Mostly BL21 strain forms your expressed proteins in inclusion forms. You can easily purify your protein and refold ( solubilizing and denaturing) it from the inclusion bodies. In case you need soluble form you better change the host or use yeast system instead.
as you don't say what your expression conditions are I can only give some general advice. You want to try to slow down the transcription & translation processes as much as possible to give your protein the best chance to fold properly. Trying different strains is also worth a try.
It is a good idea to examine a range of inducer concentrations, (if IPTG do a 0.5 to 0.05mM range). Induce at 18-20oC, there's no point in looking at higher temperatures if you are getting inclusion bodies. Induce at OD600 of 0.6.
After all that it is still possible that you get inclusion bodies, some proteins are just difficult. Check your soluble fraction as you might have a small amount of soluble protein. If all fails then looking at other expression systems could be more productive than trying to solve the E.coli expression.
The problem of body inclusions normally are the action of heat shock proteins or some hydrophobic residues in your protein. So, when do you finish your centrifugation to get your cells pellet, put some lysozyme in your buffer before you resuspend your cells and freeze to stock it. Later, perform your lysis,make more pulses of sonication to break the membrane in less time and intensity on ice, put some DTT or beta mercaptoethanol and some tween 20 or triton (solve your problem with hydrophobicity ). if you work well you dont will need to refolding (less work, and activity troubles).
For some proteins you cant get rid of those inclusion bodies. Its very difficult to get rid of them. Always play around with various IPTG concentrations and variable induction temperatures might work. Lysozyme containing buffer and giving more pulses in short time span during sonication (As said by Fabio in earlier answer) are also good ideas.
One more thing always check for the hydrophobicity using various tools (Xtalpred can be very tool for many more such kind of analyses). if your protein is highly hydrophobic, then its better to extract and purify from pellet fractions and then perform the unfolding and refolding in proper conditions for better yields, rather than trying for soluble fractions.
Sorry for insufficient information. I used 0.5 mM IPTG concentration and incubated it near 22degree C for 12 hr. Before sonication, I resuspended pellet into buffer with lysozyme conc. 1mg/ml for 1 hr at ice. But I didn't get the protein in soluble form. I have induced IPTG at 0.5 OD at 600 nm. Then I used Sarcosyl soltion to get protein into soluble form. Now the protein is in soluble form, But it seems like didnt get proper folding as protein is not binding to NiNTA column and coming in unbound sample. What to do, whether dialysis will help for proper folding? Whether i should go for purification one more time after dialysis as i will remove sarcosyl bound to protein and may come in protper folding state.
What concentration of Sarcosyl you are using, you should use low concentration only like 0.1 to .5 % in binding buffer so that it affect the protein bind to column, if you are using Imidazole and glycerol in binding buffer, don't use and use only 150 mM Nacl in binding buffer, pH also important for binding to column use buffer pH close to 7.5 to 8 either 50 mM phosphate or 20- 50 mM Tris HCl, depend on PI of protein . Once your protein of interest bind to column , you can remove nonspecific protein in washing step by increasing volume of wash buffer like more than 10 CV wash. you have to optimize washing step with increasing concentration of IMIDAZOLE. you can try gradient of imidazole for washing and elution. Also try lower concentration of IPTG at lower temperature like 18 C for proper folding of protein. and you can also try other derivative of E. coli BL21 (DE3) competent cell for expression like C41 or C43 or Rossetta origami or Origami, E. Coli strains for expression.
I guess I am also having the same problem with the same strain. So I read some protocols containing urea but could't be sure If it is a problem for the activity...
And also wonder If you came up with a solution at the end.