Comments above have certainly reviewed some of the biggest human proteins ever expressed in E. coli. All in all, it is certainly possible but it does take a lot of optimization. The physiology of bacteria and eukaryotic cells are vastly different and even if you get your protein to look perfect on SDS-PAGE it may actually turn out to be some sort of misfolded, inactive balls of junk. These risks are increased as the size of your target protein is increased. Other factors (toxicity is a big one) also limits the amounts of protein you can successfully express. I'd personally suggest other in vivo expression systems that are more physiologicall relevant, such as insect cells.

You should keep in mind that more than 50% of mammalian proteins are glycosylated and E.coli is very poor to do this kind of modification. Since the activity of your protein is mostly linked to glycosylation this is a big issue. Therefore not only the size matter.

That's a really good question, and the answer will vary depending on the protein, organism of origin, cofactors, structure, folding, etc. I have read of proteins as big as 100 kDa being successfully expressed in E.coli, but I'd say that's a very rare exception. The biggest protein I managed to express (but in micrograms per liter, not milligrams!) was about 60 kDa, expressed at low temp, and it was a bacterial enzyme (that kept activity). I tried a few human 90 kDa proteins, but with no success, despite thorough screening for strains, vectors, expression conditions. NOSs seem a rather tricky target to express, and it will likely require co-expression with cofactors, and under conditions that keep the enzyme inactive to reduce toxicity issues to your bugs. Periplasmic expression may help. In any case, it looks like a challenging target. Not impossible, but I'd say it will take quite a lot of effort, with no guarantee of success. I'd personally try another expression system, if possible. Hope it helps. Good luck.

I think that the more complicated the expressed protein (e.g. glucolysated ... etc)the better to use yeast (Saccharomyces cerevisiae) to express such protein since the mechanism of posttranslational modifications differ greatly between prokaryotes and eukaryotes.

Comments above have certainly reviewed some of the biggest human proteins ever expressed in E. coli. All in all, it is certainly possible but it does take a lot of optimization. The physiology of bacteria and eukaryotic cells are vastly different and even if you get your protein to look perfect on SDS-PAGE it may actually turn out to be some sort of misfolded, inactive balls of junk. These risks are increased as the size of your target protein is increased. Other factors (toxicity is a big one) also limits the amounts of protein you can successfully express. I'd personally suggest other in vivo expression systems that are more physiologicall relevant, such as insect cells.

Hi John,

You stated, "Human nitric oxide synthases (~1400 residues with 5 co-factors) can be expressed in E. coli to produce mg amounts by using low temperature expression". Is that expression in mg/L, or mg total with volume of culture not reported?

I remember there was a protease with about 130 kD, The authors are Rockel, Seyit, Baumeister. Paper was in EMBO or Science.

Hi Gary - I believe we get a few mg/l after purification but have not really published this work as we were trying to get the full-length structure determined and so far this is going slowly.

Hi Gary - in double checking - we did publish part of this in JBC- Elsa D. Garcin et al.,

J. Biol. Chem. 2004, 279:37918-37927. The supplemental methods may contain some info that is useful.

We could expressed several bacterial proteins of 50-60 kDa. The proteins were over-expressed at 37 C usually for 4h and collected as IBs. We could refold them successfully to monomeric or dimeric biologically active proteins, with final yield of up to 100 mg per 5 liters of flask fermentations.

Solubility of protein is a big matter as well. Hydrophobic proteins of which subcellular localization is membrane such as PM, EM, mitochondria and ER can be expressed in E. coli, but it is hard to get quite a few soluble proteins for in vitro studies.

Expression of large mammalian proteins (>60-70 kDa) in E. coli tends to be difficult. Certainly, there are some cases of plant and mammalian proteins up to about 100-150kDa expressed in soluble form in E. coli. For example a long time ago we expressed plant ClpA and ClpB like proteins (>100kDa not including His-tags). However, we were never able to express any fragments larger than about 40kDa of the DNA-dependent protein kinase (suffice it to say, full-length 480kDa protein was not stable in E. coli). In general, we have had better luck with mammalian proteins that are well expressed in their natural cell environment compared with low expressers. Further, multidomain proteins with low complexity sequence regions between the domains have not worked for us. There are however a number of one-off examples for large proteins. However, in our hands the success rate went way down. There are some things that can help. Probably the most likely alteration for increased expression of larger proteins is either optimizing the gene for E. coli codon usage or including rare tRNAs since this problem is greater with increasing size. In some cases toxicity results in plasmid loss. The Studier medium is pretty good for that or minimal media as they contain significant amounts of glucose that ensures reasonable repression of genes that ultimately are driven by the Lac operon (Bill's T7 media included). Try to stay away from LB-like media since there is often low levels of lactose or a similar compound that results in some induction and increased plasmid loss of toxic genes. Bill's expression media and minimal non-expression media are published in Protein Expression and Purification, and he has recently developed some improvements. If toxicity is not a problem then alterations in the conditions of expression can help reduce the solubility problem. Dropping the temperature at induction to 16C can be useful in increasing solubility, although it typically does not result in expression if none was seen at 37C in E. coli B (there are always counter examples but hard to predict). We typically start the culture at 37C and drop after the cells are growing well but before induction. We have never had much luck with these, but some people argue that you can use alternate strains, chaperones, or additives to the media. When we try to get proteins of this type expressed in E. coli (we typically try this first since much easier than other expression systems) we try to process as many conditions in parallel. There are a number of protocols out there. This allows you to test multiple conditions at the same time,and increasing your chances of finding one or more conditions that will work. In summary, larger than about 60-80kDa can be done, but it seems to be specific to the protein you are trying. My guess is that the physical limit is likely to be close to the largest proteins in E. coli ~200kDa

I recently listened to a talk about recombinant spider silk protein expression. I do not recall the exact numbers, but they were able to produce astonishing amounts (~8g/L (yes,that is 8000 mg/mL) of >200kDa proteins in fermenters. This is obviously a rare case and a lot of work goes into it, but it shows that expression of large proteins with high yields in E. coli is possible.