I have a plasmid containing the sequence for a protein chimera composed of (from N to C) 2 repeats of the fibronectin type III-like domain of CD45, followed by EGFP, followed by the transmembrane and cytoplasmic domain of LDLR. The construct was synthesized as a gene-block from a commercial vendor and inserted via Gibson Assembly into the peGFPN1 mammalian expression vector under control of the CMV promoter.
Upon selecting colonies and sequencing, all colonies had mutations within the first few base pairs of the second FNIII-like domain. I selected the clone with the fewest mutations (missing two base pairs causing a frameshift mutation) and attempted a 2 bp insertion via site-directed mutagenesis in two ways - one time with overlapping primers with 14 bp of 5' overlap and the 2 bp insertion contained in the middle of the overlap region, and one time with back-to-back primers with the 2bp insertion contained at the 5' end of the reverse primer. Both attempts at SDM have failed, and in the attempt with back-to-back primers, 3 of the 5 clones had even more base pairs deleted from the same region. However, my PCR yield is always excellent (based on running a gel) and my control plate (used NEB's KLD enzyme mix on my template vector) had no colonies, which tells me my DpnI digestion is working and the colonies I get on the sample plate are not just the template from the PCR reaction carrying through.
I'm wondering if it's possible that this chimera is toxic to the E. coli and I'm getting leaky expression of the protein through the CMV promoter, so the colonies that grow on my plates are the ones that have mutations, so they are selected for during the growth stage of transformation. I know there are similar threads for expression of toxic genes, but I have no reason to believe that this protein chimera would be toxic to the cells (although there is some precedent that the FNIII-like region is unstable when only some domains are expressed Article Domain organization of the extracellular region of CD45
) . But why would I only get mutations at the same location in the gene? If the protein was toxic, wouldn't all frameshift mutations be equally selected for? This makes me think there is an error in my protocol, but having obtained the same results with two different sets of primers is making me very confused. Any insight would be greatly appreciated!
Advika Kamatar Very interesting question and definitely in the 99th quality percentile of questions that are asked on Researchgate! I don't have a good answer, but I'm wondering what would happen if you would sequence the PCR product of your SDM (prior to transformation). If that sequence was clean and had your desired structure, it would definitely indicate that selection for the aberrant structures you see in the clone plasmids was due to in vivo selection. You'd just need a sequencing primer from the area of the product that is stable and at the right orientation and distance to give you a nice read through the problem zone.
David F. Barker Thank you for the suggestion! I sequenced the PCR product before transformation and, indeed, the PCR product contains the desired mutation. So, it looks like there is some in vivo selection going on! I will attempt a lower growth temperature and possibly cloning into a vector with a more tightly controlled promoter, but if you have any other suggestions they would be appreciated as well!
Advika Kamatar I agree that a strategy to reduce the expression of your construct is the best first attempt to try to recover your desired product. I'm also guessing that the reason for finding a limited variety of "mutations" is that they are not mutations that are occurring in the bacteria after transformation but are rather pre-existing as low level variants in the PCR product that you clone and are likely particular artefacts of the synthesis/construction steps. Perhaps if you closely examine the sequence traces of that PCR product you might see some evidence of them. At the position of in/dels, you would start to see a low level out-of-phase background sequence, for instance.