Protocol Online logo
Top : New Forum Archives (2009-): : Protein Expression and Purification

Problems in expression of target protein in E. coli - (May/16/2013 )

Hi everyone.
I'm facing a problem in expressing my target protein by using BLR(DE3) E. coli for months so your comments would be very valuable for me.
More than 3 years ago, I have constructed a plasmid using pET22b(+) host strain and stored in glycerol stock at -80 C conditions. The expression of my target protein had been well until last year when it began significantly decreasing in yield (like less than 10 times compared to before with the same protocol). Gradually in later expressions, I didn't get the band of the target protein when running SDS-PAGE. Seemed like E. coli expressed a different kind of protein due to it migrated at different location.
My target protein's molecular weight is 14 kDa and the unwanted protein is about 16 kDa.

So, I transformed previously constructed plasmids into BLR(DE3). And the results were the same. Strangely, the band has similar molecular weight with the unwanted protein above. I mean its molecular weight is 16 kDa.

And I constructed again a new plasmid, tested by sequencing to make sure everything is OK, transformed into a newly bought BLR(DE3) and it happened the same.

Btw, I'm using TB solution containing extra proline for nutrients. My target protein is highly hydrophobic, contains many repetitive sequences and a His-tag at the C-terminus. I purified by using Ni-NTA resin. After purification, it always contains an impurities which molecular weight is about 31 kDa besides the target. However, these impurities precipitate during dialysis and easily removed. Recently, amount of the impurities increase and hard to precipitate so this is also another problem.

I didn't get what is wrong here? I guess the problem is not on the E.coli but the expression media? Or sth contaminant?
Anyone knows this situation before?
I'm looking forward to your advice.
Sorry for my bad English and Thank you very much (from Japan)

P/s: I just use E.coli to express the target protein so I have very little background on this field (protein expression)


Are you doing each and every test expression using freshly transformed cells? Although not always the case, I have seen proteins that for some reason will not express after growing the culture from a frozen stock. Inoculate your expression cultures from a fresh colony picked directly from a fresh transformation plate.

Also, the 16 kDa vs 14 kDa may not be because the expression product is different. Since 3 years ago when the expression was working, the molecular weight markers you are using now may be different than what you were using then. The other possibility is that hydrophobic proteins have a tendency to migrate with apparent molecular weights on SDS-PAGE slightly different than their actual molecular weight, in which case the actual MW really is 14kDa even though it is running at 16kDa. To identify if this is the case, do an in-gel trypsin digest of your 16kDa band and send for mass spec sequencing. It's not too terribly difficult or time consuming to do, and would remove any doubts as to the identity of your expression product.

Also, you are getting more contaminant bands after Ni-NTA purification simply because the expression is inefficient. Spending the time to optimize expression may be worth the time and effort. Try different expression vectors and cell lines (easier said than done, I know). Also consider adding a cleavable fusion protein to assist in solubility (SUMO, GST, GFP). Try different media (LB, 2xYT, auto-induction), expression temps (37, 30, 25 and 18C) and expression times (3 hrs to overnight). I find in general that students are much too quick to settle for whatever expression construct they try first because they don't want to take the time optimize, but in the long run the extra time spent up front to get more protein per expression more reliably will save time and money. If you express your protein with a GFP tag, optimizing expression can be super easy because you can just read fluorescence of whole cells to get an idea of how much protein is expressed without having to lyse and purify the protein.

What buffer are you lysing your cells in? If it really is that hydrophobic, and you aren't adding any detergent, then you could be losing your protein when you pellet the insolubles even though you are getting decent expression.

Hope those ideas offer a little help.


Hello. Is there a reason that you need to use BLR (DE3) cells? I have used this host strain several times when either plasmid or prophage stability was a problem and the protein did not express well in BL21 (DE3) but I do think that there is a trade-off with using this strain. Expression levels can be lower and its growth is very different. If there is not a reason you have to use them, I would try BL21 (DE3). Also, I spent some time on the phone with Novagen's technical support recently about this cell line and they recommend that tetracycline is used at minimum for the glycerol stocks and perhaps for the productions too.


do you know if the 16 kDa protein has something to do with your protein? ...maybe you can identify it using mass spec? me this sounds like read-through transcription and low translation termination efficiency of your stop codon maybe? you use a single stop codon?

Has the 16 kDa the same intensity as the correct band had in your previous experiments?