Dear all,
I am using a plasmid with following organization; 5'-insert-IRES2-GFP-3' (i.e. bicistronic vector). My question; how does expression of the GFP correlate with (or influences) the expression level of the preceding cDNA insert; does high fluorescence imply high expression levels of the insert? Can I use FACS to select the brightest cells from the total cell pool and be confident that these simultaneously correspond to the best expressing cells of my cDNA insert? In other words, does the GFP fluo signal have any predictive value?
I would be very grateful to anyone with any answer, pointers, or good literature references ...
Cheers,
Jonathan
Hi Jonathan, I would recommend you to try running a Western blots of different clones you have. Probe using anti-GFP and anti-YourProtein/Tag in the SAME membrane and quantify using any imaging software (ImageJ for example) so you know if the expression of GFP in each transformation event is correlated in anyway with your protein expression level. After getting that experiment done, I think you will be able to use the FACS to get the best expressing cell clone.
Bests!
Yes, in principle this is the idea behind these vectors.
In practice there are some caveats. You may get integration of the (promoterless) GFP fragment without your Gene of Interest near a promoter in the genome, thus green cells which do not express GoI. You may also have mutations in the IRES meaning that you get expression of your GoI without much GFP. Not that common but does happen, so pick a few clones to check.
Also, if this plasmid uses a CMV promoter, this is susceptible to epigenetic silencing in certain cells (one reason to freeze a large batch as early as you can and then return to this whenever you need more protein).
Keep in mind, that if you have brighter GFP, likely you have more copies if the plasmid integrated. Also, cells that do express really high GFP won't grow particularly well.
To summarize, your plan is feasible, but once you have selected high GFP clones, you still need to check for your GoI expression and keep an eye on the expression level over time.
These papers may be useful:
Liu WM, Xiong YZ, Gossen M. Stability
and homogeneity of transgene expression
in isogenic cells. J Mol Med 2006;
84: 57–64.
McBurney MW, Mai T, Yang XF, Jardine K.
Evidence for repeat-induced gene silencing
in culturedmammalian cells: inactivation
of tandem repeats of transfected
genes. Exp Cell Res 2002; 274: 1–8.
Brooks AR, Harkins RN, Wang P, Qian
HS, Liu P, Rubanyi GM. Transcriptional
silencing is associated with extensive
methylation of the CMV promoter following
adenoviral gene delivery to muscle.
J GeneMed 2004; 6: 395–404.
Hi Xi and Petra, thanks for your insightful answers!!
I transduced 13 constructs now (varying insert sizes and gene expression levels), and used FACS to enrich the cells, taking the 10% lowest GFP expressers and 10% highest GFP expressers. I’ll do Westerns next week to verify whether these correspond to the lowest and highest expressing cells of the transgene as well. If so, it means that the fluo signal has good predictive value and hence that the approach of taking the most fluorescent cells is indeed valid. I’ll let you know.
Cheers
Jonathan
Following up in this, I made a lentiviral plasmid expressing a tandem mRuby2-IRES-EmGFP construct, and investigated correlation of red and green fluorescence. Image detailing the results is attached.
Using quantitative fluorescence microscopy, I found a Pearson's r (correlation coefficient) of 0.89 between mRuby2 and EmGFP fluorescence, with EmGFP translation under control of an IRES.
Simultaneously, I observed that IRES2-EmGFP fluorescence is >20-fold less compared to 5' cap-dependent translation of EmGFP alone. 5' cap-dependent mRuby2 fluorescence is not significantly different when comparing mRuby2 alone vs. mRuby2-IRES-EmGFP.
Similarly, using FACS, I found a Pearson's r of 0.85 between mRuby2 and EmGFP fluorescence, with EmGFP under control of IRES.
These results suggest that indeed, IRES-EmGFP fluorescence can be used as a qualitative and quantitative predictor of transgene expression, and its presence does not significantly impact the expression level of the preceding transgene.
Cheers!
Jonathan
Hello,
Have you ever tried a construct without gene 1 (promotor - IRES - gene 2), e.g. as control?
Hi Farkas,
No, I don't think I have specifically done that. However I did have a case where gene-1 barely expressed and I still got good GFP signal, so this was an instance where the relationship sort of broke down. I think that ultimately a fusion protein is best to ensure 1-to-1 stoichiometry, and 2A peptides probably function more robustly as a predictive marker. However, I'm often making protein for structural studies and don't want the C-terminal 19-aa footprint of the 2A peptide ...
I also don't want additional aa on my proteins. I wanted to use the no-gene-1 empty vector to set up and optimize the transfection and induction protocol (I have an inducible promotor) and later as control for the overexpressoin experiments but I've got no GFP expression at all.
I see ... I've uploaded here the plasmid map of my empty IRES-EmGFP backbone. The multiple cloning site is empty (12 bp stuffer sequence) but is has the secretion signal and C-terminal tags in place and everything is in frame. Maybe this setup is sufficient to successfully start transcription and then eventually translation of the IRES-EmGFP? Worth a go perhaps ...
- Badges
- Science topic