Are you using reductive NuPage gels in which there's reducing agent in the inner running buffer chamber?  There are some dimeric proteins that will oxidize in situ while running through a standard gel using standard reduction protocols.  These proteins can still associate, even fold to various extents, and form disulfide bridges in the presence of the SDS as the DTT moves away.  Some cysteine knot proteins and immunoblogulin chains are examples.  Having reducing agent in the running buffer throughout the entire run will help prevent this if dimerization is the issue.  Use of TCEP may also be very helpful in initially reducing your sample.  Have you tried to reduce and then alkylate the protein with IAA or NEM before the run?  This will definitely keep the protein in the reduced conformation but with stably alkylated cysteines that can't oxidize.  By the way, it might be interesting to find the pKa and subsequent pH of the Tris in the sample buffer at 90C.  

First, try to use increased amount of DTT - maybe such a robust dimer is linked by S-S bonds.

Regards, Azat.

Are you using reductive NuPage gels in which there's reducing agent in the inner running buffer chamber?  There are some dimeric proteins that will oxidize in situ while running through a standard gel using standard reduction protocols.  These proteins can still associate, even fold to various extents, and form disulfide bridges in the presence of the SDS as the DTT moves away.  Some cysteine knot proteins and immunoblogulin chains are examples.  Having reducing agent in the running buffer throughout the entire run will help prevent this if dimerization is the issue.  Use of TCEP may also be very helpful in initially reducing your sample.  Have you tried to reduce and then alkylate the protein with IAA or NEM before the run?  This will definitely keep the protein in the reduced conformation but with stably alkylated cysteines that can't oxidize.  By the way, it might be interesting to find the pKa and subsequent pH of the Tris in the sample buffer at 90C.  

Hi Ryan,

If you have the equipment to pour your own gels, you can run the gel in 6-8 M urea. Of course you'll have to make all your own gel-making solutions to include that amount of urea, but otherwise it will be a normal gel i.e. use standard or high amounts of SDS and beta-mercaptoethanol.

It's a bit of a mess to make these gels, since every tiny little droplet spilled will crystallize instead of just evaporating, but the resolving power of these gels is so much higher for misbehaving proteins.

Cheers,

Holger

Azat & Grant - I have tried using DTT as well with no avail; probably because between both monomers, there is only 1 Cys residue (unlikely that it's forming a S-S bond with anything). Grant, your thoughts about the pH of the Tris buffer have me wondering, could I use a very low/high pH to denature the dimer?

Holger - My apologies, I should have clarified in my initial post, I have tried using an 8M urea gel before. It looks indistinguishable from the normal SDS-PAGE gels I run in terms of the dimer band.

If there is one cysteine in a monomer molecule then another identical molecule of this monomer would also have a cysteine. If the cysteine in each of these molecules is buried into the center of each molecule's tertiary structure then they are not surface exposed and will likely not react or be able to find another cysteine with which to oxidize to form a disulfide bridge; the usual course of cysteines that are not coordinated with a cofactor or are not surface exposed.  What happens to surface exposed cysteines?  If they are not kept protonated by low pH or blocked by alkylation  or adducted then they will react with another cysteine using oxygen as the electron acceptor. If that other cysteine is within the same molecule then an INTRA-disulfide bridge is formed and the molecule has a stabilized structure.  But if that other cysteine is on another molecule then a disulfide bridge still occurs through oxidation but now a dimer forms and is INTER-disulfide.  Your molecule has only one cysteine so if it is surface exposed and oxidizing to a disulfide then it would be with another identical or non-identical molecule that also has only one cysteine available and that's mainly how dimers are formed. There are unusual ways in which non-reducible dimers form and it involves a tandem repeat that occurs at the genetic level or by desulfurylation of cysteine to form beta alanine in one monomer and then a primary amine from a lysine in another monomer will add across the beta alanine double bond. This is detected as lysinoalanine  in amino acid analysis and occurs in heating conditions.  These are unlikely mechanisms in your case.  If you want to quickly and simply obtain presumptive evidence that the upper band is a dimer of the lower band then take a sample of your protein, denature it, and then force it into dimerization chemically and see if the monomer band gets less and the dimer band increases on SDSPAGE.  If you claim that the monomer has one cysteine and it is not oxidizing to anything then you need to prove it by protein denaturation with 8 M guanidine and then testing for free thiol using  Ellmans reagent at reasonable protein concentration.  

If you want to prove the putative dimer is product related and not a host cell contaminant then Western blotting may help but only if you have a good polyclonal antibody against the product.  Do you have access to MALDI which could be the fastest and simplest mass spec that's accurate enough for you as opposed to single or multi quad LC-MS's.  Your sample appears pure enough and a nice molecular weight that you could spot it with sinapinic acid matrix and get a good TOF for the parent +1 ions.  If the upper band is truly a dimer of the lower band then the parent ion of the dimer should be a nearly exact 2x multiple of the monomer mass.  The +2 mass of the dimer may overlay the +1 mass of the monomer and you should also see the sinapinic adduct for both the monomer and dimer at the expected masses.  It's assumed that you know the theoretical molecular weight since you know there's one cysteine and therefore the amino acid sequence is probably known.  

Is there a chance that your protein belongs the class of intrinsically disordered proteins that are unfolded in their native state? (PMID: 15738986) You can check one of the disorder predictions methods like IUPred to test this on the sequence on our protein (http://iupred.enzim.hu).

Hi Ryan,

You can try a phenol treatment followed by acetone precipitation before loading your sample, as in the publication I attach. There is a simpler version of the protocol: make a 2:1:1 mix of sample/phenol/chloroform, vortex vigorously, add 2.5 vol acetone, leave 1h at -20°C, centrifuge 15 min, recover the pellet in sample buffer + reducing agent. Incubate 2 min 30 s at 100°C. Load on your gel.

Firstly, thank you for all of the feedback, your suggestions have been more than insightful!

Turns out, science is funny - the mass spec results have identified the "dimer" band as a contaminating E. coli protein, namely SlyD, a proline isomerase. The properties of this protein make it funny. My protein (integration host factor, IHF) is a thermal stable, fused to a 6xHis-tag and has no UV trace (does not possess any tryptophans). Unfortunately for me, SlyD is also thermal stable, contains a plethora of His residues and, much to my surprise, does not have a UV trace as well... It seems my purification scheme has selected for the "perfect" contaminant!

Glad you took the mass spec suggestion. Mass spec never lies; just have to interpret the data correctly.  I thought that the MALDI suggestion would give you the fastest answer but did you end up using LC/MS with triple quad?  How close was the contaminant mass to a putative dimer mass?

Classical contaminant in IMAC purification. There's an NEB strain devoid of the corresponding gene and a few other His-rich proteins, called NiCo21. This story emphasizes the need to do several purification steps before studying the protein...

Grant - I'm not sure how the mass spec was performed (I just submitted the sample to the facility at my institution). The SlyD contaminant has a MW of ~20.8 kDa and my heterodimer has a MW of ~23.2 kDa. Though, the contaminating band seemed to migrate larger than my 20 kDa marker (sometimes it even migrated above the 25 kDa marker....weird).

Olivier - You're absolutely right. We performed various controls to validate my protein prep in terms of DNA-binding (i.e. binding site mutations and DNA-binding domain mutations on the protein). However, for protein-protein interaction experiments, we will be re-purifying the protein with a new tag, namely a 2xSteptactin tag. This experience has made me more-self critical and I will definitely be improving the rigor of my purification schemes.

Thanks again everyone, your feedback has been invaluable to me!

Hi Ryan,

Here's a recent article with an example of SlyD contaminant removal by affinity chromatography:

http://www.sciencedirect.com/science/article/pii/S1046592817301651

Hope this helps,

Eric