From 2012.igem.org

Revision as of 19:35, 29 August 2012

• teams
• Page
• Discussion
• Edit
• History
• Move
• Watch
• Log in

• iGEM
• Attributions
• Data
  • Data 1
  • Data 2
  • Data 3
• Notebook
  • Week 1
  • Week 2
  • Week 3
• Safety
• Project
  • Strain
  • Catalyst
  • Protein Engineering
• Team
• Home

Protein Engineering

Beyond the general circuit outline for our degradation of PET, there are also two different protein engineering projects we performed. In the first project we constructed a more effective and higher yield form of the LC-Cutinase protein. In a second less related project, we mutated the lac-repressor in order to change its ligand specificity. We changed its specificity to bind a harmful chemical called diuron instead of the natural IPTG in order to create a unique biosensor system.

LC-Cutinase Engineering:

Confirming expectation of common results from previous mutations

We had two goals with the engineering of cutinase: replicate mutations made in the paper by Sulaiman et. al., which consisted of three residues mutated to alanines, as well as to generate our own chosen mutations. Before we decided to replicate Sulaiman’s mutations we initially wanted to validate if we could expect the same results in our protein with the same mutations. In order to do this we wanted to compare best fit models (proteins with known structure that best fit an input sequence) generated for LC-Cutinase’s sequence and Sulaiman’s Tfu_0883’s sequence. Replicating Sulaiman we used Swiss-Model to generate a best fit model for LC-Cutinase, resulting in the protein 3visB with 61% homology. In Sulaiman’s paper they received 1jfr, a model that was second best in LC-Cutinase’s output with 55% homology. This similarity between generated models and high homology rate gave support to the expectation of similar results from the same mutations.
Further validation was found using a multiple sequence alignment between: LC-Cutinase, 3visB, Tfu_0883 and 1jfr (show pic). The alignment showed strong homology in all active site residues as well as between two of the three residues targeted by Sulaiman et. al. for mutation. The third targeted residue, though significantly different in Tfu_0883 compared to in LC-Cutinase, was expected to result in similar protein activity gains upon mutation to alanine. This was found by loading both models for 3visB and 1jfr in Pymol and assessing that the residues held similar placement above the active site. Their placement, as well as the placement of the other two targeted residues, showed that mutation to a smaller residue, alanine, would result in better binding ability of the active site.

Generating our own theoretical mutations

After validating the previous mutations we wished to make our own. The first step was to generate a theoretical 3D model for LC-Cutinase. This was done by using Swiss-Model again to fit LC-Cutinase’s amino acid sequence into the structure of 3visB. The results, including the stability of each part of the sequence, are as shown. (Pic) The results show generally high stability along the sequence. Also, upon overlapping this newly generated model of LC-Cutinase with 3visB’s model in Pymol (pic) we received a low RMS value of .78A. This resulting data supported the idea that the generated 3D model for LC-Cutinase was stable and fit 3visB’s structure closely. The next stage was to get our theoretical model for LC-Cutinase and the two ligands it would interact with, PET and pNPB, into the Foldit program, a protein folding simulator (link). In Foldit we could perform and assay theoretical point mutations for increased stability and function of the protein. However, because Foldit does not directly measure function but measures stability instead, it was necessary to load LC-Cutinase with each ligand docked correctly into the active site. This would allow changes made at the active site that increased stability due to better ligand/protein interaction to also equate to better ligand/protein binding and thus better function by the protein. Docking of the protein with each of the two ligands was done through Swissdock. The above images show the docking results as well as the most stable ligand binding placement. From here we loaded each specific ligand/protein combo into foldit and generated the following mutations with each rational (link). Mutation results: Diuron biosensor: Generating Mutant Library

For the Diuron biosensor we are taking a high throughput approach in assaying a large library of 41 mutations and all the resulting combinations of them in order to change the ligand specificity of the Lac repressor. To do this we constructed two reporter type constructs, a GFP reporter and a KAN resistance reporter (show diagram). In both of these reporter constructs we transformed a library of mutants generated from by a large series of SDM reactions on the wild type LacI gene. Using the reporter constructs we assayed the large library and found those mutations and combinations of mutations that most successfully changed the Lac repressor specificity. .

Diuron biosensor

Generating Mutant Library

For the Diuron biosensor we are taking a high throughput approach in assaying a large library of 41 mutations and all the resulting combinations of them in order to change the ligand specificity of the Lac repressor. To do this we constructed two reporter type constructs, a GFP reporter and a KAN resistance reporter (show diagram). In both of these reporter constructs we transformed a library of mutants generated from by a large series of SDM reactions on the wild type LacI gene. Using the reporter constructs we assayed the large library and found those mutations and combinations of mutations that most successfully changed the Lac repressor specificity.

References

Text goes here