Meganucleases are naturally occurring endonucleases that recognize and cleave specific DNA sequence motifs 14-40 base pairs in length1. Through studying how these enzymes recognize and cleave DNA, researchers uncover the most important steps in the protein-DNA interaction, which is an essential process in all living organisms. In addition, their incredible specificity, which has earned them the alias of "homing endonucleases", makes these enzymes an attractive subject for bioengineering projects that aim to modify DNA for various applications, from commercial to therapeutic ventures.
In their recent publication in Nucleic Acids Research, scientists in the Stoddard Laboratory (Basic Sciences Division) engineered 5 new variants of the meganuclease I-OnuI that specifically and efficiently cleave new target sequences. I-OnuI is named after the fungus Ophiostoma novo-ulmi in which it was discovered. The scientists generated crystal structures of each of the variant enzymes in complex with DNA and analyzed the changes in the protein-DNA contacts in detail. Said principal investigator Dr. Barry Stoddard, "Each of the proteins harbors mutations at up to 1/6 of their amino acids positions (most localized to the protein-DNA interface) making them some of the most dramatically reengineered binding proteins described to date."
To identify functional variants of I-OnuI, the scientists expressed a combinatorial library of enzyme variants on the surface of transformed yeast. This library focused on varying the amino acids in the protein-DNA interface and included a large number of variants since it was unknown which types of mutations might disrupt the protein structure. "When mutating amino acids in the meganucleases, the sidechain contacts to the DNA are the minority of the contacts. Most of the contacts are to the protein backbone - which requires much more trial and error to alter," said Rachel Werther, a lead author of the study. "Not only that, the protein is making contacts with the DNA phosphates more often than with the unique parts of the DNA nucleotides - the bases - so it seems that the protein is recognizing the bend of the DNA foremost. These two ideas: 1. protein backbone makes the contacts with DNA, and 2. protein recognizes DNA backbone instead of bases, make the protein reengineering a time-consuming project with a lot of randomization."
The scientists screened for meganuclease variants that could cleave specific DNA substrates using a flow-cytometry based assay where the cleavage of a fluorescently labeled DNA substrate changed the cell staining intensity, enabling cell sorting and identification of the active enzyme variant. They further verified the specificity of each enzyme to the new target sequence by measuring the frequency of mutations at the endogenous DNA target sites (in cells) as well as through further in vitro cleavage assays where they also measured the ability of the enzymes to cleave similar or dissimilar sequences. Five highly specific variants were chosen for further analysis.
High-resolution structures, ranging from 2.08-3.11 Angstrom, of wild-type I-OnuI and the five variants in complex with their DNA substrates were determined by x-ray crystallography. The structures gave the scientists an atomic-level view of the specific contacts at the protein-DNA interface, which were found to be considerably different among the variants. However, the overall structure of the enzymes themselves differed by less than 1 Angstrom root-mean-square-deviation (RMSD)—an indication of high overall structural similarity—and the DNA substrates differed by 1.5 Angstrom RMSD. In the wild-type enzyme, 22 amino acids clearly contact nucleotides and in the reengineered variants a similar number, ranging from 17 to 25 amino acids make nucleotide contacts. For each of the variants, the scientists analyzed how many amino acids in the DNA interface exchanged roles, such as switching from directly contacting the DNA to playing an indirect structural role. They found that approximately 1/3 of the amino acids in the interface had changed roles in each of the reengineered variants. These changing roles resulted in clear differences in the target DNA sequence the reengineered meganucleases recognized.
"DNA-recognition specificity is derived from the combined contributions from DNA-contacting residues and from neighboring residues that influence local structural organization," said Dr. Stoddard. "Changes to specificity are facilitated by the ability of all those residues to readily 'job-swap' and exchange both form and function. The entire interface therefore behaves as a highly fluid and malleable contact region that can readily adapt to the need for new recognition specificities driven by either protein engineering experiments, or by evolutionary requirements."
In the future, engineered meganucleases such as these could potentially be used for therapeutic genome engineering. "We want our meganucleases to recognize and cut important DNA targets - in cancer or HIV, for example - while avoiding dangerous off-target cutting," said Rachel Werther. "The engineered meganucleases in this paper cut a varied population of DNA targets with high fidelity."
Werther R, Hallinan JP, Lambert AR, Havens K, Pogson M, Jarjour J, Galizi R, Windbichler N, Crisanti A, Nolan T, Stoddard BL. 2017. "Crystallographic analyses illustrate significant plasticity and efficient recoding of meganuclease target specificity." Nucleic Acids Research.
This research was supported by the National Institutes of Health, the Bill and Melinda Gates Foundation, Fred Hutchinson Cancer Research Center and Bluebird Bio, Inc.
Conflict of interest statement: JJ and KH are employees of Bluebird Bio, Inc., which uses engineered meganucleases for genome engineering applications.
1. Stoddard BL. 2005. "Homing endonuclease structure and function." Quarterly Reviews of Biophysics.