Hi Zhijie, Thank you so much for the reply. The protein I am working on is from eukaryotic cytosol. We've already tried optimize the IPTG concentration and induction temp. Co-expression of chaperons seems to be a interesting idea, I googled and found this one: http://www.clontech.com/takara/US/Products/Protein_Research/Protein_Folding_and_Expression/Chaperone_Plasmid_Set?sitex=10031:22372:US
Do u think it is good? Or do u have other recommendations? Thank you very much! Best, Xiao On 2013-11-14, at 下午5:14, Zhijie Li wrote: > Hi Xiao, > > The tail sequence does not look very hydrophobic. But until your experiments > is done we can’t really say for sure that is not the cause for the > aggregation. > > If your protein has cys residues then if the protein molecules were misfolded > and were forming intermolecular disulfide links, adding 1-10mM DTT without > denaturation may not be able to resolve the disulfide linked multimers, > because the disulfide could be buried in the misfolded proteins. You can run > SDS PAGE with samples boiled with or without DTT to see if you have > intermolecular disulfide links. > > Since your protein contains cys residues, I wonder what is the source of it. > Is it eukaryotic secreted protein or ER/Golgi protein? If the protein is > eukarytic cytosolic/nuclear protein you do not need to worry about the > oxidative folding of them, and producing in the cytosol of E. coli might be > fine. But if the protein was eukaryotic secreted protein that contain > disulfide bonds, then you have to find an expression system that have the > proper chaperone systems to let the peptide form correct disulfide linkages. > In such case, yeast, insect, or even mammalian expression system should be > tried. > > The linker between your protein and MBP is important. You do need long enough > linker for making sure the two domains are not interfering each other’s > folding. This requirement is protein-specific though. The poly Asn linker > provided by the pMal vectors is normally long enough. According to your > description that your protein family members were not even soluble with other > tags, I think it is quite clear that these proteins are not well behaving in > your current expression system. So unless you detect serious problems in your > construct design, I do not recommend you to invest too much effort on trouble > shooting the linker and so on because this is likely not the main problem > here. > > If you have to make it in e.coli, expression condition has to be optimized to > maximize their chance to fold into soluble proteins. These optimization may > include: 1) slow down the expression by using less IPTG or using lower temp; > 2) to co-expression chaperones(there are many reported cases this works like > magic); 3) to co-express the target protein’s binding partners/cofactors; 4) > to use special expression strains such as the NEB shuffle strain that allows > cytosolic oxidative folding of disulfide-containing proteins, or the rare > codon-expressing strains to resolve rare codon problems if any; 5) to send > your protein to periplasmic space (use pMAL-p vectors instead of pMAL-c) if > you believe the protein should contain disulfides > > Or you can try to refold the protein in vitro. In such case you’d better use > the non-MBP tags, and get the inclusion bodies as the start point. > > But after all, I always prefer to use expression systems that’s most close to > the protein’s natural habitat. There are proteins that would never be folded > in E coli. > > Zhijie > > > > From: Xiao Xiao > Sent: Thursday, November 14, 2013 2:18 AM > To: [email protected] > Subject: Re: [ccp4bb] Oligomerization of maltose-binding protein > > Thank you for the replies! > > Rana: Thank you for your prompt reply! I have the same problem, based on my > SDS-gel, I also have that 42kD MBP band when purifying my fusion protein. > Fortunately I can separate that from my protein of interest by size-exclusion > column. So the major issue I have now is the oligomerization of my fusion > protein. And it is also weird that I got the free-MBP as a large oligomer. > > Zhijie: Thank you for your very detailed and well-written suggestions. It is > very interesting that the MBP fusion protein can form 'micelles'. We found > very similar properties when cloning other family members into MBP vector, > they all have a very large oligomer peak, around 10-mer (several hundreds > kD), but not in void volumn of sup6 column. Biologically these proteins could > form oligomer but not this big. We have tried to purify these protein in > other tags, but most of them were insoluble. Only MBP tag gives us such an > impressive solubility. But I agree -- they may be still misfolded. > > As I replied to Rana, I agree with you that the oligomerized free-MBP is > really strange. That's why I want to figure out the reason, to exclude the > possibility that the tag itself causes aggregation. Reading-frame should be > correct since SDS-gel indicates the peptide length is correct. I didn't put > the stop codon right after precision site, so there are some restriction > cutting sites got translated. The exact C-term should be 27aa > (LEVLFQGP+HMSMGGRDIVDGSEFPAGN). Do you think this could be the reason that > MBP get oligomerized? I am trying to delete these C-term tail to see how it > will behave. > > And the answer for your last question: there is one Cys supposed to be > reduced on the surface of my protein (others are internal and really > conserved), but I include 1mM DTT during purification and have tried even > 10mM DTT when running gel-filtration but it didn't get improved. > > I have one more question. Let's say if the aggregation is really caused by > the property of this construct, could it be possible to help its folding by > modifying linker region between MBP tag and my protein? I know it could be > very tough, but is there any general idea or strategy that we can try as a > start point, such as change hydrophobicity? > > Again, thank you all and I am looking for your responds! > > Best regards, > Xiao > > 2013/11/13 Zhijie Li <[email protected]> > Hi Xiao, > > MBP usually is monomeric, unless you put something really nasty to its ends. > People have mentioned before that due to the high solubility of MBP, MBP tag > can drag otherwise insoluble/overly hydrophobic protein domains into > solution. Then this half hydrophilic half hydrophobic molecule can form > micelle-like structures. > > > "the small peaks are both free-MBP monomer, but the big peaks are fusion > protein and free-MBP respectively." > > So your fusion protein is never found in the monomer peak? That strongly > hints aggregation caused by the domain you fused to MBP. The free MBP monomer > peak you see on the MBP-fusion run is likely a proteolytic product of the > fusion protein, or is the natural MBP from E. coli (although normally far > less than the over expressed MBP fusion in quantity). Are you certain that > your target protein domain is not naturally oligomerizing? If it is not or > only dimerizes or trimerizes, you might need to consider modifying the > construct or expression strategy (including moving away from bacteria if > necessary). I would suggest against trying to rescue an inherently unhappy > construct or trying to make a protein in a system that does not fold it. > > The high MW peak for the "free" MBP is a little strange. I produced MBP-TEV > cleavage site(ENLYFQG) fusion as my control before, never observed any > non-monomer species of it. the precision site does not seem any worse than > the TEV site by sequence. (But did you put a stop codon right after the > precision site? What is the exact sequence of your C-term tail? Also double > check your plasmid sequence for frame-shift mutations, since you have > modified the vector.) Forming heterogeneous aggregates of 10+ the monomer > size is indicative of misfolding if it is not caused by a hydrophobic tail. > For robust proteins such as MBP, such bad aggregation suggests that something > in your expression or purification procedure needs to be optimized. Inducing > for too long (eg, >3hr at 37C) or harsh lysis can contribute to misfolding. > > One more question: does your protein contain cys residues? Are they supposed > to be oxidized or reduced? > > Zhijie > > > > -----Original Message----- From: Xiao Xiao > Sent: Wednesday, November 13, 2013 5:20 PM > To: [email protected] > > Subject: [ccp4bb] Oligomerization of maltose-binding protein > > Dear all, > > I am purifying a fusion protein with maltose-binding protein(MBP) as the > N-term tag. As a control, I also purified free-MBP from empty vector. The > vector I used is pMAL-c5x (from NEB) with an addition of precision protease > cutting site (8 residues, LEVLFQGP) between MBP and multi-cloning site. > > Then I run both protein through Superose6(size-exclusion chromatography) to > check oligomerization states. To my surprise, both protein has two peaks, one > is at the monomer size and another one has a big MW. SDS-gel shows from both > protein, the small peaks are both free-MBP monomer, but the big peaks are > fusion protein and free-MBP respectively. The estimated MW of the big peaks > are similar to 10 folds of their corresponding monomers but heterogeneous > (checked by DLS and native gel). > > To my knowledge, MBP should behave as a monomer. Does anyone have seen the > similar thing before or could give some explanations? I suspect maybe the > precision protease cutting site and the following multi-clonging site at the > C-term of free-MBP might cause oligomerization, could it be possible? I am > trying to delete this part to see whether it will improve, is there any other > thing I can try in order to eliminate heterogeneity? > > Any suggestion or comment will be really appreciated. > > Thank you! > > Best regards, > > Xiao Xiao >
