Designing a dual selection system
We constructed a dual selection system based on plasmid gYB2a-pobRWT-mCherry-codA-cmr. This plasmid mainly contains three parts:
1. The pobR coding sequence (CDS) which can express PobR protein;
2. The cytosine deaminase(codA) gene which can express CDase protein;
3. An engineered operon consisting of the red fluorescent protein(mCherry) and the chloramphenicol(Cm) resistant gene (cmr).
How this plasmid acts as a dual selection system is as follows:
1. Negative selection: Been activated by PobRWT, PpobA activates the expression of CDase which is available for conversion of exogenously added 5-FC to 5-FU, causing cell death;
2. Positive selection: Without no additional 5-FC, adding chloramphenicol and an aromatic compound to cultivate E. coli cells will get a library of PobR mutants. Strains with PobRmut that recognize the aromatic compound are expected to grow in the LB agar medium.
To further understand each part, please click here to view our design part.
To constructed gYB2a-pobRWT-mCherry-codA-cmr, we first tried to generate plasmid gYb2a-PobR-PpobA*2-mcherry-SacB-Cmr by Goldengate assembly. However, when applying the plasmid gYb2a-PobR-PpobA*2-mcherry-SacB-Cmr to achieve functions as a dual screen system, the results are not ideal.
We obtain the two target fragments, gYb2a-PobR-PpobA*2-mcherry-Cmr and CD by the method of PCR. Then the two fragments were ligated by using Goldengate assembly and transformed into E. coli BWΔcodA competent cells and the plasmid gYB2a-pobRWT-mCherry-codA-cmr was constructed.
The result of DNA sequencing showed that our gYB2a-pobRWT-mCherry-codA-cmr plasmid was constructed.
Directed evolution of PobR
Aiming to get strains that respond to different aromatic compounds by the dual selection system, a large PobR mutant library which is able to include as many mutants situations as possible is crucial. Therefore, we used error-prone PCR to construct the PobR mutant library.
The generated library with highly random PobR mutants was transformed into E.coli BWΔcodA to obtain transformants containing mutant plasmids. The PobR mutant library was transformed into BWΔcodA competent cells and transferred to M9 medium for culturing in shaking flasks. Ten clones were randomly picked to sequence their PobR CDS regions for the quality control, which revealed diverse mutants with an average mutants rate of about 0.36%.
Fluorescence assay and screening of the PobR mutant library
Another part which is essential to get the new ligands specificities and detection range of PobRmut is the screening part.
In the negative selection, the obtained strains were cultivated in liquid culture supplemented by 4HB and 5-FC. In the initial negative selection, we used a constant 4HB concentration of a relatively high level, 0.5 g/L, and then tested different concentrations of 5-FC to inhibit both the pseudo-positive and 4HB-responsive strains. In this selection step, we first used 50 mg/L 5-FC, and observed insufficient inhibition of the bacterial growth. Thus, we increased the 5-FC concentration to 200 mg/L, improving the selection effectiveness.
In the positive selection, we added seven aromatic compounds to LB agar medium and cultivated E. coli cells harboring a library of PobR mutants in this medium for strain selection. As for the aromatic compounds, we chose phenylethanol (2-PE), mandelate (MA), 4-hydroxymandelate (HMA), phenylpyruvate (PPA), 4-hydroxyphenylpyruvate (HPP), phenylacetaldehyde (PAld) and p-Coumaric acid.
After culturing for 30 hours, a few pink colonies were picked and cultured in a 96 deep-well plate. Since the expression of the reporter gene mCherry was positively correlated with responsiveness, all that required is to transfer the pink colonies to LB media containing only Amp after activation and used 0.5 g/L of the test ligands for initial characterization. At the same time, negative control without the addition of test ligands was used to avoid a few biosensor variants with a strong fluorescence response in the absence of any ligand. After performing fluorescence measurement，we got some strains that are highly responsive to aromatic compounds.
Besides, we used the dilution coating method to estimate the number of mutants which are capable of aromatic compounds per screen. The selection capacity for each compound was more than 900，000 clones (with at least four plates, the original density of the two-round selected bacteria was 450,000 CFU/mL.)
Please click here to view our notebook.
The specificity and detection range of PobR mutants responsive to new ligands
We obtained several responsive strains, of which the fluorescence intensity was 1.5-fold higher than that of the negative control. To further evaluate the PobR mutants obtained above, the second round of characterization experiments were carried out to individually examine their responsiveness to each candidate ligand. For each ligand, we selected a mutant strain with the highest responsive profiles and plotted the curve for their ligand response. In total, 9 potential biosensors were isolated after the second round of characterization, and all these variants were sequenced to determine the mutations in their primary sequences. Amino acids at positions 163, 177 and 234 are located near the ligand binding pocket of PobR, and amino acid at position 40 is located in the DNA binding domain.
Modeling and docking
To better understand and analyze the effects of amino acid substitutions on the response of the PobR protein, we tried to use software to simulate the structure of the protein and build a molecular docking between the PobR protein and its inducer.
We first use Homologous Model website SWISS-MODEL to construct the PobR mutant model using the PobR wild type as a template.
Secondly, we obtained the structure files of the ligands 4HB, 2-PE, MA, HMA, PAld, HPP and PPA from the organic small molecule database Pubchem. For more details you can click here.
Based on the structures of PobR monomer and ligands, the docking engine Autodock is used to simulate molecular docking. Autodock search space coordinates were set as center_x = -4.672, center_y = 3.331, center_z = -2.213. Dimensions of the search space were set as size_x = 40, size_y = 40, size_z = 40, and exhaustiveness was set at 15. The 15 conformational conditions in a score based on the lowest binding energy were listed as the docking results. To examine the accuracy of our docking, we used Ligplus to check the interaction between the small molecule and predicted protein receptors. Finally, the three-dimensional schematics of the protein and its ligand were portrayed using PyMol Version 2.2.0.
