Date of Award

Spring 5-15-2016

Author's School

Graduate School of Arts and Sciences

Author's Department

Biology & Biomedical Sciences (Computational & Systems Biology)

Degree Name

Doctor of Philosophy (PhD)

Degree Type



Cre recombinase recombines its DNA target, loxP sites, without help of accessory proteins or DNA repair systems. The simplicity of Cre-lox system has been widely utilized for genome editing, especially in mouse genetics. The goal of this dissertation is to constructCrerecombinase variants that will operate uponrecombination target sites (RTs) present within the genome, instead of perturbing the genome by inserting wildtype RTs for subsequent genome engineering. In general, the desired RTs native to the genome are asymmetric. However, the loxP sequence is pseudo-palindromic, requiring a homotetrameric formation of Cre recombinase. As a first step, I broke the symmetry of Cre tetramer so that each Cre monomer could be arranged spatially to bind distinct RT halfsites. I designed an alternative protein-protein interface for Cre. Then, I separated the mutations into a pair of Cre monomers. I could then arrange the assembly of this pair of complementary Cre monomers to form a functional heterotetramer, even though neither monomer exhibits activity alone. When combined with other mutations that confer distinct DNA specificities, the monomers preferentially formed the desired complex and recombined asymmetric DNA sequences with greater fidelity. Ive successfully found a pair of Cre monomers that do not work in isolation, but do when combined together. This has been successfully demonstrated in vivo in E. coli, mouse ES cell cultures and mouse retinal explants. As the next step, I need to change the DNA specificity of Cre recombinase to recognize native genomic sites. Surprisingly, the DNA preferences of Cre recombinase have not been thoroughly characterized. The 34 bp RT site loxP contains two palindromic arm regions and an 8 bp spacer region. The arm region is recognized by Cre monomers while homology of the spacer region determines compatibility between RT sites. While the consensus sequence of loxP is known, I performed the first high-throughput studies to determine Cres sequence specificity in the arm region. I broke the 13 bp arm region into 3 overlapping 5-6 bp small windows and used in vitro recombination and high-throughput sequencing data to generate logos for each window. I found that non-specific recombination can interfere with the analysis and careful selection of NaCl concentration is important for observing in vitro specificity. I have not only determined Cres sequence preferences, but also used similar methods to determine CreC2#4 (a Cre mutant) and VCre (a Cre homolog). In contrast to zinc finger and TAL effector domains, no modular decomposition of DNA specificity exists for Cre recombinase homologs. As a result, the RT specificity of Cre has previously been modified using directed evolution, a laborious approach. To accelerate the process, I used sequence information from homologs of Cre. By searching across genomes of different bacteria species, I found hundreds of Cre homologs. Closely related homologs share similarity in both amino acid sequence and predicted RT DNA sequence. By comparing residues that differ between close homologs in the aligned regions where Cre contacts the switched base pairs, I found candidate possible mutations for a specificity switch. The change in specificity was validated by the high-throughput sequencing assay. This demonstrates the feasibility of leveraging sequence alignment data to alter the specificity of Cre recombinase, reducing the amount of effort needed to generate mutants with novel RT preferences.


English (en)

Chair and Committee

James J. Havranek

Committee Members

Gary D. Stormo, Robi Mitra, Joseph C. Corbo, Ting Wang,


Permanent URL: https://doi.org/10.7936/K7QF8R45

Included in

Biology Commons