First check if this a protein fragment exposing large hydrophobic patches or a protein that requires specialized folding partners or post-translational modifications in order to refold properly in vivo, so as to avoid wasting time trying to get soluble expression. If the protein is supposed to form disulfide bridges forget about expression in normal strains, and try periplasmic expression or expression in a gor trx background (Origami or SHuffle strains).

Then test whether the inclusion bodies can be solubilized with mild treatments (detergents, etc. instead of the traditional chaotropes). Many proteins have not evolved to be expressed at high levels and aggregate upon overexpression in E. coli, but still retain their native conformation in inclusion bodies (see the work from the Villaverde group in Barcelona, for instance). In that case, aggregation into IBs may even be advantageous vis-a-vis downstream purification.

Then try lower temperatures (18-20*C) or poorer media (I once had a protein that was expressed in soluble form in LB but not in autoinducing media). You can try the pCold system from Takara here.

If that does not do the trick you can try co-expression of chaperonins. Several manufacturers sell ColE1-compatible plasmid systems containing the GroESL and DnaKJ chaperones.

There are other tricks you can try but in the end, if the protein is complex enough if may be better in the long run to switch to an eukaryotic host (Pichia, baculovirus or mammalian being the most well known).

Hope it helps!

Some basic questions:  Is this a membrane protein or a multimeric protein?  Does it normally have a pre or pro form?  What is the molecular weight and are there a couple or more disulfide bonds with an even or uneven number of cysteines; if so you may be stuck with insoluble expression?   How much material do you need to produce?  Is there an affinity tag?  With decent expression level and 15--20% refold efficiency, running refold matrices is quite advantageous in terms of downstream purification operations; there's far less host cell contaminants and a tag amplifies this advantage.  Some investigators make cleavable fusion constructs with other proteins such as MBP or GST to cause solubility but the mechanism isn't necessarily increasing solubility and ends up complicating the correct folding as well as necessitating a cleavage step.  

E. coli has a negative Eh (reduction potential) intracellularly so your protein's cysteines will not be oxidized to disulfides which are likely required for correct folding.  You would have to hope for correct conformation in the inclusion body so the cysteines are correctly juxtaposed for oxidation upon solubilization; GCSF is an example.   Periplasmic expression might get a positive Eh but how much expression with correct folding to you expect to get  there?  The real art in refolding is dependent upon reading SDS-PAGE gels reduced vs non-reduced in order to reasonably screen the multitude of refold conditions.  A correctly folded protein will have a distinctive band morphology and band pattern discerned on gradient gels with Coomassie staining.  

Hello! 

Thank you for detailed answer. 

1) It has a pro form. It is a single-pass type I membrane protein

2) MW is about 6,5kd. three intramolecular disulfide bonds (as far as I know, contains 6 cysteine residues that form three intramolecular disulfide bonds).

3) I will adjust my lab work to get the highest yield working with Ecoli. At first I will take all protein I can get in like from 1liter Ecoli-broth solutions. So I will need to process a lot of products in the future (bioreactor sizes). 

4) No tags

Thank you.

If the pro-region is not the signal peptide sequence then it's unusual to not have the pro-domain involved in the folding; for example, GDF15 is an exception.  You might have included this domain in your construct, especially if you're working with the ECD of a type 1 receptor.  Hopefully the N-terminus is not involved in ligand interaction because the pro-domain, if present, would need to be cleaved off.  Furin might be one possible endoproteinase to do this.  I assume that at 6.5 kDa for a receptor you might be working with the ECD domain that interacts with ligand and that the construct's C-terminus occurs in the stalk region?  If these characteristics are true then you stand a reasonable chance to fold this protein but I don't know about final functional activity by cell-based assays.  

If this is the full-length receptor then you might expect the protein to be associated with lipid from the E. coli inner membrane envelope.  In this case the centrifugal pellet of the cell lysate may be primarily a membranous layer containing your receptor.  There are methods in the literature for isolation of recombinant receptors expressed in E. coli and it usually involves peptidoglycan disruption, fragmentation of the envelope, relatively low speed centrifugation, ultracentrifugation, homogenization of the resulting pellet with maltoside-based detergents, and isolation by SEC.  This assumes correct folding in the inner membrane/periplasm.   By the way, do you actually see inclusion bodies with refractility in the post-induction cells under phase contract microscopy?  If you disrupt cells by sonication or other type I methods, how do you determine the efficiency of breakage?  If there was a significant amount of unbroken cells left and expression was soluble then the resulting pellet would contain whole cells with soluble protein thus giving the impression of insoluble expression.

Thank you very much!! I will take all these into consideration. If anyone else would like to add anything please do, so other researchers can read and have insight as well. I belive many of us struggle in refolding since its tedious work. 

Embrace the inclusion body!! Growth at lower temperatures can help.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1220970/

Lower temperature will definitely help. Even try and lookout for new techniques/process to extract the periplasmic protein. Optimization of parameters should yield an increase in the soluble protein expression pattern.