during cell division. However, the frequency of this variation is too low, and it takes a long time to achieve domestication. In order to increase the frequency of variation, scientists initially used a more "violent" method: applying some factors that can damage DNA to cells, such as ultraviolet rays. These factors will cause certain damage to the original DNA, thereby forcing cells to repair DNA. In the process of repair, it is inevitable to "make mistakes in a hurry" and achieve a higher frequency of mutations. However, such a mutagenesis process is too blind. It is easy to cause damage to cells and cannot ensure that mutations occur in the genes corresponding to the target protein, so it is easy to do a lot of useless work.

In order to focus mutagenesis on the target gene, scientists have invented a molecular biology technique similar to a "molecular scalpel" - polymerase chain reaction (PCR) . PCR is a technology that achieves the replication of specific DNA fragments in a non-cell system. Its core is the DNA polymerase that faithfully "copies" genetic information. When catalyzing the synthesis of new DNA, this type of enzyme can meticulously synthesize new DNA molecules according to the principle of complementary base pairing. Although it may occasionally "copy wrong", most DNA polymerases have their own "error correction" mechanism, which can identify and correct the "wrongly copied" parts.

The picture is copyrighted by the gallery. Reprinting and using it may lead to copyright disputes.

When we need to mutate a specific gene fragment, we can replace the DNA polymerase responsible for transcription with an enzyme that does not have an "error correction" mechanism, and increase the concentration of magnesium ions in the reaction environment to increase the probability of DNA polymerase "copying errors".

After obtaining these DNA fragments, they are linked to vectors and introduced into cells, and a large number of cell groups with random mutations of specific genes can be obtained. This precisely targeted molecular editing also brings more abundant random mutations, which increases the mutation efficiency by more than 100 times, truly realizing "surgical" genetic modification.

At present, scientists have "tamed" many enzymes through directed evolution technology, but this imitation of natural selection has an unavoidable problem: just like in nature, creating mutations in the laboratory is non-directional . If a new protein is created directly according to the blueprint in the mind, production efficiency will be greatly improved.

With the advancement of computer technology, scientists have begun to use information technology tools to predict the three-dimensional structure and corresponding functions of proteins and draw detailed "blueprints" of proteins. For example, the famous AlphaFold platform can achieve high-precision protein structure prediction. Anyone can see the corresponding protein structure by inputting a string of amino acid sequences, and can even predict the interaction between proteins and other molecules such as DNA and RNA.

On October 9, 2024, Demis Hassabis and John Jumpe of Google DeepMind shared the 2024 Nobel Prize in Chemistry with protein design pioneer David Baker for their prediction of protein structure.

Image source: Nature official website

In addition, the protein function prediction system "CLEAN" can make detailed and accurate protein function predictions in the database. These tools can help scientists to modify proteins more finely, and even create proteins with specific functions that do not exist in nature. Scientists can determine the most core component - the minimum active site - based on the function of the protein, and then use computational tools to gradually generate a complete protein skeleton blueprint. After multiple iterations and optimizations, the new protein synthesized according to the blueprint is highly consistent with the predicted structure and has a catalytic ability close to that of natural proteins.

With the help of these artificial intelligence tools, the research and creation of enzymes will inevitably become easier and more active in the future.

References:

[1]Eckert KA, Kunkel TA. DNA polymerase fidelity and the polymerase chain reaction. PCR Methods Appl. 1991 Aug;1(1):17-24.

[2].Saiki RK, Gelfand DH, Stoffel S, Scharf SJ, Higuchi R, Horn GT, Mullis KB, Erlich HA. Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase. Science. 1988 Jan 29;239(4839):487-91.

[3]KOSCHORRECK K, SCHMID RD, URLACHER VB. Improving the functional expression of a Bacillus licheniformis laccase by random and site-directed mutagenesis[J]. BMC Biotechnology, 2009, 9: 12.

[4]GUPTA N, FARINAS ET. Directed evolution of CotA laccase for increased substrate specificity using Bacillus subtilis spores[J]. Protein Engineering, Design and Selection, 2010, 23(8): 679-682.

[5]Tianhao Yu et al.Enzyme function prediction using contrastive learning.Science379,1358-1363(2023).DOI:10.1126/science.adf2465

[6]Anna Lauko et al.Computational design of serine hydrolases.Science0,eadu2454

Author: He Yiwen, undergraduate and master degree from Tsinghua University, middle school teacher

Reviewer: Li Xu, researcher at China Association for Science and Technology, associate professor at University of Science and Technology of China

Produced by: Science Popularization China

The cover image of this article comes from the copyright library. Reprinting and using it may cause copyright disputes